Publications

Show abstract This review article gives an account of the development of the MR-encephalography (MREG) method, which started as a mere ‘Gedankenexperiment’ in 2005 and gradually developed into a method for ultrafast measurement of physiological activities in the brain. After going through different approaches covering k-space with radial, rosette, and concentric shell trajectories we have settled on a stack-of-spiral trajectory, which allows full brain coverage with (nominal) 3 mm isotropic resolution in 100 ms. The very high acceleration factor is facilitated by the near-isotropic k-space coverage, which allows high acceleration in all three spatial dimensions.

Show abstract Abstract We describe the longest period of subcutaneous EEG (sqEEG) monitoring to date, in a 35-year-old female with refractory epilepsy. Over 230 days, 4791/5520 h of sqEEG were recorded (86\%, mean 20.8 [IQR 3.9] hours/day). Using an electronic diary, the patient reported 22 seizures, while automatically-assisted visual sqEEG review detected 32 seizures. There was substantial agreement between days of reported and recorded seizures (Cohen’s kappa 0.664), although multiple clustered seizures remained undocumented. Circular statistics identified significant sqEEG seizure cycles at circadian (24-hour) and multidien (5-day) timescales. Electrographic seizure monitoring and analysis of long-term seizure cycles are possible with this neurophysiological tool.

Show abstract Kurz, A, Lauber, B, Franke, S, and Leukel, C. Balance training reduces postural sway and improves sport-specific performance in visually impaired cross-country skiers. J Strength Cond Res 35(1): 247--252, 2021---Balance training is highly effective in reducing sport injuries and causes improvements in postural stability and rapid force production. So far, the positive effects of balance training have been described for healthy athletes. In the present experiments, we questioned whether athletes with disabilities of the visual system can also benefit from balance training. Fourteen visually impaired cross-country skiers participated in this randomized controlled study. The intervention group (N = 7) completed 8 sessions of balance training over a period of 4 weeks (2 times per week), whereas a waiting control group (N = 7) received no training during that time. After training, postural sway was significantly reduced in the intervention group but not in the waiting control group. In addition, sport-specific performance, which was assessed by a standardized Cooper's 12-minute test on roller skis or rollerblades, increased in the intervention group. The change in postural sway from the premeasurement to the postmeasurement correlated with the change in sport-specific performance in all subjects. Our results indicate that balance training is useful for improving postural stability and sport-specific performance in visually impaired cross-country skiers. We propose that balance training should therefore be implemented as part of the training routine in athletes with disabilities of the visual system.

Show abstract The rodent medial prefrontal cortex (mPFC) is typically considered to be involved in cognitive aspects of action control, e.g., decision making, rule learning and application, working memory and generally guiding adaptive behavior (Euston, Gruber, & McNaughton, 2012). These cognitive aspects often occur on relatively slow time scales, i.e., in the order of several trials within a block structure (Murakami, Shteingart, Loewenstein, & Mainen, 2017). In this way, the mPFC is able to set up a representational memory (Goldman-Rakic, 1987). On the other hand, the mPFC can also impact action control more directly (i.e., more on the motoric and less cognitive side). This impact on motor control manifests on faster time scales, i.e., on a single trial level (Hardung et al., 2017). While the more cognitive aspects have been reviewed previously as well as in other subchapters of this book, we explicitly focus on the latter aspect in this chapter, particularly on movement inhibition. We discuss models of prefrontal motor interactions, the impact of the behavioral paradigm, evidences for mPFC involvement in action control, and the anatomical connections between mPFC and motor cortex.

Show abstract Meta-learning, or learning to learn, is a machine learning approach that utilizes prior learning experiences to expedite the learning process on unseen tasks. As a data-driven approach, meta-learning requires meta-features that represent the primary learning tasks or datasets, and are estimated traditonally as engineered dataset statistics that require expert domain knowledge tailored for every meta-task. In this paper, first, we propose a meta-feature extractor called Dataset2Vec that combines the versatility of engineered dataset meta-features with the expressivity of meta-features learned by deep neural networks. Primary learning tasks or datasets are represented as hierarchical sets, i.e., as a set of sets, esp. as a set of predictor/target pairs, and then a DeepSet architecture is employed to regress meta-features on them. Second, we propose a novel auxiliary meta-learning task with abundant data called dataset similarity learning that aims to predict if two batches stem from the same dataset or different ones. In an experiment on a large-scale hyperparameter optimization task for 120 UCI datasets with varying schemas as a meta-learning task, we show that the meta-features of Dataset2Vec outperform the expert engineered meta-features and thus demonstrate the usefulness of learned meta-features for datasets with varying schemas for the first time.

Show abstract In case of deafness, cochlear implants bypass dysfunctional or lost hair cells by direct electrical stimulation (eCIs) of the auditory nerve. However, spectral selectivity of eCI sound coding is low as the wide current spread from each electrode activates large sets of neurons that align to a place-frequency (tonotopic) map in the cochlea. As light can be better confined in space, optical cochlear implants (oCIs) promise to overcome this shortcoming of eCIs. This requires fine-grained, fast, and power-efficient real-time sound analysis and control of multiple microscale emitters. Here, we describe the development, characterisation, and application for hearing restoration of a preclinical low-weight and wireless LED-based multichannel oCI system and its companion eCI system. The head-worn oCI system enabled deafened rats to perform a locomotion task in response to acoustic stimulation proving the concept of multichannel opto - genetic hearing restoration in rodents.Competing Interest StatementT.M. is a co-founder of OptoGenTech GmbH.

Show abstract Abstract Deep brain stimulation (DBS) in psychiatric illnesses has been clinically tested over the past 20 years. The clinical application of DBS to the superolateral branch of the medial forebrain bundle in treatment-resistant depressed patients—one of several targets under investigation—has shown to be promising in a number of uncontrolled open label trials. However, there are remain numerous questions that need to be investigated to understand and optimize the clinical use of DBS in depression, including, for example, the relationship between the symptoms, the biological substrates/projections and the stimulation itself. In the context of precision and customized medicine, the current paper focuses on clinical and experimental research of medial forebrain bundle DBS in depression or in animal models of depression, demonstrating how clinical and scientific progress can work in tandem to test the therapeutic value and investigate the mechanisms of this experimental treatment. As one of the hypotheses is that depression engenders changes in the reward and motivational networks, the review looks at how stimulation of the medial forebrain bundle impacts the dopaminergic system.

Show abstract Fiber tractography based on diffusion-weighted MRI provides a non-invasive characterization of the structural connectivity of the human brain at the macroscopic level. Quantification of structural connectivity strength is challenging and mainly reduced to “streamline counting” methods. These are however highly dependent on the topology of the connectome and the particular specifications for seeding and filtering, which limits their intra-subject reproducibility across repeated measurements and, in consequence, also confines their validity. Here we propose a novel method for increasing the intra-subject reproducibility of quantitative estimates of structural connectivity strength. To this end, the connectome is described by a large matrix in positional-orientational space and reduced by Principal Component Analysis to obtain the main connectivity “modes”. It was found that the proposed method is quite robust to structural variability of the data.

Show abstract Military drones (unmanned combat aerial vehicles) combine surveillance technology with missile equipment in a far-reaching way. In this article, I argue that military drones could and should be object for a philosophical investigation, referring in particular on Chamayou's theory of the drone, who also coined the term ``an eye turned into a weapon.'' Focusing on issues of human self-understanding, agency, and alterity, I examine the intricate human-technology relations in the context of designing and deploying military drones. For that purpose, I am drawing on the postphenomenological approach developed by Don Ihde in order to systematize the manifold aspects of human-technology relations in a four-level model (embodiment relations, hermeneutic relations, alterity relations, and background relations). This inquiry also includes a critical reflection on the (often hidden) normative implications of this technology. In doing so, I do not intent to offer an exhaustive relational ontology of military drones. I rather aim at providing a framework that is able to capture the core dimensions of this technology and their complex interrelations in a systematic way that has been missing in the philosophical debate so far.

Show abstract Stieglitz,T.: Was willst du, neue Hand? In: Labisch, M., Neuy, C. (eds.) Tales of Science - Zukunftsgeschichten aus der Mikrosystemtechnik. Ausser der Reihe, Band 50. Winnert: p.machinery Michael Haitel, pp. 134-140 (2020). ISBN 978 3 95765 186 0

Show abstract Purpose A multi-coil shim setup is designed and optimized for human brain shimming. Here, the size and position of a set of square coils are optimized to improve the shim performance without increasing the number of local coils. Utilizing such a setup is especially beneficial at ultrahigh fields where B0 inhomogeneity in the human brain is more severe. Methods The optimization started with a symmetric arrangement of 32 independent coils. Three parameters per coil were optimized in parallel, including angular and axial positions on a cylinder surface and size of the coil, which were constrained by cylinder size, construction consideration, and amplifiers specifications. B0 maps were acquired at 9.4T in 8 healthy volunteers for use as training data. The global and dynamic shimming performance of the optimized multi-coil were compared in simulations and measurements to a symmetric design and to the scanner's second-order shim setup, respectively. Results The optimized multi-coil performs better by 14.7\% based on standard deviation (SD) improvement with constrained global shimming in comparison to the symmetric positioning of the coils. Global shimming performance was comparable with a symmetric 65-channel multi-coil and full fifth-order spherical harmonic shim coils. On average, an SD of 48.4 and 31.9 Hz was achieved for in vivo measurements after global and dynamic slice-wise shimming, respectively. Conclusions An optimized multi-coil shim setup was designed and constructed for human whole-brain shimming. Similar performance of the multi-coils with many channels can be achieved with a fewer number of channels when the coils are optimally arranged around the target.

Show abstract This review presents a practical primer for functional near-infrared spectroscopy (fNIRS) with respect to technology, experimentation, and analysis software. Its purpose is to jump-start interested practitioners considering utilizing a non-invasive, versatile, nevertheless challenging window into the brain using optical methods. We briefly recapitulate relevant anatomical and optical foundations and give a short historical overview. We describe competing types of illumination (trans-illumination, reflectance, and differential reflectance) and data collection methods (continuous wave, time domain and frequency domain). Basic components (light sources, detection, and recording components) of fNIRS systems are presented. Advantages and limitations of fNIRS techniques are offered, followed by a list of very practical recommendations for its use. A variety of experimental and clinical studies with fNIRS are sampled, shedding light on many brain-related ailments. Finally, we describe and discuss a number of freely available analysis and presentation packages suited for data analysis. In conclusion, we recommend fNIRS due to its ever-growing body of clinical applications, state-of-the-art neuroimaging technique and manageable hardware requirements. It can be safely concluded that fNIRS adds a new arrow to the quiver of neuro-medical examinations due to both its great versatility and limited costs.

Show abstract Brain signals represent a communication modality that can allow users of assistive robots to specify high-level goals, such as the object to fetch and deliver. In this paper, we consider a screen-free Brain-Computer Interface (BCI), where the robot highlights candidate objects in the environment using a laser pointer, and the user goal is decoded from the evoked responses in the electroencephalogram (EEG). Having the robot present stimuli in the environment allows for more direct commands than traditional BCIs that require the use of graphical user interfaces. Yet bypassing a screen entails less control over stimulus appearances. In realistic environments, this leads to heterogeneous brain responses for dissimilar objects—posing a challenge for reliable EEG classification. We model object instances as subclasses to train specialized classifiers in the Riemannian tangent space, each of which is regularized by incorporating data from other objects. In multiple experiments with a total of 19 healthy participants, we show that our approach not only increases classification performance but is also robust to both heterogeneous and homogeneous objects. While especially useful in the case of a screen-free BCI, our approach can naturally be applied to other experimental paradigms with potential subclass structure.

Show abstract Background: In accordance with the three R principles of research, animal usage should be limited as much as possible. Especially for the training of entry-level scientists in surgical techniques underlying opto- and electrophysiology, alternative training tools are required before moving on to live animals. We have developed a cost-effective rat brain model for training a wide range of surgical techniques, including, but not limited to optogenetics, electrophysiology, and intracranial pharmacological treatments. Results: Our brain model creates a realistic training experience in animal surgery. The success of the surgeries (e.g. implantation accuracy) is readily assessable in cross sections of the model brain. Moreover, the model allows practicing electrophysiological recordings as well as testing for movement or light related artefacts. Comparison with existing method(s): The surgery and recording experience in our model closely resembles that in an actual rat in terms of the necessary techniques, considerations and time span. A few differences to an actual rat brain slightly reduce the difficulty in our model compared to a live animal. Thus, entry level scientists can first learn basic techniques in our model before moving on to the slightly more complex procedures in live animals. Conclusions: Our brain model is a useful training tool to equip scientist who are new in the field of electrophysiology and optogenetic manipulations with a basic skill set before applying it in live animals. It can be adapted to fit the desired training content or even to serve in testing and optimizing new lab equipment for more senior scientists.

Show abstract Seizures are a disruption of normal brain activity present across a vast range of species and conditions. We introduce an organizing principle that leads to the first objective Taxonomy of Seizure Dynamics (TSD) based on bifurcation theory. The ‘dynamotype’ of a seizure is the dynamic composition that defines its observable characteristics, including how it starts, evolves and ends. Analyzing over 2000 focal-onset seizures from multiple centers, we find evidence of all 16 dynamotypes predicted in TSD. We demonstrate that patients’ dynamotypes evolve during their lifetime and display complex but systematic variations including hierarchy (certain types are more common), non-bijectivity (a patient may display multiple types) and pairing preference (multiple types may occur during one seizure). TSD provides a way to stratify patients in complement to present clinical classifications, a language to describe the most critical features of seizure dynamics, and a framework to guide future research focused on dynamical properties.

Show abstract We propose Adaptive Curriculum Generation from Demonstrations (ACGD) for reinforcement learning in the presence of sparse rewards. Rather than designing shaped reward functions, ACGD adaptively sets the appropriate task difficulty for the learner by controlling where to sample from the demonstration trajectories and which set of simulation parameters to use. We show that training vision-based control policies in simulation while gradually increasing the difficulty of the task via ACGD improves the policy transfer to the real world. The degree of domain randomization is also gradually increased through the task difficulty. We demonstrate zero-shot transfer for two real-world manipulation tasks: pick-and-stow and block stacking.

Show abstract Introduction The aim of our study was to evaluate intracerebral network changes in epilepsy patients demonstrating secondary bilateral synchrony (SBS) in EEG by applying a new Diffusion Tensor Imaging (DTI) method using an energy-based global tracking algorithm. Materials and methods 10 MRI negative epilepsy patients demonstrating SBS in 10–20 surface EEG were included. EEG findings were analyzed for irritative zones characterized by focal interictal epileptiform discharges (IEDs) triggering SBS. In addition, DTI including an energy-based global tracking algorithm was applied to analyze fiber tract alterations in irritative zones. To measure the deviation of a certain cortical connection in comparison to healthy controls, normalized differences of fiber tract streamline counts (SC) and their p-values were evaluated in comparison to corresponding fibers of the control group. Results In 6 patients the irritative zone initiating SBS was located in the frontal lobe, in 3 patients in the temporal lobe and in 1 patient in the region surrounding the right central sulcus. All patients demonstrated significantly altered SC in brain lobes where the irritative zone triggering SBS was located (p ≤ 0.05). Seven out of 10 patients demonstrated SC alterations in tracts connecting brain lobes between the ipsilateral and the contralateral hemisphere (p ≤ 0.05). Conclusion Our data demonstrate that alterations in fiber tracts in irritative zones triggering SBS are not necessarily associated with intracerebral lesions visible in high resolution MRI. Our study gives evidence that diffusion tensor imaging is a promising non-invasive additive tool for intracerebral network analyses even in MRI-negative epilepsy patients.

Show abstract {Machine learning as a field of artificial intelligence is increasingly applied in medicine to assist patients and physicians. Growing datasets provide a sound basis with which to apply machine learning methods that learn from previous experiences. This review explains the basics of machine learning and its subfields of supervised learning, unsupervised learning, reinforcement learning and deep learning. We provide an overview of current machine learning applications in rheumatology, mainly supervised learning methods for e-diagnosis, disease detection and medical image analysis. In the future, machine learning will be likely to assist rheumatologists in predicting the course of the disease and identifying important disease factors. Even more interestingly, machine learning will probably be able to make treatment propositions and estimate their expected benefit (e.g. by reinforcement learning). Thus, in future, shared decision-making will not only include the patient’s opinion and the rheumatologist’s empirical and evidence-based experience, but it will also be influenced by machine-learned evidence.}

Show abstract Objective. High-frequency-oscillations (HFO) and interictal-epileptic-spikes (IES) are spatial biomarkers of the epileptogenic-zone. Those HFO spatially and temporally co-occurring with IES (IES-HFO) are potentially superior biomarkers, their use is however challenged by the difficulty in detecting the low amplitude HFO riding the high-amplitude and steep-waveform of IES. We aim to develop an automatic HFO detector with an improved performance with respect to current methods and that also correctly distinguishes IES-HFO from IES occurring in isolation (isol-IES). We also aim to validate the biomarker-value of the automatic detections by utilizing them to localize a surrogate of the epileptogenic-zone. Approach. We developed automatic-detectors of HFO-Ripples (80–250 Hz), HFO-FastRipples (250–500 Hz) and IES based on kernelized support-vector-machines (SVM). The training of the HFO-detectors was based on visually marked HFO and the parameter optimization during this training-process promoted the correct discernment between IES-HFO and isol-IES. Both HFO-detectors were benchmarked against other published detectors using databases with both visually marked and simulated gold-standards. The IES-detector was trained with a publicly available simulated database and tested against both simulated and visually marked gold-standards. To validate the detections' biomarker-value, the detectors were run on intracranial-EEGs from 8 patients and each with durations of 2–3 days, the detections' cumulated per-channel occurrence-rate was then used to predict the seizure-onset-zone as a surrogate of the epileptogenic-zone. Main results. The HFO-detectors obtained at least 21 more F1-score points than previously published algorithms at the lowest signal-to-noise-ratio. The success achieved when discerning between IES-HFO and isol-IES was comparable to that of other published algorithms. The automatically detected IES-HFO and IES-Ripples showed the best biomarker-value to localize the epileptogenic-zone. Significance. The developed detectors are publicly available and provide a reliable tool to further study HFO, IES-HFO and their value as biomarkers. The putative higher biomarker value from IES-HFO was validated on automatically-detected, long-term data.

Show abstract OBJECTIVE: High-frequency-oscillations (HFO) and interictal-epileptic-spikes (IES) are spatial biomarkers of the epileptogenic-zone. Those HFO spatially and temporally co-occurring with IES (IES-HFO) are potentially superior biomarkers, their use is however challenged by the difficulty in detecting the low amplitude HFO riding the high-amplitude and steep-waveform of IES. We aim to develop an automatic HFO detector with an improved performance with respect to current methods and that also correctly distinguishes IES-HFO from IES occurring in isolation (isol-IES). We also aim to validate the biomarker-value of the automatic detections by utilizing them to localize a surrogate of the epileptogenic-zone. APPROACH: We developed automatic-detectors of HFO-Ripples (80-250 Hz), HFO-FastRipples (250-500 Hz) and IES based on kernelized support-vector-machines (SVM). The training of the HFO-detectors was based on visually marked HFO and the parameter optimization during this training-process promoted the correct discernment between IES-HFO and isol-IES. Both HFO-detectors were benchmarked against other published detectors using databases with both visually marked and simulated gold-standards. The IES-detector was trained with a publicly available simulated database and tested against both simulated and visually marked gold-standards. To validate the detections' biomarker-value, the detectors were run on intracranial-EEGs from 8 patients and each with durations of 2-3 days, the detections' cumulated per-channel occurrence-rate was then used to predict the seizure-onset-zone as a surrogate of the epileptogenic-zone. MAIN RESULTS: The HFO-detectors obtained at least 21 more F1-score points than previously published algorithms at the lowest signal-to-noise-ratio. The success achieved when discerning between IES-HFO and isol-IES was comparable to that of other published algorithms. The automatically detected IES-HFO and IES-Ripples showed the best biomarker-value to localize the epileptogenic-zone. SIGNIFICANCE: The developed detectors are publicly available and provide a reliable tool to further study HFO, IES-HFO and their value as biomarkers. The putative higher biomarker value from IES-HFO was validated on automatically-detected, long-term data.

Show abstract Abstract Epilepsy is a heterogeneous condition with disparate etiologies and phenotypic and genotypic characteristics. Clinical and research aspects are accordingly varied, ranging from epidemiological to molecular, spanning clinical trials and outcomes, gene and drug discovery, imaging, electroencephalography, pathology, epilepsy surgery, digital technologies, and numerous others. Epilepsy data are collected in the terabytes and petabytes, pushing the limits of current capabilities. Modern computing firepower and advances in machine and deep learning, pioneered in other diseases, open up exciting possibilities for epilepsy too. However, without carefully designed approaches to acquiring, standardizing, curating, and making available such data, there is a risk of failure. Thus, careful construction of relevant ontologies, with intimate stakeholder inputs, provides the requisite scaffolding for more ambitious big data undertakings, such as an epilepsy data commons. In this review, we assess the clinical and research epilepsy landscapes in the big data arena, current challenges, and future directions, and make the case for a systematic approach to epilepsy big data.

Show abstract Cortico-basal ganglia beta oscillations (13-30Hz) are assumed to be involved in motor impairments in Parkinson{\textquoteright}s Disease (PD), especially in bradykinesia and rigidity. Various studies have utilized the unilateral 6-OHDA rat PD model to further investigate PD and test novel treatments. However, a detailed behavioral and electrophysiological characterization of the model, including analyses of popular PD treatments such as DBS, has not been documented in the literature. We hence challenged the 6-OHDA rat PD model with a series of experiments (i.e. cylinder test, open field test and rotarod test) aimed at assessing the motor impairments, analyzing the effects of Deep Brain Stimulation (DBS), and identifying under which conditions excessive beta oscillations occur. We found that hemi-PD rats presented an impaired performance in all experiments compared to the sham group, and DBS could improve their overall performance. Across all the experiments and behaviors, the power in the high beta band was observed to be an important biomarker for PD as it showed differences between healthy and lesioned hemispheres and between PD and sham rats. This all shows that the 6-OHDA PD model accurately represents many of the motor and electrophysiological symptoms of PD and makes it a useful tool for the pre-clinical testing of new treatments and further investigations into this disease.Competing Interest StatementThe authors have declared no competing interest.

Show abstract Background: Speed of performance improvements and the strength of memory consolidation in humans vary with movement expertise. Underlying neural mechanisms of behavioural differences between levels of movement expertise are so far unknown. New method: In this study, PET with [18F]fluorodeoxyglucose (FDG) was proposed as a powerful novel methodology to assess learning-related brain activity patterns during large non-restricted movements (ball throwing with a right hand). 24 male handball players ('Experts') and 24 male participants without handball experience ('Novices') performed visuomotor adaptations to prismatic glasses with or without strategic manoeuvres (i.e., explicit or implicit adaptation). Results: Regional changes in FDG uptake as a marker of neuronal activity, relative to a control condition, were assessed. Prismatic adaptation, in general, was associated with decreased occipital neuronal activity as a possible response to misleading visual information. In 'Experts', the adaptation was associated with altered neuronal activity in a network comprising the right parietal cortex and the left cerebellum. In 'Novices', implicit adaptation resulted in an activation of the middle frontal and inferior temporal gyrus. Comparison with existing methods: This study demonstrates the versatility of FDG PET for studying brain activations patterns in experimental settings with unrestricted movements that are not accessible by other techniques (e.g., fMRI or EEG). Conclusions: Observed results are consistent with the involvement of different functional networks related to strategic manoeuvres and expertise levels. This strengthens the assumption of different mechanisms underlying behavioural changes associated with movement expertise. Furthermore, the present study underscores the value of FDG PET for studying brain activation patterns during unrestricted movements. Keywords: Explicit; FDG PET; Handball; Implicit; Motor expertise; Visuomotor adaptation. Copyright © 2020 Elsevier B.V. All rights reserved.

Show abstract In Alzheimer’s disease, the tauopathy is known as a major mechanism responsible for the development of cognitive deficits. Early biomarkers of such affectations for diagnosis/stratification are crucial in Alzheimer’s disease research, and brain connectome studies increasingly show their potential establishing pathology fingerprints at the network level. In this context, we conducted an in vivo multimodal MRI study on young Thy-Tau22 transgenic mice expressing tauopathy, performing resting state functional MRI and structural brain imaging to identify early connectome signatures of the pathology, relating with histological and behavioural investigations. In the prodromal phase of tauopathy, before the emergence of cognitive impairments, Thy-Tau22 mice displayed selective modifications of brain functional connectivity involving three main centres: hippocampus (HIP), amygdala (AMG) and the isocortical areas, notably the somatosensory (SS) cortex. Each of these regions showed differential histopathological profiles. Disrupted ventral HIP-AMG functional pathway and altered dynamic functional connectivity were consistent with high pathological tau deposition and astrogliosis in both hippocampus and amygdala, and significant microglial reactivity in amygdalar nuclei. These patterns were concurrent with widespread functional hyperconnectivity of memory-related circuits of dorsal hippocampus—encompassing dorsal HIP-SS communication—in the absence of significant cortical histopathological markers. These findings suggest the coexistence of two intermingled mechanisms of response at the functional connectome level in the early phases of pathology: a maladaptive and a likely compensatory response. Captured in the connectivity patterns, such first responses to pathology could further be used in translational investigations as a lead towards an early biomarker of tauopathy as well as new targets for future treatments.

Show abstract Tracking the pose of a camera is at the core of visual localization methods used in many applications. As the observations of a camera are inherently affected by lighting, it has always been a challenge for these methods to cope with varying lighting conditions. Thus far, this issue has mainly been approached with the intent to increase robustness by choosing lighting invariant map representations. In contrast, our work aims at explicitly exploiting lighting effects for camera tracking. To achieve this, we propose a lighting adaptable map representation for indoor environments that allows real-time rendering of the scene illuminated by an arbitrary subset of the lamps contained in the model. Our method for estimating the light setting from the current camera observation enables us to adapt the model according to the lighting conditions present in the scene. As a result, lighting effects like cast shadows do no longer act as disturbances that demand robustness but rather as beneficial features when matching observations against the map. We leverage these capabilities in a direct dense camera tracking approach and demonstrate its performance in realworld experiments in scenes with varying lighting conditions.

Show abstract Leveraging machine learning methods for analyzing large amounts of digitized health-related data is one of the fastest growing areas of biomedical research and innovation. At the same time, these emerging technologies create important ethical and legal tensions. The aim of this chapter is to discuss these tensions, focusing on the following three issues: (i) increasing amount of health-related digital data creates challenges for protecting individual and group privacy; (ii) intelligent neurotechnological systems may adversely affect human agency, autonomy, and personal identity; and (iii) inherent biases in the data structures and ontologies may be replicated or amplified by algorithmic decision-support systems. In light of these tensions, multistakeholder discourse and deliberation are required to ensure effective and responsible development and implementation of these emerging technologies.

Show abstract In this lab meeting, six feminist scholars who engage with the sciences from various perspectives and have been collaborating over the last decade as members of the NeuroGenderings Network, share a sustained discussion on the responsibilities of a feminist scientist—particularly in light of our current moment. In a time when ongoing acts of anti-Black racism and police brutality have converged with a global pandemic and anti-science movements, we ask ourselves, how do we express solidarity and also hold ourselves accountable at the crossroads of science and social justice?

Show abstract Preclinical applications of resting-state functional magnetic resonance imaging (rsfMRI) offer the possibility to non-invasively probe whole-brain network dynamics and to investigate the determinants of altered network signatures observed in human studies. Mouse rsfMRI has been increasingly adopted by numerous laboratories worldwide. Here we describe a multi-centre comparison of 17 mouse rsfMRI datasets via a common image processing and analysis pipeline. Despite prominent cross-laboratory differences in equipment and imaging procedures, we report the reproducible identification of several large-scale resting-state networks (RSN), including a mouse default-mode network, in the majority of datasets. A combination of factors was associated with enhanced reproducibility in functional connectivity parameter estimation, including animal handling procedures and equipment performance. RSN spatial specificity was enhanced in datasets acquired at higher field strength, with cryoprobes, in ventilated animals, and under medetomidine-isoflurane combination sedation. Our work describes a set of representative RSNs in the mouse brain and highlights key experimental parameters that can critically guide the design and analysis of future rodent rsfMRI investigations.

Show abstract This paper reports on the development, characterization and in vivo validation of compact optical neural probes. These novel intracerebral devices comprise micro light-emitting diodes (μLEDs) integrated along their slender probe shanks with up to 20 μLEDs per device. Blue light with a peak wavelength of 455 nm is emitted from circular apertures 100 μm in diameter. The μLEDs are structured on GaN-on-sapphire wafers and subsequently transferred onto silicon (Si) carrier wafers. The wafer-scale transfer process provides the opportunity to process the functional GaN layer stack from both sides and hence enables maximizing the efficiency of the μLEDs. Combined with standard MEMS fabrication processes for Si, linear μLED arrays with small inter-μLED distances are achieved on thin probe shanks with cross-sections measuring 150 μm × 65 μm. Devices are interconnected using highly flexible polyimide cables in order to mechanically decouple them from the peripheral electronics during in vivo experiments. Assembled probes emit a peak optical radiant flux of 440 μW (emittance 56 mW mm-2) at 5 mA driving current. Thermal characterization of test probes reveals a temperature increase of 1.5K measured using an integrated thermistor. Electrical functionality stress tests have been carried out to evaluate the device passivation against the physiological environment. It is estimated to endure at least 48 h during continuously pulsed μLED operation. A compact driving circuitry enables low-noise μLED operation in in vivo optogenetic experiments. The radiant flux necessary to elicit an acceptable neuronal response is determined between 1.36 μW and 17.5μW. Probe validation successfully demonstrates the layer-specific stimulation in the cortex in multiple in vivo trials.

Show abstract This paper reports on the fabrication, assembly, characterization and validation of a novel opto-electrical cardiac stimulator designed to augment a mechanical pacing device. The integration of miniaturized electrodes and blue light-emitting diode (LED) chips on the pacer tip with a diameter of 1 mm enables the application of multimodal stimuli in one location on the surface of isolated murine hearts. The opto-electrical stimulator is based on two separate polyimide (PI) substrates each with a thickness of 10 μm combined into a functional unit based on dedicated assembly and encapsulation processes using silicone rubber. The experimental validation in isolated, whole hearts compares electrical, optical and mechanical stimuli exerted at frequencies of up to 8 Hz on Langen-dorff-perfused hearts expressing channelrhodopsin-2. The integrated iridium oxide electrodes implemented above the LED chips enable simultaneous electrical recordings of local cardiac electrical activity.

Show abstract This paper reports on the fabrication, assembly, characterization and validation of a novel opto-electrical cardiac stimulator designed to augment a mechanical pacing device. The integration of miniaturized electrodes and blue light-emitting diode (LED) chips on the pacer tip with a diameter of 1 mm enables the application of multimodal stimuli in one location on the surface of isolated murine hearts. The opto-electrical stimulator is based on two separate polyimide (PI) substrates each with a thickness of 10 μm combined into a functional unit based on dedicated assembly and encapsulation processes using silicone rubber. The experimental validation in isolated, whole hearts compares electrical, optical and mechanical stimuli exerted at frequencies of up to 8 Hz on Langen-dorff-perfused hearts expressing channelrhodopsin-2. The integrated iridium oxide electrodes implemented above the LED chips enable simultaneous electrical recordings of local cardiac electrical activity.

Show abstract Structural biocompatibility is a fundamental requirement for chronically stable bioelectronic devices. Newest neurotechnologies are increasingly focused on minimizing the foreign body response through the development of devices that match the mechanical properties of the implanted tissue and mimic its surface composition, often compromising on their robustness. In this study, an analytical approach is proposed to determine the threshold of conformability for polyimide-based electrocorticography devices. A finite element model was used to quantify the depression of the cortex following the application of devices mechanically above or below conformability threshold. Findings were validated in vivo on rat animal models. Impedance measurements were performed for 40 days after implantation to monitor the status of the biotic/abiotic interface with both conformable and non-conformable implants. Multi-unit activity was then recorded for 12 weeks after implantation using the most compliant device type. It can therefore be concluded that conformability is an essential prerequisite for steady and reliable implants which does not only depend on the Young's modulus of the device material: it strongly relies on the relation between tissue curvature at the implantation site and corresponding device's thickness and geometry, which eventually define the moment of inertia and the interactions at the material-tissue interface.

Show abstract [This corrects the article DOI: 10.3389/fnins.2020.00543.]. Keywords: HARDI; MR microscopy; hippocampus; kainate; mesial temporal lobe epilepsy; tractography. Copyright © 2020 Göbel-Guéniot, Gerlach, Kamberger, Leupold, von Elverfeldt, Hennig, Korvink, Haas and LeVan.

Show abstract Objective To study motor cortex plasticity after a period of training with a new prototype of bidirectional hand prosthesis in three left trans-radial amputees, correlating these changes with the modification of Phantom Limb Pain (PLP) in the same period. Methods Each subject underwent a brain motor mapping with Transcranial Magnetic Stimulation (TMS) and PLP evaluation with questionnaires during a six-month training with a prototype of bidirectional hand prosthesis. Results The baseline motor maps showed in all three amputees a smaller area of muscles representation of the amputated side compared to the intact limb. After training, there was a partial reversal of the baseline asymmetry. The two subjects affected by PLP experienced a statistically significant reduction of pain. Conclusions Two apparently opposite findings, the invasion of the “deafferented” cortex by neighbouring areas and the “persistence” of neural structures after amputation, could vary according to different target used for measurement. Our results do not support a correlation between PLP and motor cortical changes. Significance The selection of the target and of the task is essential for studies investigating motor brain plasticity. This study boosts against a direct and unique role of motor cortical changes on PLP genesis.

Show abstract This paper reports on the customized thinning of neural probes based on silicon (Si) using deep reactive ion etching (DRIE) as a post-processing step. The reduced probe dimensions are expected to minimize local tissue trauma, while guaranteeing probe integrity during implantation. For DRIE, the probes are partially masked by a micromachined Si cover chip comprising tailored cavities enabling any desired thinned length l and probe thickness t by a proper choice of cover chip design and DRIE parameters, respectively. A broad variety of probe designs were realized with shank tip thicknesses ranging from 35 μm down to 2 μm. All probes could successfully be implanted into a brain tissue phantom, demonstrating a pronounced reduction in insertion force from 0.55 mN for unprocessed probes to 0.08 mN for 2-μm-thin shanks. When the dura mater was mimicked by a polyethylene (PE) membrane, forces were reduced from 28.9 mN to 16.6 mN for 15-μm-thin shanks.

Show abstract OBJECTIVE: A major challenge in neuroprosthetics is the restoration of sensory-motor hand functions in upper-limb amputees. Neuroprostheses based on the direct re-connection of the peripheral nerves may be an interesting approach for re-establishing the natural and effective bidirectional control of hand prostheses. Recent results have shown that transverse intrafascicular multi-channel electrodes (TIMEs) can restore natural and sophisticated sensory feedback. However, the potential of using TIME-recorded motor intraneural signals to decode grasping tasks has not as yet been explored. APPROACH: In this study, we show that several hand-movement intentions can be decoded from intraneural signals recorded using four TIMEs implanted in the median and ulnar nerves of an upper limb amputee. Experimental sessions were performed over a week, from day 16 to day 23 after the surgical operation. Intraneural activity was recorded during several hand motor tasks imagined by the subject and processed offline. MAIN RESULTS: We obtained a very high decoding accuracy considering 11 class states (up to 83%). These results confirm that neural signals recorded by multi-channel intraneural electrodes can be used to decode several movement intentions with high accuracy. Moreover, we were able to use same TIME channels for decoding over one week within the first month, even if the stability has to be confirmed during long-term experiments. SIGNIFICANCE: Therefore, TIMEs could be used in the future to achieve a complete bidirectional approach exploiting neural pathways, to make a more natural and intuitive new generation of hand prostheses that have a closer resemblance to a healthy hand.

Show abstract BACKGROUND: Deep brain stimulation (DBS) of the medial forebrain bundle (MFB) can reverse depressive-like symptoms clinically and in experimental models of depression, but the mechanisms of action are unknown. OBJECTIVES: This study investigated the role of dopaminergic mechanisms in MFB stimulation-mediated behavior changes, in conjunction with raclopride administration and micropositron emission tomography (micro-PET). METHODS: Flinders Sensitive Line (FSL) rats were allocated into 4 groups: FSL (no treatment), FSL+ (DBS), FSL.R (FSL with raclopride), and FSL.R+ (FSL with raclopride and DBS). Animals were implanted with bilateral electrodes targeting the MFB and given 11 days access to raclopride in the drinking water with or without concurrent continuous bilateral DBS over the last 10 days. Behavioral testing was conducted after stimulation. A PET scan using [18F]desmethoxyfallypride was performed to determine D2 receptor availability before and after raclopride treatment. Changes in gene expression in the nucleus accumbens and the hippocampus were assessed using quantitative polymerase chain reaction. RESULTS: Micro-PET imaging showed that raclopride administration blocked 36% of the D2 receptor in the striatum, but the relative level of blockade was reduced/modulated by stimulation. Raclopride treatment enhanced depressive-like symptoms in several tasks, and the MFB DBS partially reversed the depressive-like phenotype. The raclopride-treated MFB DBS animals had increased levels of mRNA coding for dopamine receptor D1 and D2 suggestive of a stimulation-mediated increase in dopamine receptors. CONCLUSION: Data suggest that chronic and continuous MFB DBS could act via the modulation of the midbrain dopaminergic transmission, including impacting on the postsynaptic dopamine receptor profile.

Show abstract Popular Maximum Entropy Inverse Reinforcement Learning approaches require the computation of expected state visitation frequencies for the optimal policy under an estimate of the reward function. This usually requires intermediate value estimation in the inner loop of the algorithm, slowing down convergence considerably. In this work, we introduce a novel class of algorithms that only needs to solve the MDP underlying the demonstrated behavior once to recover the expert policy. This is possible through a formulation that exploits a probabilistic behavior assumption for the demonstrations within the structure of Q-learning. We propose Inverse Actionvalue Iteration which is able to fully recover an underlying reward of an external agent in closed-form analytically. We further provide an accompanying class of sampling-based variants which do not depend on a model of the environment. We show how to extend this class of algorithms to continuous state-spaces via function approximation and how to estimate a corresponding action-value function, leading to a policy as close as possible to the policy of the external agent, while optionally satisfying a list of predefined hard constraints. We evaluate the resulting algorithms called Inverse Action-value Iteration, Inverse Q-learning and Deep Inverse Qlearning on the Objectworld benchmark, showing a speedup of up to several orders of magnitude compared to (Deep) Max-Entropy algorithms. We further apply Deep Constrained Inverse Q-learning on the task of learning autonomous lane-changes in the open-source simulator SUMO achieving competent driving after training on data corresponding to 30 minutes of demonstrations.

Show abstract Understanding the semantic characteristics of the environment is a key enabler for autonomous robot operation. In this paper, we propose a deep convolutional neural network (DCNN) for semantic segmentation of a LiDAR scan into the classes car, pedestrian and bicyclist. This architecture is based on dense blocks and efficiently utilizes depth separable convolutions to limit the number of parameters while still maintaining the state-of-the-art performance. To make the predictions from the DCNN temporally consistent, we propose a Bayes filter based method. This method uses the predictions from the neural network to recursively estimate the current semantic state of a point in a scan. This recursive estimation uses the knowledge gained from previous scans, thereby making the predictions temporally consistent and robust towards isolated erroneous predictions. We compare the performance of our proposed architecture with other state-of-the-art neural network architectures and report substantial improvement. For the proposed Bayes filter approach, we shows results on various sequences in the KITTI tracking benchmark.

Show abstract The dentate gyrus (DG) has a key role in hippocampal memory formation. Intriguingly, DG lesions impair many, but not all, hippocampus-dependent mnemonic functions, indicating that the rest of the hippocampus (CA1-CA3) can operate autonomously under certain conditions. An extensive body of theoretical work has proposed how the architectural elements and various cell types of the DG may underlie its function in cognition. Recent studies recorded and manipulated the activity of different neuron types in the DG during memory tasks and have provided exciting new insights into the mechanisms of DG computational processes, particularly for the encoding, retrieval and discrimination of similar memories. Here, we review these DG-dependent mnemonic functions in light of the new findings and explore mechanistic links between the cellular and network properties of, and the computations performed by, the DG.

Show abstract Visual neglect and extinction are two distinct visuospatial attention deficits that frequently occur after right hemisphere cerebral stroke. However, their different lesion profiles remain a matter of debate. In the left hemisphere, a domain-general dual-loop model with distinct computational abilities onto which several cognitive functions may project, has been proposed: a dorsal stream for sensori-motor mapping in time and space and a ventral stream for comprehension and representation of concepts. We wondered whether such a distinction may apply to visual extinction and neglect in left hemisphere lesions. Of 165 prospectively studied patients with acute left hemispheric ischemic stroke with a single lesion on MRI, 122 had no visuospatial attention deficit, 10 had extinction, 31 neglect and 2 had both, visual extinction and neglect. Voxel-based-lesion-symptom mapping (VLSM, FDR<.05) showed a clear anatomical dissociation. Extinction occurred after damage to the parietal cortex (anterior bank of the intraparietal sulcus, inferior parietal lobe, and supramarginal gyrus), while visual neglect occurred after damage mainly to the temporal lobe (superior and middle temporal lobe, anterior temporal pole), inferior ventral premotor cortex, frontal operculum, angular gyrus, and insula. Direct comparison of both conditions linked extinction to intraparietal sulcus and supramarginal gyrus (FDR<.05). Thus, in the left hemisphere extinction seems to be related to dorsal stream lesions, whereas neglect maps more on the ventral stream. These data cannot be generalized to the right hemisphere. However, a domain-general point-of-view may stimulate discussion on visuospatial attention processing also in the right hemisphere.

Show abstract Introduction Increased expression of hyperphosphorylated tau and the formation of neurofibrillary tangles are associated with neuronal loss and white matter damage. Using high resolution ex vivo diffusion tensor imaging (DTI), we investigated microstructural changes in the white and grey matter in the P301L mouse model of human tauopathy at 8.5 months-of-age. For unbiased computational analysis, we implemented a pipeline for voxel-based analysis (VBA) and atlas-based analysis (ABA) of DTI mouse brain data.Methods Hemizygous and homozygous transgenic P301L mice and non-transgenic littermates were used. DTI data were acquired for generation of fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), axial diffusivity (AD) maps. VBA on the entire brain were performed using SPM8 and SPM Mouse toolbox. Initially, all DTI maps were co-registered with Allen mouse brain atlas to bring them to one common coordinate space. In VBA, co-registered DTI maps were normalized and smoothed in order to perform two-sample t-tests to compare hemizygotes with non-transgenic littermates, homozygotes with non-transgenic littermates, hemizygotes with homozygotes on each DTI parameter map. In ABA, the average values for selected regions-of-interests were computed with co-registered DTI maps and labels in Allen mouse brain atlas. After, the same two-sample t-tests were executed on the estimated average values.Results We made reconstructed DTI data and VBA and ABA pipeline publicly available. With VBA, we found microstructural changes in the white matter in hemizygous P301L mice compared to non-transgenic littermates. In contrast, more pronounced and brain-wide spread changes were observed in VBA when comparing homozygous P301L mice with non-transgenic littermates. Statistical comparison of DTI metrics in selected brain regions by ABA corroborated findings from VBA. FA was found to be decreased in most brain regions, while MD, RD and AD were increased compared to hemizygotes and non-transgenic littermates.Discussion/Conclusion High resolution ex vivo DTI demonstrated brain-wide microstructural changes in the P301L mouse model of human tauopathy. The comparison between hemizygous and homozygous P301L mice revealed a gene-dose dependent effect on DTI metrics. The publicly available computational data analysis pipeline can provide a platform for future mechanistic and longitudinal studies.Competing Interest StatementThe authors have declared no competing interest.

Show abstract The common pipeline in autonomous driving systems is highly modular and includes a perception component which extracts lists of surrounding objects and passes these lists to a high-level decision component. In this case, leveraging the benefits of deep reinforcement learning for high-level decision making requires special architectures to deal with multiple variable-length sequences of different object types, such as vehicles, lanes or traffic signs. At the same time, the architecture has to be able to cover interactions between traffic participants in order to find the optimal action to be taken. In this work, we propose the novel Deep Scenes architecture, that can learn complex interaction-aware scene representations based on extensions of either 1) Deep Sets or 2) Graph Convolutional Networks. We present the Graph-Q and DeepScene-Q off-policy reinforcement learning algorithms, both outperforming state-ofthe-art methods in evaluations with the publicly available traffic simulator SUMO.

Show abstract {The loss and recovery of language functions are still incompletely understood. This longitudinal functional MRI study investigated the neural mechanisms underlying language recovery in patients with post-stroke aphasia putting particular emphasis on the impact of lesion site. To identify patterns of language-related activation, an auditory functional MRI sentence comprehension paradigm was administered to patients with circumscribed lesions of either left frontal (n = 17) or temporo-parietal (n = 17) cortex. Patients were examined repeatedly during the acute (≤1 week, t1), subacute (1–2 weeks, t2) and chronic phase (\\&gt;6 months, t3) post-stroke; healthy age-matched control subjects (n = 17) were tested once. The separation into two patient groups with circumscribed lesions allowed for a direct comparison of the contributions of distinct lesion-dependent network components to language reorganization between both groups. We hypothesized that activation of left hemisphere spared and perilesional cortex as well as lesion-homologue cortex in the right hemisphere varies between patient groups and across time. In addition, we expected that domain-general networks serving cognitive control independently contribute to language recovery. First, we found a global network disturbance in the acute phase that is characterized by reduced functional MRI language activation including areas distant to the lesion (i.e. diaschisis) and subsequent subacute network reactivation (i.e. resolution of diaschisis). These phenomena were driven by temporo-parietal lesions. Second, we identified a lesion-independent sequential activation pattern with increased activity of perilesional cortex and bilateral domain-general networks in the subacute phase followed by reorganization of left temporal language areas in the chronic phase. Third, we observed involvement of lesion-homologue cortex only in patients with frontal but not temporo-parietal lesions. Fourth, irrespective of lesion location, language reorganization predominantly occurred in pre-existing networks showing comparable activation in healthy controls. Finally, we detected different relationships of performance and activation in language and domain-general networks demonstrating the functional relevance for language recovery. Our findings highlight that the dynamics of language reorganization clearly depend on lesion location and hence open new perspectives for neurobiologically motivated strategies of language rehabilitation, such as individually-tailored targeted application of neuro-stimulation.}

Show abstract A network consisting of excitatory and inhibitory (EI) neurons is a canonical model for understanding local cortical network activity. In this study, we extended the local circuit model and investigated how its dynamical landscape can be enriched when it interacts with another excitatory (E) population with long transmission delays. Through analysis of a rate model and numerical simulations of a corresponding network of spiking neurons, we studied the transition from stationary to oscillatory states by analyzing the Hopf bifurcation structure in terms of two network parameters: (1) transmission delay between the EI subnetwork and the E population and (2) inhibitory couplings that induced oscillatory activity in the EI subnetwork. We found that the critical coupling strength can strongly modulate as a function of transmission delay, and consequently the stationary state can be interwoven intricately with the oscillatory state. Such a dynamical landscape gave rise to an isolated stationary state surrounded by multiple oscillatory states that generated different frequency modes, and cross-frequency coupling developed naturally at the bifurcation points. We identified the network motifs with short- and long-range inhibitory connections that underlie the emergence of oscillatory states with multiple frequencies. Thus, we provided a mechanistic explanation of how the transmission delay to and from the additional E population altered the dynamical landscape. In summary, our results demonstrated the potential role of long-range connections in shaping the network activity of local cortical circuits.

Show abstract A network consisting of excitatory and inhibitory (EI) neurons is a canonical model for understanding local cortical network activity. In this study, we extended the local circuit model and investigated how its dynamical landscape can be enriched when it interacts with another excitatory (E) population with long transmission delays. Through analysis of a rate model and numerical simulations of a corresponding network of spiking neurons, we studied the transition from stationary to oscillatory states by analyzing the Hopf bifurcation structure in terms of two network parameters: (1) transmission delay between the EI subnetwork and the E population and (2) inhibitory couplings that induced oscillatory activity in the EI subnetwork. We found that the critical coupling strength can strongly modulate as a function of transmission delay, and consequently the stationary state can be interwoven intricately with the oscillatory state. Such a dynamical landscape gave rise to an isolated stationary state surrounded by multiple oscillatory states that generated different frequency modes, and cross-frequency coupling developed naturally at the bifurcation points. We identified the network motifs with short- and long-range inhibitory connections that underlie the emergence of oscillatory states with multiple frequencies. Thus, we provided a mechanistic explanation of how the transmission delay to and from the additional E population altered the dynamical landscape. In summary, our results demonstrated the potential role of long-range connections in shaping the network activity of local cortical circuits.

Show abstract In the field of Neuroelectronic Interfaces it seems as though the lines between reality and science-fiction/fantasy are often blurred. One of the inspirations for our most recent Gordon Research Conference in March 2018 aimed at “Bridging the Gap in Neuroelectronic Interfaces” dates back to 1999 when Chapin et al. (1999) described their ability to predict movement trajectories of rodents and non-human primates by “eavesdropping” on groups of neurons. Many in the field felt that science fiction seemed to become reality and the future of prosthetics appeared on the horizon. Less than a decade later invasive micro-electrode arrays and the latest jewels of micro-machining found their way into the brains of a few human patients as well (Hochberg et al., 2006), very much triggering expectations of a coming Golden Age of Brain-Machine-Interfacing for severely handicapped patients.

Show abstract Resilience to stress is critical for the development of depression. Enhanced adenosine A1 receptor (A1R) signaling mediates the antidepressant effects of acute sleep deprivation (SD). However, chronic SD causes long-lasting upregulation of brain A1R and increases the risk of depression. To investigate the effects of A1R on mood, we utilized two transgenic mouse lines with inducible A1R overexpression in forebrain neurons. These two lines have identical levels of A1R increase in the cortex, but differ in the transgenic A1R expression in the hippocampus. Switching on the transgene promotes robust antidepressant and anxiolytic effects in both lines. The mice of the line without transgenic A1R overexpression in the hippocampus (A1Hipp-) show very strong resistance towards development of stress-induced chronic depression-like behavior. In contrast, the mice of the line in which A1R upregulation extends to the hippocampus (A1Hipp+), exhibit decreased resilience to depression as compared to A1Hipp-. Similarly, automatic analysis of reward behavior of the two lines reveals that depression resistant A1Hipp-transgenic mice exhibit high sucrose preference, while mice of the vulnerable A1Hipp + line developed stress-induced anhedonic phenotype. The A1Hipp + mice have increased Homer1a expression in hippocampus, correlating with impaired long-term potentiation in the CA1 region, mimicking the stressed mice. Furthermore, virus-mediated overexpression of Homer1a in the hippocampus decreases stress resilience. Taken together our data indicate for first time that increased expression of A1R and Homer1a in the hippocampus modulates the resilience to stress-induced depression and thus might potentially mediate the detrimental effects of chronic sleep restriction on mood.

Show abstract Spatial reasoning is essential for an agent's navigation and the cognitive processing of abstract arrangements. Meta-analyses of neuroimaging data reveal that both the right posterior parietal cortex and left dorsolateral prefrontal cortex (PPC and DLPFC, respectively) show increased activation during spatial relational reasoning. To investigate whether participants' reasoning performance can be modified and potentially enhanced, anodal transcranial direct current stimulation (tDCS) was applied over either region. 51 healthy adult participants solved spatial reasoning problems after the application of either anodal tDCS over the right PPC, the left DLPFC or a sham stimulation. We expect anodal stimulation to enhance cortical excitability which would be reflected by enhanced reasoning performance in participants receiving stimulation. The results demonstrate that anodal stimulation applied over the right PPC enhances participants' performance in indeterminate reasoning problems, compared to sham and anodal stimulation over the left DLPFC. This finding is highly relevant for clarifying the cognitive mechanisms of relational reasoning and for clinical applications, e.g., enhancing or restoring higher cognitive functions for spatial representation and reasoning.

Show abstract Social distancing measures due to the COVID-19 pandemic have accelerated the adoption and implementation of digital mental health tools. Psychiatry and therapy sessions are being conducted via videoconferencing platforms, and the use of digital mental health tools for monitoring and treatment has grown. This rapid shift to telehealth during the pandemic has given added urgency to the ethical challenges presented by digital mental health tools. Regulatory standards have been relaxed to allow this shift to socially distanced mental health care. It is imperative to ensure that the implementation of digital mental health tools, especially in the context of this crisis, is guided by ethical principles and abides by professional codes of conduct. This paper examines key areas for an ethical path forward in this digital mental health revolution: privacy and data protection, safety and accountability, and access and fairness.

Show abstract Key points Discrete and rhythmic dynamics are inherent components of (human) movements. We provide evidence that distinct human motor cortex circuits contribute to discrete and rhythmic movements. Excitability of supragranular layer circuits of the human motor cortex was higher during discrete movements than during rhythmic movements. Conversely, more complex corticospinal circuits showed higher excitability during rhythmic movements than during discrete movements. No task-specific differences existed for corticospinal output neurons at infragranular layers. The excitability differences were found to be time(phase)-specific and could not be explained by the kinematic properties of the movements. The same task-specific differences were found between the last cycle of a rhythmic movement period and ongoing rhythmic movements. Abstract Human actions entail discrete and rhythmic movements (DM and RM, respectively). Recent insights from human and animal studies indicate different neural control mechanisms for DM and RM, emphasizing the intrinsic nature of the task. However, how distinct human motor cortex circuits contribute to these movements remains largely unknown. In the present study, we tested distinct primary motor cortex and corticospinal circuits and proposed that they show differential excitability between DM and RM. Human subjects performed either 1) DM or 2) RM using their right wrist. We applied an advanced electrophysiological approach involving transcranial magnetic stimulation and peripheral nerve stimulation to test the excitability of the neural circuits. Probing was performed at different movement phases: movement initiation (MI, 20 ms after EMG onset) and movement execution (ME, 200 ms after EMG onset) of the wrist flexion. At MI, excitability at supragranular layers was significantly higher in DM than in RM. Conversely, excitability of more complex corticospinal circuits was significantly lower in DM than RM at ME. No task-specific differences were found for direct corticospinal output neurons at infragranular layers. The neural differences could not be explained by the kinematic properties of the movements and also existed between ongoing RM and the last cycle of RM. Our results therefore strengthen the hypothesis that different neural control mechanisms engage in DM and RM.

Show abstract Objective: Report simple reference structure fabrication and validate the precise localization of subdural micro- and standard electrodes in magnetic resonance imaging (MRI) in phantom experiments. Approach: Electrode contacts with diameters of 0.3 mm and 4 mm are localized in 1.5 T MRI using reference structures made of silicone and iron oxide nanoparticle doping. The precision of the localization procedure was assessed for several standard MRI sequences and implant orientations in phantom experiments and compared to common clinical localization procedures. Main results: A localization precision of 0.41 ± 0.20 mm could be achieved for both electrode diameters compared to 1.46 ± 0.69 mm that was achieved for 4 mm standard electrode contacts localized using a common clinical standard method. The new reference structures are intrinsically bio-compatible, and they can be detected with currently available feature detection software so that a clinical implementation of this technology should be feasible. Significance: Neuropathologies are increasingly diagnosed and treated with subdural electrodes, where the exact localization of the electrode contacts with respect to the patient's cortical anatomy is a prerequisite for the procedure. Post-implantation electrode localization using MRI may be advantageous compared to the common alternative of CT-MRI image co-registration, as it avoids systematic localization errors associated with the co-registration itself, as well as brain shift and implant movement. Additionally, MRI provides superior soft tissue contrast for the identification of brain lesions without exposing the patient to ionizing radiation. Recent studies show that smaller electrodes and high-density electrode grids are ideal for clinical and research purposes, but the localization of these devices in MRI has not been demonstrated.

Show abstract We present a novel, to the best of our knowledge, fabrication process for highly aspherical lenses based on surface deformation due to thermal expansion of a soft polymer, polydimethylsiloxane (PDMS), using laser-structuring, molding, and precise shape optimization. Our fabrication process can be used for almost any lens shape with a large degree of freedom---both individual lenses and dense arrays. We present the design, fabrication, and characterization with examples of four different lenses with 1 mm apertures and surface deviations below 100 nm.

Show abstract Flexible, cloth-like, polymer-derived carbon fiber mats, that are embedded in polyimide, are presented in article number 1900713 by Swati Sharma, Thomas Stieglitz and co-workers. This scalable, monolithic manufacturing method eliminates any joints or metal interconnects and creates compliable, implantable electrocorticography (ECoG) electrode arrays that show good biocompatibility and performance in rat models.

Show abstract One-shot imitation is the vision of robot programming from a single demonstration, rather than by tedious construction of computer code. We present a practical method for realizing one-shot imitation for manipulation tasks, exploiting modern learning-based optical flow to perform real-time visual servoing. Our approach, which we call FlowControl, continuously tracks a demonstration video, using a specified foreground mask to attend to an object of interest. Using RGB-D observations, FlowControl requires no 3D object models, and is easy to set up. FlowControl inherits great robustness to visual appearance from decades of work in optical flow. We exhibit FlowControl on a range of problems, including ones requiring very precise motions, and ones requiring the ability to generalize.

Show abstract Abstract Objective Seizure unpredictability is rated as one of the most challenging aspects of living with epilepsy. Seizure likelihood can be influenced by a range of environmental and physiological factors that are difficult to measure and quantify. However, some generalizable patterns have been demonstrated in seizure onset. A majority of people with epilepsy exhibit circadian rhythms in their seizure times, and many also show slower, multiday patterns. Seizure cycles can be measured using a range of recording modalities, including self-reported electronic seizure diaries. This study aimed to develop personalized forecasts from a mobile seizure diary app. Methods Forecasts based on circadian and multiday seizure cycles were tested pseudoprospectively using data from 50 app users (mean of 109 seizures per subject). Individuals' strongest cycles were estimated from their reported seizure times and used to derive the likelihood of future seizures. The forecasting approach was validated using self-reported events and electrographic seizures from the Neurovista dataset, an existing database of long-term electroencephalography that has been widely used to develop forecasting algorithms. Results The validation dataset showed that forecasts of seizure likelihood based on self-reported cycles were predictive of electrographic seizures for approximately half the cohort. Forecasts using only mobile app diaries allowed users to spend an average of 67.1\% of their time in a low-risk state, with 14.8\% of their time in a high-risk warning state. On average, 69.1\% of seizures occurred during high-risk states and 10.5\% of seizures occurred in low-risk states. Significance Seizure diary apps can provide personalized forecasts of seizure likelihood that are accurate and clinically relevant for electrographic seizures. These results have immediate potential for translation to a prospective seizure forecasting trial using a mobile diary app. It is our hope that seizure forecasting apps will one day give people with epilepsy greater confidence in managing their daily activities.

Show abstract The increasing awareness of the impact of spontaneous movements on neuronal activity has raised the need to track behavior. We present FreiPose, a versatile learning-based framework to directly capture 3D motion of freely definable points with high precision (median error < 3.5% body length, 41.9% improvement compared to state-of-the-art) and high reliability (82.8% keypoints within < 5% body length error boundary, 72.0% improvement). The versatility of FreiPose is demonstrated in two experiments: (1) By tracking freely moving rats with simultaneous electrophysiological recordings in motor cortex, we identified neuronal tuning to behavioral states and individual paw trajectories. (2) We inferred time points of optogenetic stimulation in rat motor cortex from the measured pose across individuals and attributed the stimulation effect automatically to body parts. The versatility and accuracy of FreiPose open up new possibilities for quantifying behavior of freely moving animals and may lead to new ways of setting up experiments.

Show abstract The increasing awareness of the impact of spontaneous movements on neuronal activity has raised the need to track behavior. We present FreiPose, a versatile learning-based framework to directly capture 3D motion of freely definable points with high precision (median error \&lt; 3.5\% body length, 41.9\% improvement compared to state-of-the-art) and high reliability (82.8\% keypoints within \&lt; 5\% body length error boundary, 72.0\% improvement). The versatility of FreiPose is demonstrated in two experiments: (1) By tracking freely moving rats with simultaneous electrophysiological recordings in motor cortex, we identified neuronal tuning to behavioral states and individual paw trajectories. (2) We inferred time points of optogenetic stimulation in rat motor cortex from the measured pose across individuals and attributed the stimulation effect automatically to body parts. The versatility and accuracy of FreiPose open up new possibilities for quantifying behavior of freely moving animals and may lead to new ways of setting up experiments.

Show abstract The vertebrate brain comprises a plethora of cell types connected by intertwined pathways. Optogenetics enriches the neuroscientific tool set for disentangling these neuronal circuits in a manner which exceeds the spatio-temporal precision of previously existing techniques. Technically, optogenetics can be divided in three types of optical and genetic combinations: (1) it is primarily understood as the manipulation of the activity of genetically modified cells (typically neurons) with light, i.e. optical actuators. (2) A second combination refers to visualizing the activity of genetically modified cells (again typically neurons), i.e. optical sensors. (3) A completely different interpretation of optogenetics refers to the light activated expression of a genetically induced construct. Here, we focus on the first two types of optogenetics, i.e. the optical actuators and sensors in an attempt to give an overview into the topic. We first cover methods to express opsins into neurons and introduce strategies of targeting specific neuronal populations in different animal species. We then summarize combinations of optogenetics with behavioral read out and neuronal imaging. Finally, we give an overview of the current state-of-the-art and an outlook on future perspectives.

Show abstract Recent experiments have shown that neural stimulation can successfully restore sensory feedback in upper-limb amputees improving their ability to control the prosthesis. However, the potential advantages of invasive sensory feedback with respect to non-invasive solutions have not been yet identified. Our hypothesis was that a difference would appear when the subject cannot focus all the attention to the use of the prosthesis, but some additional activities require his/her cognitive attention, which is a quite common situation in real-life conditions. To verify this hypothesis, we asked a trans-radial amputee, equipped with a bidirectional hand prosthesis, to perform motor tasks also in combination with a cognitive task. Sensory feedback was provided via intraneural (invasive) or electro-tactile (non-invasive) stimulation. We collected also data related to self-confidence. While both approaches were able to significantly improve the motor performance of the subject when no additional cognitive effort was asked, the manual accuracy was not affected by the cognitive task only when intraneural feedback was provided. The highest self-confidence was obtained when intraneural sensory feedback was provided. Our findings show that intraneural sensory feedback is more robust to dual tasks than non-invasive feedback. This is the first direct comparison between invasive and non-invasive approaches for restoring sensory feedback and it could suggest an advantage of using invasive solutions.Clinical Trial Registration:<ext-link ext-link-type="uri" xlink:href="http://www.ClinicalTrials.gov" xmlns:xlink="http://www.w3.org/1999/xlink">www.ClinicalTrials.gov</ext-link>, identifier NCT02848846.

Show abstract The performance of gradient-based optimization strategies depends heavily on the initial weights of the parametric model. Recent works show that there exist weight initializations from which optimization procedures can find the task-specific parameters faster than from uniformly random initializations and that such a weight initialization can be learned by optimizing a specific model architecture across similar tasks via MAML (Model-Agnostic Meta-Learning). Current methods are limited to populations of classification tasks that share the same number of classes due to the static model architectures used during meta-learning. In this paper, we present HIDRA, a meta-learning approach that enables training and evaluating across tasks with any number of target variables. We show that Model-Agnostic Meta-Learning trains a distribution for all the neurons in the output layer and a specific weight initialization for the ones in the hidden layers. HIDRA explores this by learning one master neuron, which is used to initialize any number of output neurons for a new task. Extensive experiments on the Miniimagenet and Omniglot data sets demonstrate that HIDRA improves over standard approaches while generalizing to tasks with any number of target variables. Moreover, our approach is shown to robustify low-capacity models in learning across complex tasks with a high number of classes for which regular MAML fails to learn any feasible initialization.

Show abstract Objective. The analysis of interactions among local populations of neurons in the cerebral cortex (e.g. within cortical microcolumns) requires high resolution and high channel count recordings from chronically implanted laminar microelectrode arrays. The request for high-density recordings of a large number of recording sites can presently only be accomplished by probes realized using complementary metal-oxide-semiconductor (CMOS) technology. In preparation for their use in non-human primates, we aimed for neural probe validation in a head-fixed approach analyzing the long-term recording capability. Approach. We examined chronically implanted silicon-based laminar probes, realized using a CMOS technology in combination with micromachining, to record from the primary visual cortex (V1) of a monkey. We used a passive CMOS probe that had 128 electrodes arranged at a pitch of 22.5 µm in four columns and 32 rows on a slender shank. In order to validate the performance of a dedicated microdrive, the overall dimensions of probe and interface boards were chosen to be compatible with the final active CMOS probe comprising integrated circuitry. Main results. Using the passive probe, we recorded simultaneously local field potentials (LFP) and spiking multiunit activity (MUA) in V1 of an awake behaving macaque monkey. We found that an insertion through the dura and subsequent readjustments of the chronically implanted neural probe was possible and allowed us to record stable LFPs for more than five months. The quality of MUA degraded within the first month but remained sufficiently high to permit mapping of receptive fields during the full recording period. Significance. We conclude that the passive silicon probe enables semi-chronic recordings of high quality of LFP and MUA for a time span exceeding five months. The new microdrive compatible with a commercial recording chamber successfully demonstrated the readjustment of the probe position while the implemented plug structure effectively reduced brain tissue movement relative to the probe.

Show abstract The interest in dry electroencephalography (EEG) electrodes has increased in recent years, especially as everyday suitability earplugs for measuring drowsiness or focus of auditory attention. However, the challenge is still the need for a good electrode material, which is reliable and can be easily processed for highly personalized applications. Laser processing, as used here, is a fast and very precise method to produce personalized electrode configurations that meet the high requirements of in-ear EEG electrodes. The arrangement of the electrodes on the flexible and compressible mats allows an exact alignment to the ear mold and contributes to high wearing comfort, as no edges or metal protrusions are present. For better transmission properties, an adapted coating process for surface enlargement of platinum electrodes is used, which can be controlled precisely. The resulting porous platinum-copper alloy is chemically very stable, shows no exposed copper residues, and enlarges the effective surface area by 40. In a proof-of-principle experiment, these porous platinum electrodes could be used to measure the Berger effect in a dry state using just one ear of a test person. Their signal-to-noise ratio and the frequency transfer function is comparable to gel-based silver/silver chloride electrodes.

Show abstract A key challenge for an agent learning to interact with the world is to reason about physical properties of objects and to foresee their dynamics under the effect of applied forces. In order to scale learning through interaction to many objects and scenes, robots should be able to improve their own performance from real-world experience without requiring human supervision. To this end, we propose a novel approach for modeling the dynamics of a robot's interactions directly from unlabeled 3D point clouds and images. Unlike previous approaches, our method does not require ground-truth data associations provided by a tracker or any pre-trained perception network. To learn from unlabeled real-world interaction data, we enforce consistency of estimated 3D clouds, actions and 2D images with observed ones. Our joint forward and inverse network learns to segment a scene into salient object parts and predicts their 3D motion under the effect of applied actions. Moreover, our object-centric model outputs action-conditioned 3D scene flow, object masks and 2D optical flow as emergent properties. Our extensive evaluation both in simulation and with real-world data demonstrates that our formulation leads to effective, interpretable models that can be used for visuomotor control and planning.

Show abstract Summary The current study was designed to further clarify the influence of brain morphology, sleep oscillatory activity and age on memory consolidation. Specifically, we hypothesized, that a smaller volume of hippocampus, parahippocampal and medial prefrontal cortex negatively impacts declarative, but not procedural, memory consolidation. Explorative analyses were conducted to demonstrate whether a decrease in slow-wave activity negatively impacts declarative memory consolidation, and whether these factors mediate age effects on memory consolidation. Thirty-eight healthy participants underwent an acquisition session in the evening and a retrieval session in the morning after night-time sleep with polysomnographic monitoring. Declarative memory was assessed with the paired-associate word list task, while procedural memory was tested using the mirror-tracing task. All participants underwent high-resolution magnetic resonance imaging. Participants with smaller hippocampal, parahippocampal and medial prefrontal cortex volumes displayed a reduced overnight declarative, but not procedural memory consolidation. Mediation analyses showed significant age effects on overnight declarative memory consolidation, but no significant mediation effects of brain morphology on this association. Further mediation analyses showed that the effects of age and brain morphology on overnight declarative memory consolidation were not mediated by polysomnographic variables or sleep electroencephalogram spectral power variables. Thus, the results suggest that the association between age, specific brain area volume and overnight memory consolidation is highly relevant, but does not necessarily depend on slow-wave sleep as previously conceptualized.

Show abstract Mesial temporal lobe epilepsy (MTLE) is the most common form of focal, pharmacoresistant epilepsy in adults and is often associated with hippocampal sclerosis. Here, we established the efficacy of optogenetic and electrical low-frequency stimulation (LFS) in interfering with seizure generation in a mouse model of MTLE. Specifically, we applied LFS in the sclerotic hippocampus to study the effects on spontaneous subclinical and evoked generalized seizures. We found that stimulation at 1 Hz for 1 hr resulted in an almost complete suppression of spontaneous seizures in both hippocampi. This seizure-suppressive action during daily stimulation remained stable over several weeks. Furthermore, LFS for 30 min before a pro-convulsive stimulus successfully prevented seizure generalization. Finally, acute slice experiments revealed a reduced efficacy of perforant path transmission onto granule cells upon LFS. Taken together, our results suggest that hippocampal LFS constitutes a promising approach for seizure control in MTLE.

Show abstract Mesial temporal lobe epilepsy (MTLE) is the most common form of focal, pharmacoresistant epilepsy in adults and is often associated with hippocampal sclerosis. Here, we established the efficacy of optogenetic and electrical low-frequency stimulation (LFS) in interfering with seizure generation in a mouse model of MTLE. Specifically, we applied LFS in the sclerotic hippocampus to study the effects on spontaneous subclinical and evoked generalized seizures. We found that stimulation at 1 Hz for 1 hr resulted in an almost complete suppression of spontaneous seizures in both hippocampi. This seizure-suppressive action during daily stimulation remained stable over several weeks. Furthermore, LFS for 30 min before a pro-convulsive stimulus successfully prevented seizure generalization. Finally, acute slice experiments revealed a reduced efficacy of perforant path transmission onto granule cells upon LFS. Taken together, our results suggest that hippocampal LFS constitutes a promising approach for seizure control in MTLE.

Show abstract Mesial temporal lobe epilepsy (MTLE) is the most common type of focal epilepsy. It is frequently associated with abnormal MRI findings, which are caused by underlying cellular, structural, and chemical changes at the micro-scale. In the current study, it is investigated to which extent these alterations correspond to imaging features detected by high resolution magnetic resonance imaging in the intrahippocampal kainate mouse model of MTLE. Fixed hippocampal and whole-brain sections of mouse brain tissue from nine animals under physiological and chronically epileptic conditions were examined using structural and diffusion-weighted MRI. Microstructural details were investigated based on a direct comparison with immunohistochemical analyses of the same specimen. Within the hippocampal formation, diffusion streamlines could be visualized corresponding to dendrites of CA1 pyramidal cells and granule cells, as well as mossy fibers and Schaffer collaterals. Statistically significant changes in diffusivities, fractional anisotropy, and diffusion orientations could be detected in tissue samples from chronically epileptic animals compared to healthy controls, corresponding to microstructural alterations (degeneration of pyramidal cells, dispersion of the granule cell layer, and sprouting of mossy fibers). The diffusion parameters were significantly correlated with histologically determined cell densities. These findings demonstrate that high-resolution diffusion-weighted MRI can resolve subtle microstructural changes in epileptic hippocampal tissue corresponding to histopathological features in MTLE.

Show abstract Mesial temporal lobe epilepsy (MTLE) is the most common type of focal epilepsy. It is frequently associated with abnormal MRI findings, which are caused by underlying cellular, structural, and chemical changes at the micro-scale. In the current study, it is investigated to which extent these alterations correspond to imaging features detected by high resolution magnetic resonance imaging in the intrahippocampal kainate mouse model of MTLE. Fixed hippocampal and whole-brain sections of mouse brain tissue from nine animals under physiological and chronically epileptic conditions were examined using structural and diffusion-weighted MRI. Microstructural details were investigated based on a direct comparison with immunohistochemical analyses of the same specimen. Within the hippocampal formation, diffusion streamlines could be visualized corresponding to dendrites of CA1 pyramidal cells and granule cells, as well as mossy fibers and Schaffer collaterals. Statistically significant changes in diffusivities, fractional anisotropy, and diffusion orientations could be detected in tissue samples from chronically epileptic animals compared to healthy controls, corresponding to microstructural alterations (degeneration of pyramidal cells, dispersion of the granule cell layer, and sprouting of mossy fibers). The diffusion parameters were significantly correlated with histologically determined cell densities. These findings demonstrate that high-resolution diffusion-weighted MRI can resolve subtle microstructural changes in epileptic hippocampal tissue corresponding to histopathological features in MTLE.

Show abstract Brain networks store new memories using functional and structural synaptic plasticity. Memory formation is generally attributed to Hebbian plasticity, while homeostatic plasticity is thought to have an ancillary role in stabilizing network dynamics. Here we report that homeostatic plasticity alone can also lead to the formation of stable memories. We analyze this phenomenon using a new theory of network remodeling, combined with numerical simulations of recurrent spiking neural networks that exhibit structural plasticity based on firing rate homeostasis. These networks are able to store repeatedly presented patterns and recall them upon the presentation of incomplete cues. Storing is fast, governed by the homeostatic drift. In contrast, forgetting is slow, driven by a diffusion process. Joint stimulation of neurons induces the growth of associative connections between them, leading to the formation of memory engrams. In conclusion, homeostatic structural plasticity induces a specific type of “silent memories”, different from conventional attractor states.

Show abstract Brain networks store new memories using functional and structural synaptic plasticity. Memory formation is generally attributed to Hebbian plasticity, while homeostatic plasticity is thought to have an ancillary role in stabilizing network dynamics. Here we report that homeostatic plasticity alone can also lead to the formation of stable memories. We analyze this phenomenon using a new theory of network remodeling, combined with numerical simulations of recurrent spiking neural networks that exhibit structural plasticity based on firing rate homeostasis. These networks are able to store repeatedly presented patterns and recall them upon the presentation of incomplete cues. Storing is fast, governed by the homeostatic drift. In contrast, forgetting is slow, driven by a diffusion process. Joint stimulation of neurons induces the growth of associative connections between them, leading to the formation of memory engrams. In conclusion, homeostatic structural plasticity induces a specific type of {\textquotedblleft}silent memories{\textquotedblright}, different from conventional attractor states.

Show abstract The identification of oscillatory neural markers of Parkinson’s disease (PD) can contribute not only to the understanding of functional mechanisms of the disorder, but may also serve in adaptive deep brain stimulation (DBS) systems. These systems seek online adaptation of stimulation parameters in closed-loop as a function of neural markers, aiming at improving treatment’s efficacy and reducing side effects. Typically, the identification of PD neural markers is based on group-level studies. Due to the heterogeneity of symptoms across patients, however, such group-level neural markers, like the beta band power of the subthalamic nucleus, are not present in every patient or not informative about every patient’s motor state. Instead, individual neural markers may be preferable for providing a personalized solution for the adaptation of stimulation parameters. Fortunately, data-driven bottom-up approaches based on machine learning may be utilized. These approaches have been developed and applied successfully in the field of brain-computer interfaces with the goal of providing individuals with means of communication and control. In our contribution, we present results obtained with a novel supervised data-driven identification of neural markers of hand motor performance based on a supervised machine learning model. Data of 16 experimental sessions obtained from seven PD patients undergoing DBS therapy show that the supervised patient-specific neural markers provide improved decoding accuracy of hand motor performance, compared to group-level neural markers reported in the literature. We observed that the individual markers are sensitive to DBS therapy and thus, may represent controllable variables in an adaptive DBS system.

Show abstract Electroencephalogram data used in the domain of brain-computer interfaces typically has subpar signal-to-noise ratio and data acquisition is expensive. An effective and commonly used classifier to discriminate event-related potentials is the linear discriminant analysis which, however, requires an estimate of the feature distribution. While this information is provided by the feature covariance matrix its large number of free parameters calls for regularization approaches like Ledoit-Wolf shrinkage. Assuming that the noise of event-related potential recordings is not time-locked, we propose to decouple the time component from the covariance matrix of event-related potential data in order to further improve the estimates of the covariance matrix for linear discriminant analysis. We compare three regularized variants thereof and a feature representation based on Riemannian geometry against our proposed novel linear discriminant analysis with time-decoupled covariance estimates. Extensive evaluations on 14 electroencephalogram datasets reveal, that the novel approach increases the classification performance by up to four percentage points for small training datasets, and gracefully converges to the performance of standard shrinkage-regularized LDA for large training datasets. Given these results, practitioners in this field should consider using our proposed time-decoupled covariance estimation when they apply linear discriminant analysis to classify event-related potentials, especially when few training data points are available.

Show abstract Robots perceive their environment using various sensor modalities, e.g., vision, depth, sound or touch. Each modality provides complementary information for perception. However, while it can be assumed that all modalities are available for training, when deploying the robot in real-world scenarios the sensor setup often varies. In order to gain flexibility with respect to the deployed sensor setup we propose a new multimodal approach within the framework of contrastive learning. In particular, we consider the case of learning from RGB-D images while testing with one modality available, i.e., exclusively RGB or depth. We leverage contrastive learning to capture high-level information between different modalities in a compact feature embedding. We extensively evaluate our multimodal contrastive learning method on the Falling Things dataset and learn representations that outperform prior methods for RGB-D object recognition on the NYU-D dataset. Our code and details on the used datasets are available at: https://github.com/meyerjo/MultiModalContrastiveLearning.

Show abstract The human temporal lobe is a multimodal association area which plays a key role in various aspects of cognition, particularly in memory formation and spatial navigation. Functional and anatomical connectivity of temporal structures is thus a subject of intense research, yet by far underexplored in humans due to ethical and technical limitations. We assessed intratemporal cortico-cortical interactions in the living human brain by means of single pulse electrical stimulation, an invasive method allowing mapping effective intracortical connectivity with a high spatiotemporal resolution. Eighteen subjects with normal anterior-mesial temporal MR imaging undergoing intracranial presurgical epilepsy diagnostics with multiple depth electrodes were included. The investigated structures were temporal pole, hippocampus, amygdala and parahippocampal gyrus. Intratemporal cortical connectivity was assessed as a function of amplitude of the early component of the cortico-cortical evoked potentials (CCEP). While the analysis revealed robust interconnectivity between all explored structures, a clear asymmetry in bidirectional connectivity was detected for the hippocampal network and for the connections between the temporal pole and parahippocampal gyrus. The amygdala showed bidirectional asymmetry only to the hippocampus. The provided evidence of asymmetrically weighed intratemporal effective connectivity in humans in vivo is important for understanding of functional interactions within the temporal lobe since asymmetries in the brain connectivity define hierarchies in information processing. The findings are in exact accord with the anatomical tracing studies in non-human primates and open a translational route for interventions employing modulation of temporal lobe function.

Show abstract Rationale: Patients with dual pathology have two potentially epileptogenic lesions: One in the hippocampus and one in the neocortex. If epilepsy surgery is considered, stereotactic electroencephalography (SEEG) may reveal which of the lesions is seizure-generating, but frequently, some uncertainty remains. We aimed to investigate whether interictal high-frequency oscillations (HFOs), which are a promising biomarker of epileptogenicity, are associated with the primary focus. Methods: We retrospectively analyzed 16 patients with dual pathology. They were grouped according to their seizure-generating lesion, as suggested by ictal SEEG. An automated detector was applied to identify interictal epileptic spikes, ripples (80-250 Hz), ripples co-occurring with spikes (IES-ripples) and fast ripples (250-500 Hz). We computed a ratio R to obtain an indicator of whether rates were higher in the hippocampal lesion (R close to 1), higher in the neocortical lesion (R close to -1), or more or less similar (R close to 0). Results: Spike and HFO rates were higher in the hippocampal than in the neocortical lesion (p < 0.001), particularly in seizure onset zone channels. Seizures originated exclusively in the hippocampus in 5 patients (group 1), in both lesions in 7 patients (group 2), and exclusively in the neocortex in 4 patients (group 3). We found a significant correlation between the patients' primary focus and the ratio R(fast ripples), i.e., the proportion of interictal fast ripples detected in this lesion (p < 0.05). No such correlation was observed for interictal epileptic spikes (p = 0.69), ripples (p = 0.60), and IES-ripples (p = 0.54). In retrospect, interictal fast ripples would have correctly "predicted" the primary focus in 69% of our patients (p < 0.01). Conclusions: We report a correlation between interictal fast ripple rate and the primary focus, which was not found for epileptic spikes. Fast ripple analysis could provide helpful information for generating a hypothesis on seizure-generating networks, especially in cases with few or no recorded seizures.

Show abstract Systemic inflammation is associated with alterations in complex brain functions such as learning and memory. However, diagnostic approaches to functionally assess and quantify inflammation-associated alterations in synaptic plasticity are not well-established. In previous work, we demonstrated that bacterial lipopolysaccharide (LPS)-induced systemic inflammation alters the ability of hippocampal neurons to express synaptic plasticity, i.e., the long-term potentiation (LTP) of excitatory neurotransmission. Here, we tested whether synaptic plasticity induced by repetitive magnetic stimulation (rMS), a non-invasive brain stimulation technique used in clinical practice, is affected by LPS-induced inflammation. Specifically, we explored brain tissue cultures to learn more about the direct effects of LPS on neural tissue, and we tested for the plasticity-restoring effects of the anti-inflammatory cytokine interleukin 10 (IL10). As shown previously, 10 Hz repetitive magnetic stimulation (rMS) of organotypic entorhino-hippocampal tissue cultures induced a robust increase in excitatory neurotransmission onto CA1 pyramidal neurons. Furthermore, LPS-treated tissue cultures did not express rMS-induced synaptic plasticity. Live-cell microscopy in tissue cultures prepared from a novel transgenic reporter mouse line [C57BL/6-Tg(TNFa-eGFP)] confirms that ex vivo LPS administration triggers microglial tumor necrosis factor alpha (TNFα) expression, which is ameliorated in the presence of IL10. Consistent with this observation, IL10 hampers the LPS-induced increase in TNFα, IL6, IL1β, and IFNγ and restores the ability of neurons to express rMS-induced synaptic plasticity in the presence of LPS. These findings establish organotypic tissue cultures as a suitable model for studying inflammation-induced alterations in synaptic plasticity, thus providing a biological basis for the diagnostic use of transcranial magnetic stimulation in the context of brain inflammation.

Show abstract Systemic inflammation is associated with alterations in complex brain functions such as learning and memory. However, diagnostic approaches to functionally assess and quantify inflammation-associated alterations in synaptic plasticity are not well-established. In previous work, we demonstrated that bacterial lipopolysaccharide (LPS)-induced systemic inflammation alters the ability of hippocampal neurons to express synaptic plasticity, i.e., the long-term potentiation (LTP) of excitatory neurotransmission. Here, we tested whether synaptic plasticity induced by repetitive magnetic stimulation (rMS), a non-invasive brain stimulation technique used in clinical practice, is affected by LPS-induced inflammation. Specifically, we explored brain tissue cultures to learn more about the direct effects of LPS on neural tissue, and we tested for the plasticity-restoring effects of the anti-inflammatory cytokine interleukin 10 (IL10). As shown previously, 10 Hz repetitive magnetic stimulation (rMS) of organotypic entorhino-hippocampal tissue cultures induced a robust increase in excitatory neurotransmission onto CA1 pyramidal neurons. Furthermore, LPS-treated tissue cultures did not express rMS-induced synaptic plasticity. Live-cell microscopy in tissue cultures prepared from a novel transgenic reporter mouse line [C57BL/6-Tg(TNFa-eGFP)] confirms that ex vivo LPS administration triggers microglial tumor necrosis factor alpha (TNFα) expression, which is ameliorated in the presence of IL10. Consistent with this observation, IL10 hampers the LPS-induced increase in TNFα, IL6, IL1β, and IFNγ and restores the ability of neurons to express rMS-induced synaptic plasticity in the presence of LPS. These findings establish organotypic tissue cultures as a suitable model for studying inflammation-induced alterations in synaptic plasticity, thus providing a biological basis for the diagnostic use of transcranial magnetic stimulation in the context of brain inflammation.

Show abstract The dynamics of the adjustment of center of pressure (CoP) has been utilized to understand motor control in human pathologies characterized by impairments in postural balance. The control mechanisms that maintain balance can be investigated via the analysis of muscle recruitment using electromyography (EMG) signals. In this work, we combined these two techniques to investigate balance control during upright standing in transfemoral unilateral amputees wearing a prosthesis. The dynamics of the CoP adjustments and EMG-EMG coherence between four muscles of the trunk and lower limb of 5 unilateral transfemoral amputees and 5 age-matched able-bodied participants were quantified during 30 s of quiet standing using the entropic half-life (EnHL) method. Two visual conditions, eyes open and eyes closed, were tested. Overall, the group of amputees presented lower EnHL values (higher dynamics) in their CoP adjustments than controls, especially in their intact limb. The EnHL values of the EMG-EMG coherence time series in the amputee group were lower than the control group for almost all muscle pairs under both visual conditions. Different correlations between the EnHL values of the CoP data and the EMG-EMG coherence data were observed in the amputee and control groups. These preliminary results suggest the onset of distinct neuromuscular adaptations following a unilateral amputation.Clinical Relevance - Understanding neuromuscular adaptation mechanisms after an amputation may serve to design better rehabilitation treatments and novel prosthetic devices with sensory feedback.

Show abstract Towards a biocompatible HySiF, in this work a customized Chip-Film Patch system is presented, in which an ultra-thin silicon chip is embedded in spin-coated polyimide. In addition, packaged with ceramic encapsulation layers, using atomic layer deposition (ALD) method thus achieving a prolonged long-term performance of the sensors. We have thoroughly investigated ALD coated polyimide foils with varying sequential metal oxide layers. More specifically, Al 2 O 3 -TiO 2 thin-film layers are deposited on polyimide foils in the range from 20 nm -to 80 nm thickness. ALD thin-film is investigated with means of scanning electron microscopy and atomic force microscopy. Intrinsic stress of Al 2 O 3 -TiO 2 thin-films on polyimide foils has been investigated and the water vapor transmission rate (WVTR) of the compound layers have been determined to assess hermeticity of the ALD layers. The long-term electrical leakage performance of the CFP substrates is evaluated, in physiological environment under cyclic mechanical load. The results provide a basis for processing flexible Hybrid Systems-in-foil for long-term utilization.

Show abstract Robots coexisting with humans in their environment and performing services for them need the ability to interact with them. One particular requirement for such robots is that they are able to understand spatial relations and can place objects in accordance with the spatial relations expressed by their user. In this work, we present a convolutional neural network for estimating pixelwise object placement probabilities for a set of spatial relations from a single input image. During training, our network receives the learning signal by classifying hallucinated high-level scene representations as an auxiliary task. Unlike previous approaches, our method does not require ground truth data for the pixelwise relational probabilities or 3D models of the objects, which significantly expands the applicability in practical applications. Our results obtained using real-world data and human-robot experiments demonstrate the effectiveness of our method in reasoning about the best way to place objects to reproduce a spatial relation.

Show abstract Microfabricated neuroprosthetic devices have made possible important observations on neuron activity; however, long-term high-fidelity recording performance of these devices has yet to be realized. Tissue-device interactions appear to be a primary source of lost recording performance. The current state of the art for visualizing the tissue response surrounding brain implants in animals is Immunohistochemistry + Confocal Microscopy, which is mainly performed after sacrificing the animal. Monitoring the tissue response as it develops could reveal important features of the response which may inform improvements in electrode design. Optical Coherence Tomography (OCT), an imaging technique commonly used in ophthalmology, has already been adapted for imaging of brain tissue. Here, we use OCT to achieve real-time, in vivo monitoring of the tissue response surrounding chronically implanted neural devices. The employed tissue-response-provoking implants are coated with a plasma-deposited nanofilms, which have been demonstrated as a biocompatible and anti-inflammatory interface for indwelling devices. We evaluate the method by comparing the OCT results to traditional histology qualitatively and quantitatively. The differences in OCT signal across the implantation period between the plasma group and the control reveal that the Parylene-type coating of otherwise rigid brain probes (glass and silicon) does not improve the glial encapsulation in the brain parenchyma.

Show abstract Machine learning (ML) methods have the potential to automate clinical EEG analysis. They can be categorized into feature-based (with handcrafted features), and end-to-end approaches (with learned features). Previous studies on EEG pathology decoding have typically analyzed a limited number of features, decoders, or both. For a I) more elaborate feature-based EEG analysis, and II) in-depth comparisons of both approaches, here we first develop a comprehensive feature-based framework, and then compare this framework to state-of-the-art end-to-end methods. To this aim, we apply the proposed feature-based framework and deep neural networks including an EEG-optimized temporal convolutional network (TCN) to the task of pathological versus non-pathological EEG classification. For a robust comparison, we chose the Temple University Hospital (TUH) Abnormal EEG Corpus (v2.0.0), which contains approximately 3000 EEG recordings. The results demonstrate that the proposed feature-based decoding framework can achieve accuracies on the same level as state-of-the-art deep neural networks. We find accuracies across both approaches in an astonishingly narrow range from 81 to 86%. Moreover, visualizations and analyses indicated that both approaches used similar aspects of the data, e.g., delta and theta band power at temporal electrode locations. We argue that the accuracies of current binary EEG pathology decoders could saturate near 90% due to the imperfect inter-rater agreement of the clinical labels, and that such decoders are already clinically useful, such as in areas where clinical EEG experts are rare. We make the proposed feature-based framework available open source and thus offer a new tool for EEG machine learning research.

Show abstract BACKGROUND: Medial forebrain bundle (MFB) deep brain stimulation (DBS) has anti-depressant effects clinically and in depression models. Currently, therapeutic mechanisms of MFB DBS or how stimulation parameters acutely impact neurotransmitter release, particularly dopamine, are unknown. Experimentally, MFB DBS has been shown to evoke dopamine response in healthy controls, but not yet in a rodent model of depression. OBJECTIVE: The study investigated the impact of clinically used stimulation parameters on the dopamine induced response in a validated rodent depression model and in healthy controls. METHOD: The stimulation-induced dopamine response in Flinders Sensitive Line (FSL, n = 6) rat model of depression was compared with Sprague Dawley (SD, n = 6) rats following MFB DSB, using Fast Scan Cyclic Voltammetry to assess the induced response in the nucleus accumbens. Stimulation parameters were 130 Hz ("clinically" relevant) with pulse widths between 100 and 350 μs. RESULTS: Linear mixed model analysis showed significant impact in both models following MFB DBS both at 130 and 60 Hz with 100 μs pulse width in inducing dopamine response. Furthermore, at 130 Hz the evoked dopamine responses were different across the groups at the different pulse widths. CONCLUSION: The differential impact of MFB DBS on the induced dopamine response, including different response patterns at given pulse widths, is suggestive of physiological and anatomical divergence in the MFB in the pathological and healthy state. Studying how varying stimulation parameters affect the physiological outcome will promote a better understanding of the biological substrate of the disease and the possible anti-depressant mechanisms at play in clinical MFB DBS.

Show abstract BMT 2020 – Leipzig, September 29 – October 1 • DOI 10.1515/bmt-2020-6006 Biomed. Eng.-Biomed. Tech. 2020; 65(s1): S20–S24 • © by Walter de Gruyter • Berlin • Boston

Show abstract Real-time object detection in videos using lightweight hardware is a crucial component of many robotic tasks. Detectors using different modalities and with varying computational complexities offer different trade-offs. One option is to have a very lightweight model that can predict from all modalities at once for each frame. However, in some situations (e.g., in static scenes) it might be better to have a more complex but more accurate model and to extrapolate from previous predictions for the frames coming in at processing time. We formulate this task as a sequential decision making problem and use reinforcement learning (RL) to generate a policy that decides from the RGB input which detector out of a portfolio of different object detectors to take for the next prediction. The objective of the RL agent is to maximize the accuracy of the predictions per image. We evaluate the approach on the Waymo Open Dataset and show that it exceeds the performance of each single detector.

Show abstract Electric stimulators with precise and reliable outputs are an indispensable part of electrophysiological research. From single cells to deep brain or neuromuscular tissue, there are diverse targets for electrical stimulation. Even though commercial systems are available, we state the need for a low-cost, high precision, functional, and modular (hardware, firmware, and software) current stimulation system with the capacity to generate stable and complex waveforms for pre-clinical research. The system presented in this study is a USB controlled 4-channel modular current stimulator that can be expanded and generate biphasic arbitrary waveforms with 16-bit resolution, high temporal precision (μs), and passive charge balancing: the NES STiM (Neuro Electronic Systems Stimulator). We present a detailed description of the system’s structural design, the controlling software, reliability test, and the pre-clinical studies [deep brain stimulation (DBS) in hemi-PD rat model] in which it was utilized. The NES STiM has been tested with MacOS and Windows operating systems. Interfaces to MATLAB source codes are provided. The system is inexpensive, relatively easy to build and can be assembled quickly. We hope that the NES STiM will be used in a wide variety of neurological applications such as Functional Electrical Stimulation (FES), DBS and closed loop neurophysiological research.

Show abstract Recently, cortical correlates of specific dream contents have been reported, such as the activation of the sensorimotor cortex during dreamed hand clenching. Yet, despite a close resemblance of such activation patterns to those seen during the corresponding wakeful behaviour, the causal mechanisms underlying specific dream contents remain largely elusive. Here, we aimed to investigate the causal role of the sensorimotor cortex in generating movement and bodily sensations during REM sleep dreaming. Following bihemispheric transcranial direct current stimulation (tDCS) or sham stimulation, guided by functional mapping of the primary motor cortex, naive participants were awakened from REM sleep and responded to a questionnaire on bodily sensations in dreams. Electromyographic (EMG) and electroencephalographic (EEG) recordings were used to quantify physiological changes during the preceding REM period. We found that tDCS, compared to sham stimulation, significantly decreased reports of dream movement, especially of repetitive actions. Other types of bodily experiences, such as tactile or vestibular sensations, were not affected by tDCS, confirming the specificity of stimulation effects to movement sensations. In addition, tDCS reduced EEG interhemispheric coherence in parietal areas and affected the phasic EMG correlation between both arms. These findings show that a complex temporal reorganization of the motor network co-occurred with the reduction of dream movement, revealing a link between central and peripheral motor processes and movement sensations of the dream self. tDCS over the sensorimotor cortex interferes with dream movement during REM sleep, which is consistent with a causal contribution to dream experience and has broader implications for understanding the neural basis of self-experience in dreams.

Show abstract Humans rely on their sense of touch to interact with the environment. Thus, restoring lost tactile sensory capabilities in amputees would advance their quality of life. In particular, texture discrimination is an important component for the interaction with the environment, but its restoration in amputees has been so far limited to simplified gratings. Here we show that naturalistic textures can be discriminated by trans-radial amputees using intraneural peripheral stimulation and tactile sensors located close to the outer layer of the artificial skin. These sensors exploit the morphological neural computation (MNC) approach, i.e., the embodiment of neural computational functions into the physical structure of the device, encoding normal and shear stress to guarantee a faithful neural temporal representation of stimulus spatial structure. Two trans-radial amputees successfully discriminated naturalistic textures via the MNC-based tactile feedback. The results also allowed to shed light on the relevance of spike temporal encoding in the mechanisms used to discriminate naturalistic textures. Our findings pave the way to the development of more natural bionic limbs.

Show abstract Cochlear implants provide the brain with auditory information as a treatment for hearing impairment, but electrical implants suffer from channel cross-talk, which can limit the quality of hearing. Keppeler et al. developed multichannel LED cochlear implants that use optogenetics to stimulate spiral ganglion neurons. Testing in rats and gerbils demonstrated improved spectral selectivity and hearing restoration. This study provides the groundwork for further design optimization and scaling of LED-based optical cochlear implants toward clinical translation.When hearing fails, electrical cochlear implants (eCIs) provide the brain with auditory information. One important bottleneck of CIs is the poor spectral selectivity that results from the wide current spread from each of the electrode contacts. Optical CIs (oCIs) promise to make better use of the tonotopic order of spiral ganglion neurons (SGNs) inside the cochlea by spatially confined stimulation. Here, we established multichannel oCIs based on light-emitting diode (LED) arrays and used them for optical stimulation of channelrhodopsin (ChR)-expressing SGNs in rodents. Power-efficient blue LED chips were integrated onto microfabricated 15-μm-thin polyimide-based carriers comprising interconnecting lines to address individual LEDs by a stationary or mobile driver circuitry. We extensively characterized the optoelectronic, thermal, and mechanical properties of the oCIs and demonstrated stability over weeks in vitro. We then implanted the oCIs into ChR-expressing rats and gerbils, and characterized multichannel optogenetic SGN stimulation by electrophysiological and behavioral experiments. Improved spectral selectivity was directly demonstrated by recordings from the auditory midbrain. Long-term experiments in deafened ChR-expressing rats and in nontreated control animals demonstrated specificity of optogenetic stimulation. Behavioral studies on animals carrying a wireless oCI sound processor revealed auditory percepts. This study demonstrates hearing restoration with improved spectral selectivity by an LED-based multichannel oCI system.

Show abstract Bioelectronic devices, interfacing neural tissue for therapeutic, diagnostic, or rehabilitation purposes, rely on small electrode contacts in order to achieve highly sophisticated communication at the neural interface. Reliable recording and safe stimulation with small electrodes, however, are limited when conventional electrode metallizations are used, demanding the development of new materials to enable future progress within bioelectronics. In this study, we present a versatile process for the realization of nanostructured platinum (nanoPt) coatings with a high electrochemically active surface area, showing promising biocompatibility and providing low impedance, high charge injection capacity, and outstanding long-term stability both for recording and stimulation. The proposed electrochemical fabrication process offers exceptional control over the nanoPt deposition, allowing the realization of specific coating morphologies such as small grains, pyramids, or nanoflakes, and can moreover be scaled up to wafer level or batch fabrication under economic process conditions. The suitability of nanoPt as a coating for neural interfaces is here demonstrated, in vitro and in vivo, revealing superior stimulation performance under chronic conditions. Thus, nanoPt offers promising qualities as an advanced neural interface coating which moreover extends to the numerous application fields where a large (electro)chemically active surface area contributes to increased efficiency.

Show abstract Bioelectronic devices, interfacing neural tissue for therapeutic, diagnostic, or rehabilitation purposes, rely on small electrode contacts in order to achieve highly sophisticated communication at the neural interface. Reliable recording and safe stimulation with small electrodes, however, are limited when conventional electrode metallizations are used, demanding the development of new materials to enable future progress within bioelectronics. In this study, we present a versatile process for the realization of nanostructured platinum (nanoPt) coatings with a high electrochemically active surface area, showing promising biocompatibility and providing low impedance, high charge injection capacity, and outstanding long-term stability both for recording and stimulation. The proposed electrochemical fabrication process offers exceptional control over the nanoPt deposition, allowing the realization of specific coating morphologies such as small grains, pyramids, or nanoflakes, and can moreover be scaled up to wafer level or batch fabrication under economic process conditions. The suitability of nanoPt as a coating for neural interfaces is here demonstrated, in vitro and in vivo, revealing superior stimulation performance under chronic conditions. Thus, nanoPt offers promising qualities as an advanced neural interface coating which moreover extends to the numerous application fields where a large (electro)chemically active surface area contributes to increased efficiency.

Show abstract Objective: Chronic neural implants require energy and signal supply. The objective of this work was to evaluate a multichannel transcutaneous coupling approach in an ex vivo split-concept study, which minimizes the invasiveness of such an implant by externalizing the processing electronics. Methods: Herein, the experimental work focused on the transcutaneous energy and signal transmission. The performance was discussed with widely evaluated concepts of neural interfaces in the literature. Results: The performance of the transcutaneous coupling approach increased with higher channel count and higher electrode pitches. Electrical crosstalk among channels was present, but acceptable for the stimulation of peripheral nerves. Conclusions: Transcutaneous coupling with extracorporeal transmitting arrays and subcutaneous counterparts provide a promising alternative to the inductive concept particularly when a fully integration of the system in a prosthetic shaft is intended. The relocation of the electronics can potentially prevent pressure sores, improve accessibility for maintenance and increase lifetime of the implant.

Show abstract Hermetic packaging is a suitable means to achieve long lifetime for implantable medical devices. In order to determine the time to failure of the hermetic packages, the prediction of lifetime by helium leakage measurements prior to implantation is state of the art. However, these methods are not applicable to packages with very small internal cavity volumes, as these established methods reach their resolution limits in this case. In these packages, even the smallest amounts of water can lead to corrosion or short circuits and thus to a malfunctioning system with unexpected behavior. Online monitoring of humidity offers the advantage that critical conditions are detected before a failure occurs. A typical maximum limit is 5000 ppm water vapor inside the package, which corresponds to a relative humidity of around 8.1 %RH at 37 °C. This paper presents the design and implementation of capacitive micromachined humidity detectors with a desiccant-based dielectric in different designs. These allow the electrical detection of moisture absorption, especially in the very low humidity regime ≤ 8 %RH. For this humidity range design-dependent sensitivities of 0.08 pF/%RH and 0.25 pF/%RH were achieved. Presented sensor devices facilitate active monitoring of current moisture content in the package cavity and at the same time provide a buffer for ingressing moisture, thus increasing the overall lifetime of the system. The current sensors are designed for miniaturized active medical implants, but can also be transferred to distributed sensors in harsh environments for Internet of Things applications or machine condition monitoring in Industry 4.0 scenarios.

Show abstract We present a novel, to the best of our knowledge, fabrication process for highly aspherical lenses based on surface deformation due to thermal expansion of a soft polymer, polydimethylsiloxane (PDMS), using laser-structuring, molding, and precise shape optimization. Our fabrication process can be used for almost any lens shape with a large degree of freedom—both individual lenses and dense arrays. We present the design, fabrication, and characterization with examples of four different lenses with 1 mm apertures and surface deviations below 100 nm.

Show abstract Prosthetic technology for people with missing limbs has made great progress in recent decades. However, acceptance rates and user satisfaction are not only dependent on technical aspects, but also to a great extent on social and psychological factors. We propose that these factors should receive greater attention in order to improve prosthetic care and give recommendations how to incorporate the findings from social science in research and development (R&amp;D) and in care practice. Limited access due to high costs of new prosthetic technology combined with rising costs in health care systems in general is a further issue we address. Our legal and ethical analysis of the reimbursement process in Germany shows that this issue requires further empirical investigation, a stakeholder dialogue and maybe even policy changes. Social science knowledge and participatory methods are of high relevance to answer questions about the benefit of prosthetics for users, based on individual needs and preferences, which should be at the core of debates on ethical resource allocation.

Show abstract Emerging technological developments in nano- and microsystems engineering have delivered powerful tools for neuroscience research over the last 50 years. However, only a few neural implants have been transferred into clinical practice. Challenges and opportunities for translational research are discussed herein.

Show abstract Diffusion-weighted magnetic resonance imaging (dMRI) allows non-invasive investigation of whole-brain connectivity, which can reveal the brain’s global network architecture and also abnormalities involved in neurological and mental disorders. However, the reliability of connection inferences from dMRI-based fiber tracking is still debated, due to low sensitivity, dominance of false positives, and inaccurate and incomplete reconstruction of long-range connections. Furthermore, parameters of tracking algorithms are typically tuned in a heuristic way, which leaves room for manipulation of an intended result. Here we propose a general data-driven framework to optimize and validate parameters of dMRI-based fiber tracking algorithms using neural tracer data as a reference. Japan’s Brain/MINDS Project provides invaluable datasets containing both dMRI and neural tracer data from the same primates. A fundamental difference when comparing dMRI-based tractography and neural tracer data is that the former cannot specify the direction of connectivity; therefore, evaluating the fitting of dMRI-based tractography becomes challenging. The framework implements multi-objective optimization based on the non-dominated sorting genetic algorithm II. Its performance is examined in two experiments using data from ten subjects for optimization and six for testing generalization. The first uses a seed-based tracking algorithm, iFOD2, and objectives for sensitivity and specificity of region-level connectivity. The second uses a global tracking algorithm and a more refined set of objectives: distance-weighted coverage, true/false positive ratio, projection coincidence, and commissural passage. In both experiments, with optimized parameters compared to default parameters, fiber tracking performance was significantly improved in coverage and fiber length. Improvements were more prominent using global tracking with refined objectives, achieving an average fiber length from 10 to 17 mm, voxel-wise coverage of axonal tracts from 0.9 to 15%, and the correlation of target areas from 40 to 68%, while minimizing false positives and impossible cross-hemisphere connections. Optimized parameters showed good generalization capability for test brain samples in both experiments, demonstrating the flexible applicability of our framework to different tracking algorithms and objectives. These results indicate the importance of data-driven adjustment of fiber tracking algorithms and support the validity of dMRI-based tractography, if appropriate adjustments are employed.

Show abstract The study design of PIMIDES, a trial based on patient-individualized transcranial electric neurostimulation of epileptic foci, is reported. Inclusion criteria include a predominant epileptic focus and pharmacoresistance to two antiepileptic drug treatments. The study is prospective, unblinded, and serves to assess the safety of subgaleal implantation and transcranial stimulation.

Show abstract Application-specific integrated circuits (ASICs) embedded in polymers have been subject in implant manufacturing for the recent years. The increased functionality combined with good biocompatibility due to flexibility of thin implants makes them interesting for further studies. Thin-film ASICs can be used for the recording and processing of a high amount of biological signals, improving the performance of neural implants. Fabrication and analysis of gold and platinum thin-film connections are subject of this study, especially their capability as high frequency data transmission lines. Three layers of polyimide are used as flexible substrate and insulator of the traces. Various test structures were designed and fabricated, to investigate the resistance and reactance up to GHz frequencies, crosstalk and influence of vias between metallization layers. All conducting structures have a comparable design with a length of 50 mm and a metal thickness of 300 nm, while the line widths were varied. In this configuration gold and platinum thinfilm conductors are both suitable for high-frequency data transmission up to 100 MHz. This transmission frequency limit and impedances are unaffected by a wet environment and in accelerated aging tests. However, both metals show a high pass filter behavior, whose frequency behavior is mostly dependent by the self-inductance and resistance. A simplified ideal transmission model predicts the electrical behavior sufficiently and can be used to design the favored line impedance matching input impedances of the connected ASICs.

Show abstract Horizontally scanning 2D laser rangefinders are a popular approach for indoor robot localization because of the high accuracy of the sensors and the compactness of the required 2D maps. As the scanners in this configuration only provide information about one slice of the environment, the measurements typically do not capture the full extent of a large variety of obstacles, including chairs or tables. Accordingly, obstacle avoidance based on laser scanners mounted in such a fashion is likely to fail. In this paper, we propose a learning-based approach to predict collisions in 2D occupancy maps. Our approach is based on a convolutional neural network which is trained on a 2D occupancy map and collision events recorded with a bumper while the robot is navigating in its environment. As the network operates on local structures only, it can generalize to new environments. In addition, the robot can collect and integrate new collision examples after an initial training phase. Extensive experiments carried out in simulation and a realistic real-world environment confirm that our approach allows robots to learn from collision events to avoid collisions in the future.

Show abstract Information processing in cortical neuronal networks relies on properly balanced excitatory and inhibitory neurotransmission. A ubiquitous motif for maintaining this balance is the somatostatin interneuron (SOM-IN) feedback microcircuit. Here, we investigated the modulation of this microcircuit by presynaptic GABAB receptors (GABABRs) in the rodent hippocampus. Whole-cell recordings from SOM-INs revealed that both excitatory and inhibitory synaptic inputs are strongly inhibited by GABABRs, while optogenetic activation of the interneurons shows that their inhibitory output is also strongly suppressed. Electron microscopic analysis of immunogold-labelled freeze-fracture replicas confirms that GABABRs are highly expressed presynaptically at both input and output synapses of SOM-INs. Activation of GABABRs selectively suppresses the recruitment of SOM-INs during gamma oscillations induced in vitro. Thus, axonal GABABRs are positioned to efficiently control the input and output synapses of SOM-INs and can functionally uncouple them from local network with implications for rhythmogenesis and the balance of entorhinal versus intrahippocampal afferents.

Show abstract Neural oscillations as important information carrier in the brain, are increasingly interpreted as transient bursts rather than as sustained oscillations. Short (<150 ms) bursts of beta-waves (15-30 Hz) have been documented in humans, monkeys and mice. These events were correlated with memory, movement and perception, and were even suggested as the primary ingredient of all beta-band activity. However, a method to measure these short-lived events in real-time and to investigate their impact on behaviour is missing. Here we present a real-time data analysis system, capable to detect short narrowband bursts, and demonstrate its usefulness to increase the beta-band burst-rate in rats. This neurofeedback training induced changes in overall oscillatory power, and bursts could be decoded from the movement of the rats, thus enabling future investigation of the role of oscillatory bursts.

Show abstract {Resting-state functional MRI has shown potential for detecting changes in cerebral blood oxygen level-dependent signal in patients with epilepsy, even in the absence of epileptiform activity. Furthermore, it has been suggested that coefficient of variation mapping of fast functional MRI signal may provide a powerful tool for the identification of intrinsic brain pulsations in neurological diseases such as dementia, stroke and epilepsy. In this study, we used fast functional MRI sequence (magnetic resonance encephalography) to acquire ten whole-brain images per second. We used the functional MRI data to compare physiological brain pulsations between healthy controls (n = 102) and patients with epilepsy (n = 33) and furthermore to drug-naive seizure patients (n = 9). Analyses were performed by calculating coefficient of variation and spectral power in full band and filtered sub-bands. Brain pulsations in the respiratory-related frequency sub-band (0.11–0.51 Hz) were significantly (P \\&lt; 0.05) increased in patients with epilepsy, with an increase in both signal variance and power. At the individual level, over 80\\% of medicated and drug-naive seizure patients exhibited areas of abnormal brain signal power that correlated well with the known clinical diagnosis, while none of the controls showed signs of abnormality with the same threshold. The differences were most apparent in the basal brain structures, respiratory centres of brain stem, midbrain and temporal lobes. Notably, full-band, very low frequency (0.01–0.1 Hz) and cardiovascular (0.8–1.76 Hz) brain pulses showed no differences between groups. This study extends and confirms our previous results of abnormal fast functional MRI signal variance in epilepsy patients. Only respiratory-related brain pulsations were clearly increased with no changes in either physiological cardiorespiratory rates or head motion between the subjects. The regional alterations in brain pulsations suggest that mechanisms driving the cerebrospinal fluid homeostasis may be altered in epilepsy. Magnetic resonance encephalography has both increased sensitivity and high specificity for detecting the increased brain pulsations, particularly in times when other tools for locating epileptogenic areas remain inconclusive.}

Show abstract Self-supervised learning is a promising new technique for learning representative features in the absence of manual annotations. It is particularly efficient in cases where labeling the training data is expensive and tedious, naturally linking it to the semi-supervised learning paradigm. In this work, we propose a new semi-supervised time series classification model that leverages features learned from the self-supervised task on unlabeled data. The idea is to exploit the unlabeled training data with a forecasting task which provides a strong surrogate supervision signal for feature learning. We draw from established multi-task learning approaches and model forecasting as an auxiliary task to be optimized jointly with the main task of classification. We evaluate our proposed method on benchmark time series classification datasets in semi-supervised setting and are able to show that it significantly outperforms the state-of-the-art baselines.

Show abstract Recent studies have shown that neural stimulation can be used to provide artificial sensory feedback to amputees eliciting sensations referred on the amputated hand. The temporal properties of the neural stimulation modulate aspects of evoked sensations that can be exploited in a bidirectional hand prosthesis.

Show abstract is dedicated to building social movements and political struggles around progressive and radical perspectives on science and society. We are workers, educators, and students in STEM and related fields committed to the democratic practice of science for the benefit of humanity and the planet.

Show abstract Abstract Noninvasive wearable devices have great potential to aid the management of epilepsy, but these devices must have robust signal quality, and patients must be willing to wear them for long periods of time. Automated machine learning classification of wearable biosensor signals requires quantitative measures of signal quality to automatically reject poor-quality or corrupt data segments. In this study, commercially available wearable sensors were placed on patients with epilepsy undergoing in-hospital or in-home electroencephalographic (EEG) monitoring, and healthy volunteers. Empatica E4 and Biovotion Everion were used to record accelerometry (ACC), photoplethysmography (PPG), and electrodermal activity (EDA). Byteflies Sensor Dots were used to record ACC and PPG, the Activinsights GENEActiv watch to record ACC, and Epitel Epilog to record EEG data. PPG and EDA signals were recorded for multiple days, then epochs of high-quality, marginal-quality, or poor-quality data were visually identified by reviewers, and reviewer annotations were compared to automated signal quality measures. For ACC, the ratio of spectral power from 0.8 to 5 Hz to broadband power was used to separate good-quality signals from noise. For EDA, the rate of amplitude change and prevalence of sharp peaks significantly differentiated between good-quality data and noise. Spectral entropy was used to assess PPG and showed significant differences between good-, marginal-, and poor-quality signals. EEG data were evaluated using methods to identify a spectral noise cutoff frequency. Patients were asked to rate the usability and comfort of each device in several categories. Patients showed a significant preference for the wrist-worn devices, and the Empatica E4 device was preferred most often. Current wearable devices can provide high-quality data and are acceptable for routine use, but continued development is needed to improve data quality, consistency, and management, as well as acceptability to patients.

Show abstract Spiking activity in cortical networks is nonlinear in nature. The linear-nonlinear cascade model, some versions of which are also known as point-process generalized linear model, can efficiently capture the nonlinear dynamics exhibited by such networks. Of particular interest in such models are theoretical predictions of spike train statistics. However, due to the moment-closure problem, approximations are inevitable. We suggest here a series expansion that explains how higher-order moments couple to lower-order ones. Our approach makes predictions in terms of certain integrals, the so-called loop integrals. In previous studies these integrals have been evaluated numerically, but numerical instabilities are sometimes encountered rendering the results unreliable. Analytic solutions are presented here to overcome this problem, and to arrive at more robust evaluations. We were able to deduce these analytic solutions by switching to Fourier space and making use of complex analysis, specifically Cauchy's residue theorem. We formalized the loop integrals and explicitly solved them for specific response functions. To quantify the importance of these corrections for spike train cumulants, we numerically simulated spiking networks and compared their sample statistics to our theoretical predictions. Our results demonstrate that the magnitude of the nonlinear corrections depends on the working point of the nonlinear network dynamics, and that it is related to the eigenvalues of the mean-field stability matrix. For our example, the corrections for the firing rates are in the range between 4 % and 21 % on average. Precise and robust predictions of spike train statistics accounting for nonlinear effects are, for example, highly relevant for theories involving spike-timing dependent plasticity (STDP).

Show abstract The functional role of spontaneous brain activity, especially in relation to external events, is a longstanding key question in neuroscience. Intrinsic and externally-evoked activities were suggested to be anticorrelated, yet inferring an antagonistic mechanism between them remains a challenge. Here, we used beta-band (15-30 Hz) power as a proxy of spontaneous activity in the rat somatosensory cortex during a detection task. Beta-power anticorrelated with sensory-evoked-responses, and high rates of spontaneously occurring beta-bursts predicted reduced detection. By applying a burst-rate detection algorithm in real-time and trial-by-trial stimulus-intensity adjustment, this influence could be counterbalanced. Mechanistically, bursts in all bands indicated transient synchronization of cell assemblies, but only beta-bursts were followed by a reduction in firing-rate. Our findings reveal that spontaneous beta-bursts reflect a dynamic state that competes with external stimuli.Competing Interest StatementThe authors have declared no competing interest.

Show abstract Objective. Micro-fabricated neural interfaces based on polyimide (PI) are achieving increasing importance in translational research. The ability to produce well-defined micro-structures with properties that include chemical inertness, mechanical flexibility and low water uptake are key advantages for these devices. Approach. This paper reports the development of the transverse intrafascicular multichannel electrode (TIME) used to deliver intraneural sensory feedback to an upper-limb amputee in combination with a sensorized hand prosthesis. A failure mode analysis on the explanted devices was performed after a first-in-human study limited to 30 d. Main results. About 90% of the stimulation contact sites of the TIMEs maintained electrical functionality and stability during the full implant period. However, optical analysis post-explantation revealed that 62.5% of the stimulation contacts showed signs of delamination at the metallization-PI interface. Such damage likely occurred due to handling during explantation and subsequent analysis, since a significant change in impedance was not observed in vivo. Nevertheless, whereas device integrity is mandatory for long-term functionality in chronic implantation, measures to increase the bonding strength of the metallization-PI interface deserve further investigation. We report here that silicon carbide (SiC) is an effective adhesion-promoting layer resisting heavy electrical stimulation conditions within a rodent animal trial. Optical analysis of the new electrodes revealed that the metallization remained unaltered after delivering over 14 million pulses in vivo without signs of delamination at the metallization-PI interface. Significance. Failure mode analysis guided implant stability optimization. Reliable adhesion of thin-film metallization to substrate has been proven using SiC, improving the potential transfer of micro-fabricated neural electrodes for chronic clinical applications. (Document number of Ethical Committee: P/905/CE/2012; Date of approval: 2012–10-04)

Show abstract {High-frequency oscillations are markers of epileptic tissue. Recently, different patterns of EEG background activity were described from which high-frequency oscillations occur: high-frequency oscillations with continuously oscillating background were found to be primarily physiological, those from quiet background were linked to epileptic tissue. It is unclear, whether these interactions remain stable over several days and during different sleep-wake stages. High-frequency oscillation patterns (oscillatory vs. quiet background) were analysed in 23 patients implanted with depth and subdural grid electrodes. Pattern scoring was performed on every channel in 10 s intervals in three separate day- and night-time EEG segments. An entropy value, measuring variability of patterns per channel, was calculated. A low entropy value indicated a stable occurrence of the same pattern in one channel, whereas a high value indicated pattern instability. Differences in pattern distribution and entropy were analysed for 143 280 10 s intervals with allocated patterns from inside and outside the seizure onset zone, different electrode types and brain regions. We found a strong association between high-frequency oscillations out of quiet background activity, and channels of the seizure onset zone (35.2\\% inside versus 9.7\\% outside the seizure onset zone, P \\&lt; 0.001), no association was found for high-frequency oscillations from continuous oscillatory background (P = 0.563). The type of background activity remained stable over the same brain region over several days and was independent of sleep stage and recording technique. Stability of background activity was significantly higher in channels of the seizure onset zone (entropy mean value 0.56 ± 0.39 versus 0.64 ± 0.41; P \\&lt; 0.001). This was especially true for the presumed epileptic high-frequency oscillations out of quiet background (0.57 ± 0.39 inside versus 0.72 ± 0.37 outside the seizure onset zone; P \\&lt; 0.001). In contrast, presumed physiological high-frequency oscillations from continuous oscillatory backgrounds were significantly more stable outside the seizure onset zone (0.72 ± 0.45 versus 0.48 ± 0.53; P \\&lt; 0.001). The overall low entropy values suggest that interactions between high-frequency oscillations and background activity are a stable phenomenon specific to the function of brain regions. High-frequency oscillations occurring from a quiet background are strongly linked to the seizure onset zone whereas high-frequency oscillations from an oscillatory background are not. Pattern stability suggests distinct underlying mechanisms. Analysing short time segments of high-frequency oscillations and background activity could help distinguishing epileptic from physiologically active brain regions.}

Show abstract The fabrication of stretchable conductive material through vapor phase polymerization of poly(3,4-ethylenedioxythiophene) (PEDOT) is presented alongside a method to easily pattern these materials with nanosecond laser structuring. The devices were constructed from sheets of vapor phase polymerized PEDOT doped with tosylate on pre-stretched elastomeric substrates followed by laser structuring to achieve the desired geometrical shape. Devices were characterized for electrical conductivity, morphology, and electrical integrity in response to externally applied strain. Fabricated PEDOT sheets displayed a conductivity of 53.1 &plusmn; 1.2 S cm&minus;1; clear buckling in the PEDOT microstructure was observed as a result of pre-stretching the underlying elastomeric substrate; and the final stretchable electronic devices were able to remain electrically conductive with up to 100% of externally applied strain. The described polymerization and fabrication steps achieve highly processable and patternable functional conductive polymer films, which are suitable for stretchable electronics due to their ability to withstand externally applied strains of up to 100%.

Show abstract Electrical stimulation of the subthalamic nucleus (STN) is clinically employed to ameliorate several symptoms of manifest Parkinson’s Disease (PD). Stimulation parameters utilized by chronically implanted pulse generators comprise biphasic rectangular short (60–100 μs) pulses with a repetition frequency between 130 and 180 Hz. A better insight into the effect of electrical stimulation parameters could potentially reveal new possibilities for the improvement of deep brain stimulation (DBS) as a treatment. To this end, we employed single-sided 6-hydroxidopamine (6-OHDA) lesioning of the medial forebrain bundle (MFB) in rats to systematically investigate alternative stimulation parameters. These hemi-parkinsonian (hemi-PD) rats underwent individualized, ipsilateral electrical stimulation to the STN of the lesioned hemisphere, while the transiently induced contralateral rotational behavior was quantified to assess the effect of DBS parameter variations. The number of induced rotations during 30 s of stimulation was strongly correlated with the amplitude of the stimulation pulses. Despite a general linear relation between DBS frequency and rotational characteristics, a plateau effect was observed in the rotation count throughout the clinically used frequency range. Alternative waveforms to the conventional biphasic rectangular (Rect) pulse shapes [Triangular (Tri), Sinusoidal (Sine), and Sawtooth (Lin.Dec.)] required higher charges per phase to display similar behavior in rats as compared to the conventional pulse shape. The Euclidean Distance (ED) was used to quantify similarities between different angular trajectories. Overall, our study confirmed that the effect of different amplitude and frequency parameters of STN-DBS in the hemi-PD rat model was similar to those in human PD patients. This shows that induced contralateral rotation is a valuable readout in testing stimulation parameters. Our study supports the call for more pre-clinical studies using this measurement to assess the effect of other DBS parameters such as pulse-width and interphase intervals.

Show abstract Apraxia is frequently described after left hemisphere stroke and results from lesions to a complex network for motor cognition with dorso-dorsal, ventro-dorsal and ventral processing streams. Apraxia also occurs after right hemisphere stroke, but lesion correlates and underlying mechanisms remain to be elucidated. To clarify the role of the right hemisphere in apraxic deficits and the influence of neglect, we prospectively examined apraxia (imitation of meaningless postures and pantomime of tool use) and neglect in 138 acute right hemisphere stroke patients with first-ever ischemic stroke in the middle cerebral artery territory and identified corresponding lesion correlates using voxel-based lesion-symptom mapping. Imitation of meaningless postures was impaired as frequently as after left hemisphere stroke (38.4%) and was significantly associated with neglect. Imitation of meaningless postures was related to temporal (middle temporal gyrus, temporoparietal junction, superior temporal gyrus and sulcus), parietal (angular gyrus, parieto-occpitpial sulcus), secondary sensorimotor cortex and (peri-)insular lesions. Presence of neglect dichotomized the results: a lesion correlate for isolated imitation without neglect was found in the right parieto-occipital cortex, while imitation deficits, when co-occurring with neglect, were related to lateral occipito-temporal, superior temporal sulcus and (peri-)insular lesions. Pantomime of tool use deficits, typical for apraxia after left hemisphere lesions, were found in only 5 cases (3.6%) and only in the context of neglect, and were associated with occipital lobe, ventral and anterior temporal lobe, and inferior frontal (areas 45/47) lesions. The syndrome of apraxia after right hemisphere stroke differs from apraxia after left hemisphere stroke. Imitation deficits are found in both hemispheres after dorso-dorsal stream lesions. Neglect also leads to and explains deficits in imitation and pantomime in patients with right ventral stream lesions. Therefore, in right hemisphere lesions, apraxia can either be explained as impaired visuomotor transformation or as a result of visuospatial deficits.

Show abstract {Inferring knowledge is a core aspect of human cognition. We can form complex sentences connecting different pieces of information, such as in conditional statements like “if someone drinks alcohol, then they must be older than 18.” These are relevant for causal reasoning about our environment and allow us to think about hypothetical scenarios. Another central aspect to forming complex statements is to quantify about sets, such as in “some apples are green.” Reasoning in terms of the ability to form these statements is not yet fully understood, despite being an active field of interdisciplinary research. On a theoretical level, several conceptual frameworks have been proposed, predicting diverging brain activation patterns during the reasoning process. We present a meta-analysis comprising the results of 32 neuroimaging experiments about reasoning, which we subdivided by their structure, content, and requirement for world knowledge. In conditional tasks, we identified activation in the left middle and rostrolateral pFC and parietal regions, whereas syllogistic tasks elicit activation in Broca's complex, including the BG. Concerning the content differentiation, abstract tasks exhibit activation in the left inferior and rostrolateral pFC and inferior parietal regions, whereas content tasks are in the left superior pFC and parieto-occipital regions. The findings clarify the neurocognitive mechanisms of reasoning and exhibit clear distinctions between the task's type and content. Overall, we found that the activation differences clarify inconsistent results from accumulated data and serve as useful scaffolding differentiations for theory-driven interpretations of the neuroscientific correlates of human reasoning.}

Show abstract Biomarkers sensitive to prodromal or early pathophysiological changes in Alzheimer's disease (AD) symptoms could improve disease detection and enable timely interventions. Changes in brain hemodynamics may be associated with the main clinical AD symptoms. To test this possibility, we measured the variability of blood oxygen level-dependent (BOLD) signal in individuals from three independent datasets (totaling 80 AD patients and 90 controls). We detected a replicable increase in brain BOLD signal variability in the AD populations, which constituted a robust biomarker for clearly differentiating AD cases from controls. Fast BOLD scans showed that the elevated BOLD signal variability in AD arises mainly from cardiovascular brain pulsations. Manifesting in abnormal cerebral perfusion and cerebrospinal fluid convection, present observation presents a mechanism explaining earlier observations of impaired glymphatic clearance associated with AD in humans.

Show abstract Blind men and the elephant-comment on "The dentato-rubro-thalamic tract as the potential common deep brain stimulation target for tremor of various origin: an observational case series"

Show abstract Hermetic and non-hermetic packages of active implantable medical devices are often fabricated of ceramics like alumina. Screen printed PtAu paste is the state of the art metallization for functional structures. Due to solid state and liquid diffusion of Au at thermal exposure, solder times are limited; otherwise metal structures tend to delaminate. Moreover, it was shown that PtAu with solder fails after 37.4 years. We established a thin film metallization on alumina process to overcome these disadvantages. The metallization consists of sputtered platinum with an underlying adhesion layer made of tungsten-titanium to increase the adhesion strength to the alumina substrate. We avoided using gold in this work due to its high diffusion tendency. Instead, the materials in use provide relatively low diffusion properties, which potentially increases the long term mechanical performance and usability during assembly and packaging.Utilizing the Design of Experiment (DoE) methodology, we derived an optimal Pt thickness of 500 nm with 43 nm of WTi as adhesion promoting layer. After accelerated aging at 150 {\textdegree}C, corresponding to 125 years at body temperature (37 {\textdegree}C), the contact pad adhesion strength was with 32.75 MPa {\textpm} 7.08 MPa still significantly higher than the safety limit of 17 MPa, following the recommendations for a robust screen-printing metallization process. Moreover, soldering times of up to 120 s did not influence the adhesive strength. The new process reduced the minimum track distance to 50\% of screen printing values and is capable of rapid prototyping. It helps to make the assembly process independent of the manufacturing person, to increase the yield of device fabrication and -most important in implantable device manufacturing-to make it more robust and thereby more safe for the patient.Competing Interest StatementThe authors have declared no competing interest.

Show abstract Synaptic plasticity is the mechanistic basis of development, aging, learning and memory, both in the healthy and pathological brain. Structural plasticity is rarely accounted for in computational network models, due to a lack of insight into the underlying neuronal mechanisms and processes. Little is known about how the rewiring of networks is dynamically regulated. In our current study, we characterized the time course of neural activity, neural morphology, and the expression of synaptic proteins employing an in vivo optogenetic mouse model. We stimulated pyramidal neurons in the anterior cingulate cortex of mice and harvested their brains at 1.5 h, 24 h, and 48 h after stimulation. Stimulus-induced cortical hyperactivity persisted up to 1.5 h and decayed to baseline after 24 h, indicated by c-Fos expression. The synaptic proteins VGLUT1 and PSD-95, in contrast, were upregulated at 24 h and downregulated at 48 h, respectively. Spine density and spine head volume were also increased at 24 h and decreased at 48 h. This specific sequence of events reflects a continuous joint evolution of activity and connectivity that is typical of homeostatic structural plasticity. In this computational model, the turnover of dendritic spines and axonal boutons is regulated via firing rate homeostasis of individual neurons.Competing Interest StatementThe authors have declared no competing interest.

Show abstract Purpose To improve the reconstruction efficiency (i.e., computational load) and stability of iterative reconstruction for non-Cartesian fMRI when using high undersampling rates and/or in the presence of strong off-resonance effects. Theory and Methods The magnetic resonance encephalography (MREG) sequence with 3D non-Cartesian trajectory and 0.1s repetition time (TR) was applied to acquire fMRI datasets. Different from a conventional time-point-by-time-point sequential reconstruction (SR), the proposed time-domain principal component reconstruction (tPCR) performs three steps: (1) decomposing the k-t-space fMRI datasets into time-domain principal component space using singular value decomposition, (2) reconstructing each principal component with redistributed computation power according to their weights, and (3) combining the reconstructed principal components back to image-t-space. The comparison of reconstruction accuracy was performed by simulation experiments and then verified in real fMRI data. Results The simulation experiments showed that the proposed tPCR was able to significantly reduce reconstruction errors, and subsequent functional activation errors, relative to SR at identical computational cost. Alternatively, at fixed reconstruction accuracy, computation time was greatly reduced. The improved performance was particularly obvious for L1-norm nonlinear reconstructions relative to L2-norm linear reconstructions and robust to different regularization strength, undersampling rates, and off-resonance effects intensity. By examining activation maps, tPCR was also found to give similar improvements in real fMRI experiments. Conclusion The proposed proof-of-concept tPCR framework could improve (1) the reconstruction efficiency of iterative reconstruction, and (2) the reconstruction stability especially for nonlinear reconstructions. As a practical consideration, the improved reconstruction speed promotes the application of highly undersampled non-Cartesian fast fMRI.

Show abstract Paralysis is a frequent phenomenon in many diseases, and to date, only functional electrical stimulation (FES) mediated via the innervating nerve can be employed to restore skeletal muscle function in patients. Despite recent progress, FES has several technical limitations and significant side effects. Optogenetic stimulation has been proposed as an alternative, as it may circumvent some of the disadvantages of FES enabling cell type-specific, spatially and temporally precise stimulation of cells expressing light-gated ion channels, commonly Channelrhodopsin2. Two distinct approaches for the restoration of skeletal muscle function with optogenetics have been demonstrated: indirect optogenetic stimulation through the innervating nerve similar to FES and direct optogenetic stimulation of the skeletal muscle. Although both approaches show great promise, both have their limitations and there are several general hurdles that need to be overcome for their translation into clinics. These include successful gene transfer, sustained optogenetic protein expression, and the creation of optically active implantable devices. Herein, a comprehensive summary of the underlying mechanisms of electrical and optogenetic approaches is provided. With this knowledge in mind, we substantiate a detailed discussion of the advantages and limitations of each method. Furthermore, the obstacles in the way of clinical translation of optogenetic stimulation are discussed, and suggestions on how they could be overcome are provided. Finally, four specific examples of pathologies demanding novel therapeutic measures are discussed with a focus on the likelihood of direct versus indirect optogenetic stimulation.

Show abstract Background Major depression (MD) and obsessive-compulsive disorder (OCD) are psychiatric diseases with a huge impact on individual well-being. Despite optimal treatment regiments a subgroup of patients remains treatment resistant and stereotactic surgery (stereotactic lesion surgery, SLS or Deep Brain Stimulation, DBS) might be an option. Recent research has described four networks related to MD and OCD (affect, reward, cognitive control, default network) but only on a cortical and the adjacent sub-cortical level. Despite the enormous impact of comparative neuroanatomy, animal science and stereotactic approaches a holistic theory of subcortical and cortical network interactions is elusive. Because of the dominant hierarchical rank of the neocortex, corticofugal approaches have been used to identify connections in subcortical anatomy without anatomical priors and in part confusing results. We here propose a different corticopetal approach by identifying subcortical networks and search for neocortical convergences thereby following the principle of phylogenetic and ontogenetic network development. Material and methods This work used a diffusion tensor imaging data from a normative cohort (Human Connectome Project, HCP; n = 200) to describe eight subcortical fiber projection pathways (PPs) from subthalamic nucleus (STN), substantia nigra (SNR), red nucleus (RN), ventral tegmental area (VTA), ventrolateral thalamus (VLT) and mediodorsal thalamus (MDT) in a normative space (MNI). Subcortical and cortical convergences were described including an assignment of the specific pathways to MD/OCD-related networks. Volumes of activated tissue for different stereotactic stimulation sites and procedures were simulated to understand the role of the distinct networks, with respect to symptoms and treatment of OCD and MD. Results The detailed course of eight subcortical PPs (stnPP, snrPP, rnPP, vlATR, vlATRc, mdATR, mdATRc, vtaPP/slMFB) were described together with their subcortical and cortical convergences. The anterior limb of the internal capsule can be subdivided with respect to network occurrences in ventral-dorsal and medio-lateral gradients. Simulation of stereotactic procedures for OCD and MD showed dominant involvement of mdATR/mdATRc (affect network) and vtaPP/slMFB (reward network). Discussion Corticofugal search strategies for the evaluation of stereotactic approaches without anatomical priors often lead to confusing results which do not allow for a clear assignment of a procedure to an involved network. According to our simulation of stereotactic procedures in the treatment of OCD and MD, most of the target regions directly involve the reward (and affect) networks, while side-effects can in part be explained with a co-modulation of the control network. Conclusion The here proposed corticopetal approach of a hierarchical description of 8 subcortical PPs with subcortical and cortical convergences represents a new systematics of networks found in all different evolutionary and distinct parts of the human brain.

Show abstract Key points The primary motor cortex (M1) is fundamentally important for the acquisition of skilled motor behaviours. We tested the excitability changes of distinct M1 circuits at movement onset with TMS H-reflex conditioning. Human subjects trained a discrete spatiotemporal motor skill. Practice was associated with reduced kinematic variability and improved motor performance. Performance improvements were paralleled by task-specific excitability increases of the fastest corticospinal connections at infragranular layer 5b of M1. No task-related changes in excitability were observed at supragranular layers. Excitability changes in the fastest corticospinal connections were not directly related to changes in motor performance. Abstract The primary motor cortex (M1) is fundamentally important for the acquisition of skilled motor behaviours. Recent advances in imaging and electrophysiological techniques have improved our understanding of M1 neural circuit modulation in rodents and non-human primates during motor learning. However, little remains known about the learning-related changes of distinct elements in the human brain. In this study, we tested excitability changes of different neural circuits (infragranular and supragranular layers) in the M1 of human subjects who underwent training in a discrete spatiotemporal motor skill. Excitability modulations were assessed by recording H-reflex facilitation from transcranial magnetic stimulation at movement onset. Motor practice improved the consistency of movements and was accompanied by an excitability increase of the fastest corticospinal connections during the initial stages of motor practice. No such excitability changes were observed for training in a simple motor skill and circuits at supragranular layers of M1. Notably, changes in excitability were not associated with changes in motor performance. Our findings could reflect learning-related increases in the recruitment and/or reorganisation of the fastest corticospinal connections.

Show abstract Implantable neural interfaces advance the possibilities for neuroscientists to study the brain. They are also promising for use in a multitude of bioelectronic therapies. Electrode technology plays a central role in these developments, as the electrode surfaces form the physical interfaces between technology and the biological targets. Despite this, a common understanding of how electrodes should best be evaluated and compared with respect to their efficiency in recording and stimulation is currently lacking. Without broadly accepted performance tests, it is difficult to rank the many suggestions for electrode materials available in the literature, or to identify where efforts should be focused to advance the field most efficiently. This tutorial critically discusses the most relevant performance tests for characterization of neural interface electrodes and explains their implementation, interpretation and respective limitations. We propose a unified standard to facilitate transparent reporting on electrode performance, promote efficient scientific process and ultimately accelerate translation into clinical practice.

Show abstract We present a highly compact and fast varifocal lens with aspherical tunability based on an active piezo-glass-piezo sandwich membrane. Using an optimized geometry, improved fabrication and compliant elastomer structures together with an index-matched optical fluid, we achieved an outer diameter of just 9 mm (10 mm packaged) for a clear aperture of 7.6 mm. The range of the focal power was -7 m−1 to +6 m−1, with a wavefront error around 100 nm and a response time between 0.1 and 0.15 ms.

Show abstract We present a highly compact and fast varifocal lens with aspherical tunability based on an active piezo-glass-piezo sandwich membrane. Using an optimized geometry, improved fabrication and compliant elastomer structures together with an index-matched optical fluid, we achieved an outer diameter of just 9 mm (10 mm packaged) for a clear aperture of 7.6 mm. The range of the focal power was -7 m\&\#x2212;1 to $+$6 m\&\#x2212;1, with a wavefront error around 100 nm and a response time between 0.1 and 0.15 ms.

Show abstract Deep generative networks trained via maximum likelihood on a natural image dataset like CIFAR10 often assign high likelihoods to images from datasets with different objects (e.g., SVHN). We refine previous investigations of this failure at anomaly detection for invertible generative networks and provide a clear explanation of it as a combination of model bias and domain prior: Convolutional networks learn similar low-level feature distributions when trained on any natural image dataset and these low-level features dominate the likelihood. Hence, when the discriminative features between inliers and outliers are on a high-level, e.g., object shapes, anomaly detection becomes particularly challenging. To remove the negative impact of model bias and domain prior on detecting high-level differences, we propose two methods, first, using the log likelihood ratios of two identical models, one trained on the in-distribution data (e.g., CIFAR10) and the other one on a more general distribution of images (e.g., 80 Million Tiny Images). We also derive a novel outlier loss for the in-distribution network on samples from the more general distribution to further improve the performance. Secondly, using a multi-scale model like Glow, we show that low-level features are mainly captured at early scales. Therefore, using only the likelihood contribution of the final scale performs remarkably well for detecting high-level feature differences of the out-of-distribution and the in-distribution. This method is especially useful if one does not have access to a suitable general distribution. Overall, our methods achieve strong anomaly detection performance in the unsupervised setting, and only slightly underperform state-of-the-art classifier-based methods in the supervised setting.

Show abstract ABSTRACT The online usage of brain-computer interfaces (BCI) generates unlabeled data. This data in combination with the rich structure contained in BCI applications based on event-related potentials allow to design novel unsupervised classification approaches like learning from label proportions (LLP) or its combination with expectation-maximization (EM) into a mixed model. In this work, we explore the feasibility of unsupervised classification in a BCI chess application. We propose an LLP extension based on weighted least squares regression. It requires randomization of timing parameters but overcomes the dependency on additional symbols. Simulations on electroencephalogram data obtained from six subjects playing BCI-controlled chess show that a combination of unsupervised LLP with EM (despite not using any labels) by constant adaptation quickly reaches and on the long run outperforms the average performance level of non-adaptive supervised classifiers. With our contribution, we increase the scope for which unsupervised learning methods can successfully be applied in BCI.

Show abstract Der menschliche K{\"o}rper funktioniert elektrisch, zumindest teilweise. In der Medizintechnik hat dies zu Ger{\"a}ten zur Aufnahme von Herz- und Hirnsignalen und zu Herzschrittmachern gef{\"u}hrt. Elektrisch aktive Implantate, die Signale nat{\"u}rlicher Sensoren und Organe im menschlichen K{\"o}rper ersetzen oder {\"u}berschreiben, k{\"o}nnen alternative Behandlungsmethoden zu pharmazeutischen L{\"o}sungen bieten. „Neuronale Implantate``, „Elektrozeutika``, „elektronische Pillen`` und „bioelektronische Medizin`` sind Begriffe, die dieses Forschungsfeld beschreiben. Klinische Anwendungen im Bereich dieser Neurotechnik erm{\"o}glichen das H{\"o}ren mit Hilfe von Cochlea Implantaten oder das Unterdr{\"u}cken von Symptomen bei Parkinson Patienten mit Tiefer Hirn Stimulatoren. Die Forschung ist bereits einen Schritt weiter, steuert Hilfsmittel mit Hirnsignalen, stattet Prothesen mit Gef{\"u}hl aus. Dieser Beitrag gibt einen {\"U}berblick {\"u}ber grundlegende Ideen und Ans{\"a}tze sowie M{\"o}glichkeiten und Herausforderungen auf dem Gebiet der Neurotechnik anhand der Beispiele zur Blutdrucksenkung durch Vagusnervstimulation und zur Wiederherstellung sensorischen Feedbacks nach Amputation.

Show abstract Abstract Electrical cochlear implants (eCIs) partially restore hearing and enable speech comprehension to more than half a million users, thereby re-connecting deaf patients to the auditory scene surrounding them. Yet, eCIs suffer from limited spectral selectivity, resulting from current spread around each electrode contact and causing poor speech recognition in the presence of background noise. Optogenetic stimulation of the auditory nerve might overcome this limitation as light can be conveniently confined in space. Here, we combined virus-mediated optogenetic manipulation of cochlear spiral ganglion neurons (SGNs) and microsystems engineering to establish acute multi-channel optical cochlear implant (oCI) stimulation in adult Mongolian gerbils. oCIs based on 16 microscale thin-film light-emitting diodes (μLEDs) evoked tonotopic activation of the auditory pathway with high spectral selectivity and modest power requirements in hearing and deaf gerbils. These results prove the feasibility of μLED-based oCIs for spectrally selective activation of the auditory nerve.

Show abstract Current myoelectric prostheses allow transradial amputees to regain voluntary motor control of their artificial limb by exploiting residual muscle function in the forearm. However, the overreliance on visual cues resulting from a lack of sensory feedback is a common complaint. Recently, several groups have provided tactile feedback in upper limb amputees using implanted electrodes, surface nerve stimulation, or sensory substitution. These approaches have led to improved function and prosthesis embodiment. Nevertheless, the provided information remains limited to a subset of the rich sensory cues available to healthy individuals. More specifically, proprioception, the sense of limb position and movement, is predominantly absent from current systems. Here, we show that sensory substitution based on intraneural stimulation can deliver position feedback in real time and in conjunction with somatotopic tactile feedback. This approach allowed two transradial amputees to regain high and close-to-natural remapped proprioceptive acuity, with a median joint angle reproduction precision of 9.1° and a median threshold to detection of passive movements of 9.5°, which was comparable with results obtained in healthy participants. The simultaneous delivery of position information and somatotopic tactile feedback allowed both amputees to discriminate the size and compliance of four objects with high levels of performance (75.5%). These results demonstrate that tactile information delivered via somatotopic neural stimulation and position information delivered via sensory substitution can be exploited simultaneously and efficiently by transradial amputees. This study paves a way to more sophisticated bidirectional bionic limbs conveying richer, multimodal sensations.

Show abstract Key challenges for the deployment of reinforcement learning (RL) agents in the real world are the discovery, representation and reuse of skills in the absence of a reward function. To this end, we propose a novel approach to learn a task-agnostic skill embedding space from unlabeled multi-view videos. Our method learns a general skill embedding independently from the task context by using an adversarial loss. We combine a metric learning loss, which utilizes temporal video coherence to learn a state representation, with an entropy regularized adversarial skill-transfer loss. The metric learning loss learns a disentangled representation by attracting simultaneous viewpoints of the same observations and repelling visually similar frames from temporal neighbors. The adversarial skill-transfer loss enhances re-usability of learned skill embeddings over multiple task domains. We show that the learned embedding enables training of continuous control policies to solve novel tasks that require the interpolation of previously seen skills. Our extensive evaluation with both simulation and real world data demonstrates the effectiveness of our method in learning transferable skills from unlabeled interaction videos and composing them for new tasks.

Show abstract Intrinsically generated patterns of coupled neuronal activity are associated with the dynamics of specific brain states. Sensory inputs are extrinsic factors that can perturb these intrinsic coupling modes, creating a complex scenario in which forthcoming stimuli are processed. Studying this intrinsic-extrinsic interplay is necessary to better understand perceptual integration and selection. Here, we show that this interplay leads to a reconfiguration of functional cortical connectivity that acts as a mechanism to facilitate stimulus processing. Using audiovisual stimulation in anesthetized ferrets, we found that this reconfiguration of coupling modes is context specific, depending on long-term modulation by repetitive sensory inputs. These reconfigured coupling modes lead to changes in latencies and power of local field potential responses that support multisensory integration. Our study demonstrates that this interplay extends across multiple time scales and involves different types of intrinsic coupling. These results suggest a previously unknown large-scale mechanism that facilitates multisensory integration.

Show abstract Intrinsically generated patterns of coupled neuronal activity are associated with the dynamics of specific brain states. Sensory inputs are extrinsic factors that can perturb these intrinsic coupling modes, creating a complex scenario in which forthcoming stimuli are processed. Studying this intrinsic-extrinsic interplay is necessary to better understand perceptual integration and selection. Here, we show that this interplay leads to a reconfiguration of functional cortical connectivity that acts as a mechanism to facilitate stimulus processing. Using audiovisual stimulation in anesthetized ferrets, we found that this reconfiguration of coupling modes is context specific, depending on long-term modulation by repetitive sensory inputs. These reconfigured coupling modes lead to changes in latencies and power of local field potential responses that support multisensory integration. Our study demonstrates that this interplay extends across multiple time scales and involves different types of intrinsic coupling. These results suggest a previously unknown large-scale mechanism that facilitates multisensory integration.

Show abstract Fast-spiking parvalbumin-expressing interneurons (PVIs) and granule cells (GCs) of the dentate gyrus receive layer-specific dendritic inhibition. Its impact on PVI and GC excitability is, however, unknown. By applying whole-cell recordings, GABA uncaging and single-cell-modeling, we show that proximal dendritic inhibition in PVIs is less efficient in lowering perforant path-mediated subthreshold depolarization than distal inhibition but both are highly efficient in silencing PVIs. These inhibitory effects can be explained by proximal shunting and distal strong hyperpolarizing inhibition. In contrast, GC proximal but not distal inhibition is the primary regulator of their excitability and recruitment. In GCs inhibition is hyperpolarizing along the entire somato-dendritic axis with similar strength. Thus, dendritic inhibition differentially controls input-output transformations in PVIs and GCs. Dendritic inhibition in PVIs is suited to balance PVI discharges in dependence on global network activity thereby providing strong and tuned perisomatic inhibition that contributes to the sparse representation of information in GC assemblies.

Show abstract Advances in sensing technology raise the possibility of creating neural interfaces that can more effectively restore or repair neural function and reveal fundamental properties of neural information processing. To realize the potential of these bioelectronic devices, it is necessary to understand the capabilities of emerging technologies and identify the best strategies to translate these technologies into products and therapies that will improve the lives of patients with neurological and other disorders. Here, we discuss emerging technologies for sensing brain activity, anticipated challenges for translation, and perspectives for how to best transition these technologies from academic research labs to useful products for neuroscience researchers and human patients.

Show abstract Striatal projection neurons, the medium spiny neurons (MSNs), play a crucial role in various motor and cognitive functions. MSNs express either D1 or D2 type dopamine receptors and initiate the direct-pathway (dMSNs) or indirect pathways (iMSNs) of the basal ganglia, respectively. dMSNs have been shown to receive more inhibition than iMSNs from intrastriatal sources. Based on these findings, computational modelling of the striatal network has predicted that under healthy conditions dMSNs should receive more total input than iMSNs. To test this prediction, we analyzed in vivo whole-cell recordings from dMSNs and iMSNs in healthy and dopamine-depleted (6OHDA) anaesthetized mice. By comparing their membrane potential fluctuations, we found that dMSNs exhibited considerably larger membrane potential fluctuations over a wide frequency range. Furthermore, by comparing the spike-triggered average membrane potentials, we found that dMSNs depolarized towards the spike threshold significantly faster than iMSNs did. Together, these findings (in particular the STA analysis) corroborate the theoretical prediction that direct-pathway MSNs receive stronger total input than indirect-pathway neurons. Finally, we found that dopamine-depleted mice exhibited no difference between the membrane potential fluctuations of dMSNs and iMSNs. These data provide new insights into the question how the lack of dopamine may lead to behavioral deficits associated with Parkinson's disease.

Show abstract Current treatment concepts for epilepsy are based on continuous drug delivery or electrical stimulation to prevent the occurrence of seizures, exposing the brain and body to a mostly unneeded risk of adverse effects. To address the infrequent occurrence and short duration of epileptic seizures, intelligent implantable closed-loop devices are needed which are based on a refined analysis of ongoing brain activity with highly specific and fast detection algorithms to allow for timely, ictal interventions. Since the development and FDA approval of a first closed loop neurostimulation device relying on simple threshold-based approaches, machine learning approaches became widely available, probably outperformed in the near future by deep convolutional neural networks, which already showed to be extremely successful in pattern recognition in images and partly in signal analysis. Handcrafted features or rules defined by experts become replaced by systematic feature selection procedures and systematic hyperparameter search approaches. Training of these classifiers augments the need of large databases with intracranial EEG recordings, which is partly given by existing databases but potentially can be replaced by continuously transferring data from implanted devices and their publication for research purposes. Already in early design states, the final target hardware must be taken into account for algorithm development. Size, power consumption and, as a consequence, limited computational resources given by low power microcontrollers, FPGAs and ASICS limit the complexity of feature computation, classifier complexity, and the numbers and complexity of layers of deep neuronal networks. Novel approaches for early seizure detection will be a key module for new generations of closed-loop devices together with improved low power implant hardware and will provide together with more efficient intervention paradigms new treatment options for patients with difficult to treat epilepsy.

Show abstract Technological advances in electrically active implantable devices have increased the complexity of hardware design. In particular, the increasing number of stimulation and recording channels requires innovative approaches for connectors that interface electrodes with the implant circuitry. This work aims to provide a common theoretical ground for implantable connector development with a focus on neural applications. Aspects and experiences from several disciplines are compiled from an engineering perspective to discuss the state of the art of connector solutions. Whenever available, we also present general design guidelines. Degradation mechanisms, material stability and design rules in terms of biocompatibility and biostability are introduced. Considering contact physics, we address the design and characterization of the contact zone and review contaminants, wear and contact degradation. For high-channel counts and body-like environments, insulation can be even more crucial than the electrical connection itself. Therefore, we also introduce the requirements for electrical insulation to prevent signal loss and distortion and discuss its impact on the practical implementation. A final review is dedicated to the state of the art connector concepts, their mechanical setup, electrical performance and the interface to other implant components. We conclude with an outlook for possible approaches for the future generations of implants.

Show abstract In this study, the release of fluorescein from a photo‐crosslinked conducting polymer hydrogel made from a hydrogel precursor poly(dimethylacrylamide‐co‐4‐methacryloyloxy benzophenone (5%)‐co‐4‐styrenesulfonate (2.5%)) (PDMAAp) and the conducting polymer poly(3,4‐ethylenedioxythiophene) (PEDOT) is investigated. Fluorescein, here used as a model for a drug, is actively released through application of an electrical trigger signal. The detected quantity is more than six times higher in comparison to that released from a conventional PEDOT/polysterene sulfonate (PSS) system. Release profiles, drug dose, and timing can be tailored by the application of different trigger signals and pretreatments. To demonstrate that the novel drug release system can be used for a drug relevant for local delivery to a neural interface, experiments are furthermore performed with the anti‐inflammatory drug dexamethasone (Dex). The conducting polymer hydrogel facilitates the active release of Dex, in comparison to the previously used PEDOT/Dex. It is suggested that PEDOT/PDMAAp is an interesting alternative for conducting polymer based drug release systems, with the potential to offer more volume for storage, yet retaining the excellent electrochemical properties known for PEDOT electrodes.

Show abstract Lower limb amputation (LLA) destroys the sensory communication between the brain and the external world during standing and walking. Current prostheses do not restore sensory feedback to amputees, who, relying on very limited haptic information from the stump-socket interaction, are forced to deal with serious issues: the risk of falls, decreased mobility, prosthesis being perceived as an external object (low embodiment), and increased cognitive burden. Poor mobility is one of the causes of eventual device abandonment. Restoring sensory feedback from the missing leg of above-knee (transfemoral) amputees and integrating the sensory feedback into the sensorimotor loop would markedly improve the life of patients. In this study, we developed a leg neuroprosthesis, which provided real-time tactile and emulated proprioceptive feedback to three transfemoral amputees through nerve stimulation. The feedback was exploited in active tasks, which proved that our approach promoted improved mobility, fall prevention, and agility. We also showed increased embodiment of the lower limb prosthesis (LLP), through phantom leg displacement perception and questionnaires, and ease of the cognitive effort during a dual-task paradigm, through electroencephalographic recordings. Our results demonstrate that induced sensory feedback can be integrated at supraspinal levels to restore functional abilities of the missing leg. This work paves the way for further investigations about how the brain interprets different artificial feedback strategies and for the development of fully implantable sensory-enhanced leg neuroprostheses, which could drastically ameliorate life quality in people with disability.

Show abstract Emerging virtual reality systems offer intriguing therapeutic possibilities, but their development and use should be guided by ethical priorities that account for the specific vulnerabilities of patients.

Show abstract Cognitive control is proposed to rely on a rostral-to-caudal hierarchy of neural processing within the prefrontal cortex (PFC), with more rostral parts exerting control over more caudal parts. Anatomical and functional data suggest that this hierarchical organization of the PFC may be separated into a ventral and a dorsal component. Furthermore, recent studies indicate that the apex of the hierarchy resides within the mid-lateral rather the rostral PFC. However, investigating the hierarchical aspect of rostro-to-caudal processing requires quantification of the directed interactions between PFC regions. Using functional near-infrared spectroscopy (fNIRS) in a sample of healthy young adults we analyzed directed interactions between rostral and caudal PFC during passive watching of nature documentaries. Directed coherence (DC) as a measure of directed interaction was computed pairwise between 38 channels evenly distributed over the lateral prefrontal convexity. Results revealed an overall predominance of rostral-to-caudal directed interactions in the PFC that further dissociated along a ventro-dorsal axis: Dorsal regions exerted stronger rostro-caudally directed interactions on dorsal than on ventral regions and vice versa. Interactions between ventral and dorsal PFC were stronger from ventral to dorsal areas than vice versa. Results further support the notion that the mid-dorsolateral PFC constitutes the apex of the prefrontal hierarchy. Taken together these data provide novel evidence for parallel dorsal and ventral streams within the rostro-caudal hierarchical organization of the PFC. FNIRS-based analyses of directed interactions put forward a new perspective on the functional architecture of the prefrontal hierarchy and complement previous insights from functional magnetic resonance imaging.

Show abstract Selective gene delivery into subtypes of interneurons remains an important challenge in vector development. Adeno-associated virus (AAV) vector particles are especially promising for intracerebral injections. For cell entry, AAV2 particles are supposed to attach to heparan-sulfate proteoglycans (HSPGs) followed by endocytosis via the AAV receptor (AAVR). Here, we assessed engineered AAV particles deficient in HSPG attachment but competent in recognizing the glutamate receptor 4 (GluA4, also known as GluRD or GRIA4) through a displayed GluA4-specific DARPin (designed ankyrin repeat protein). When injected into the mouse brain, histological evaluation revealed that in various regions, more than 90% of the transduced cells were interneurons, mainly of the parvalbumin-positive subtype. Although part of the selectivity was mediated by the DARPin, the chosen spleen focus-forming virus (SFFV) promoter had contributed as well. Further analysis revealed that the DARPin mediated selective attachment to GluA4-positive cells, whereas gene delivery required expression of AAVR. Our data suggest that cell selectivity of AAV particles can be modified rationally and efficiently through DARPins, but expression of the AAV entry receptor remains essential.

Show abstract Objective: Resonant vibrations of implanted structures during an MRI procedure pose a risk to the patient in the form of soft tissue irritation and degradation of the implant. In this study, the mechanical behaviour of implant structures in air, water, and viscoelastic materials was explored. Methods: The static and dynamic transfer functions of various test samples in air, and immersed in both water and hydrogels, were analysed. The laser-based acquisition method allowed for high angular resolution (10 μDeg) and high dynamic range (0 to 6 kHz) measurements. Additional MRI experiments were conducted to investigate the dependence of vibration strength on MR sequence parameters in combination with the obtained transfer functions. Results: The largest forces were found to be in the μN to mN range, which is comparable to forces applied during implantation. Of additional concern was the damping introduced by viscoelastic tissue, which was less than expected, leading to an underdamped system. In contrast to current wisdom, the imaging experiments demonstrated drastically different vibration amplitudes for identical gradient slope but different timing parameters TR, mainly due to resonant amplification. Conclusion: The results showed that a safe, force-free MR procedure depends not only on the gradient slope, but also and more drastically on the choice of secure timing parameters. Significance: These findings delineate design improvements to achieve longevity of implants, and will lead to increased patient safety during MRI. A prudent choice of mechanical characteristics of implanted structures is sufficient to avoid resonant excitation due to mismatched MR sequence parameters.

Show abstract Smiling, laughing, and overt speech production are fundamental to human everyday communication. However, little is known about how the human brain achieves the highly accurate and differentiated control of such orofacial movement during natural conditions. Here, we utilized the high spatiotemporal resolution of subdural recordings to elucidate how human motor cortex is functionally engaged during control of real-life orofacial motor behaviour. For each investigated movement class—lip licking, speech production, laughing and smiling—our findings reveal a characteristic brain activity pattern within the mouth motor cortex with both spatial segregation and overlap between classes. Our findings thus show that motor cortex relies on sparse and action-specific activation during real-life orofacial behaviour, apparently organized in distinct but overlapping subareas that control different types of natural orofacial movements.

Show abstract Objective. Tactile afferents in the human hand provide fundamental information about hand-environment interactions, which is used by the brain to adapt the motor output to the physical properties of the object being manipulated. A hand amputation disrupts both afferent and efferent pathways from/to the hand, completely invalidating the individual's motor repertoire. Although motor functions may be partially recovered by using a myoelectric prosthesis, providing functionally effective sensory feedback to users of prosthetics is a largely unsolved challenge. While past studies using invasive stimulation suggested that sensory feedback may help in handling fragile objects, none explored the underpinning, relearned, motor coordination during grasping. In this study, we aimed at showing for the first time that intraneural sensory feedback of the grip force (GF) improves the sensorimotor control of a transradial amputee controlling a myoelectric prosthesis. Approach. We performed a longitudinal study testing a single subject (clinical trial registration number NCT02848846). A stacking cups test (CUP) performed over two weeks aimed at measuring the subject's ability to finely regulate the GF applied with the prosthesis. A pick and lift test (PLT), performed at the end of the study, measured the level of motor coordination, and whether the subject transferred the motor skills learned in the CUP to an alien task. Main results. The results show that intraneural sensory feedback increases the subject's ability in regulating the GF and allows for improved performance over time. Additionally, the PLT demonstrated that the subject was able to generalize and transfer her manipulation skills to an unknown task and to improve her motor coordination. Significance. Our findings suggest that intraneural sensory feedback holds the potential of restoring functionally effective tactile feedback. This opens up new possibilities to improve the quality of life of amputees using a neural prosthesis.

Show abstract Neurological disorders are increasingly analysed and treated with implantable electrodes, and patients with such electrodes are studied with MRI despite the risk of radio-frequency (RF) induced heating during the MRI exam. Recent clinical research suggests that electrodes with smaller diameters of the electrical interface between implant and tissue are beneficial; however, the influence of this electrode contact diameter on RF-induced heating has not been investigated. In this work, electrode contact diameters between 0.3 and 4 mm of implantable electrodes appropriate for stimulation and electrocorticography were evaluated in a 1.5 T MRI system. In situ temperature measurements adapted from the ASTM standard test method were performed and complemented by simulations of the specific absorption rate (SAR) to assess local SAR values, temperature increase and the distribution of dissipated power. Measurements showed temperature changes between 0.8 K and 53 K for different electrode contact diameters, which is well above the legal limit of 1 K. Systematic errors in the temperature measurements are to be expected, as the temperature sensors may disturb the heating pattern near small electrodes. Compared to large electrodes, simulations suggest that small electrodes are subject to less dissipated power, but more localized power density. Thus, smaller electrodes might be classified as safe in current certification procedures but may be more likely to burn adjacent tissue. To assess these local heating phenomena, smaller temperature sensors or new non-invasive temperature sensing methods are needed.

Show abstract Neuroprostheses aimed to restore lost functions after a limb amputation are based on the interaction with the nervous system by means of neural interfaces. Among the different designs, intraneural electrodes implanted in peripheral nerves represent a good strategy to stimulate nerve fibers to send sensory feedback and to record nerve signals to control the prosthetic limb. However, intraneural electrodes, as any device implanted in the body, induce a foreign body reaction (FBR) that results in the tissue encapsulation of the device. The FBR causes a progressive decline of the electrode functionality over time due to the physical separation between the electrode active sites and the axons to interface. Modulation of the inflammatory response has arisen as a good strategy to reduce the FBR and maintain electrode functionality. In this study transversal intraneural multi-channel electrodes (TIMEs) were implanted in the rat sciatic nerve and tested for 3 months to evaluate stimulation and recording capabilities under chronic administration of dexamethasone. Dexamethasone treatment significantly reduced the threshold for evoking muscle responses during the follow-up compared to saline-treated animals, without affecting the selectivity of stimulation. However, dexamethasone treatment did not improve the signal-to-noise ratio of the recorded neural signals. Dexamethasone treatment allowed to maintain more working active sites along time than saline treatment. Thus, systemic administration of dexamethasone appears as a useful treatment in chronically implanted animals with neural electrodes as it increases the number of functioning contacts of the implanted TIME and reduces the intensity needed to stimulate the nerve.

Show abstract Transcranial direct current stimulation (tDCS) is a variant of noninvasive neuromodulation, which promises treatment for brain diseases like major depressive disorder. In experiments, long-lasting aftereffects were observed, suggesting that persistent plastic changes are induced. The mechanism underlying the emergence of lasting aftereffects, however, remains elusive. Here we propose a model, which assumes that tDCS triggers a homeostatic response of the network involving growth and decay of synapses. The cortical tissue exposed to tDCS is conceived as a recurrent network of excitatory and inhibitory neurons, with synapses subject to homeostatically regulated structural plasticity. We systematically tested various aspects of stimulation, including electrode size and montage, as well as stimulation intensity and duration. Our results suggest that transcranial stimulation perturbs the homeostatic equilibrium and leads to a pronounced growth response of the network. The stimulated population eventually eliminates excitatory synapses with the unstimulated population, and new synapses among stimulated neurons are grown to form a cell assembly. Strong focal stimulation tends to enhance the connectivity within new cell assemblies, and repetitive stimulation with well-chosen duty cycles can increase the impact of stimulation even further. One long-term goal of our work is to help in optimizing the use of tDCS in clinical applications.

Show abstract Transcranial direct current stimulation (tDCS) is a variant of noninvasive neuromodulation, which promises treatment for brain diseases like major depressive disorder. In experiments, long-lasting aftereffects were observed, suggesting that persistent plastic changes are induced. The mechanism underlying the emergence of lasting aftereffects, however, remains elusive. Here we propose a model, which assumes that tDCS triggers a homeostatic response of the network involving growth and decay of synapses. The cortical tissue exposed to tDCS is conceived as a recurrent network of excitatory and inhibitory neurons, with synapses subject to homeostatically regulated structural plasticity. We systematically tested various aspects of stimulation, including electrode size and montage, as well as stimulation intensity and duration. Our results suggest that transcranial stimulation perturbs the homeostatic equilibrium and leads to a pronounced growth response of the network. The stimulated population eventually eliminates excitatory synapses with the unstimulated population, and new synapses among stimulated neurons are grown to form a cell assembly. Strong focal stimulation tends to enhance the connectivity within new cell assemblies, and repetitive stimulation with well-chosen duty cycles can increase the impact of stimulation even further. One long-term goal of our work is to help in optimizing the use of tDCS in clinical applications.

Show abstract Behavioral deficits after stroke like apraxia can be related to structural lesions and to a functional state of the underlying network - three factors, reciprocally influencing each other. Combining lesion data, behavioral performance and passive functional activation of the network-of-interest, this study aims to disentangle those mutual influences and to identify 1) activation patterns associated with the presence or absence of acute apraxia in tool-associated actions and 2) the specific impact of lesion location on those activation patterns. Brain activity of 48 patients (63.31 ± 13.68 years, 35 male) was assessed in a fMRI paradigm with observation of tool-related actions during the acute phase after first-ever left-hemispheric stroke (4.83 ± 2.04 days). Behavioral assessment of apraxia in tool-related tasks was obtained independently. Brain activation was compared between patients versus healthy controls and between patient with versus without apraxia. Interaction effects of lesion location (frontal vs parietal) and behavioral performance (apraxia vs no apraxia) were assessed in a 2 × 2 factorial design. Action observation activated the ventro-dorsal parts of the network for cognitive motor function; activation was globally downregulated after stroke. Apraxic compared to non-apraxic patients showed relatively increased activity in bilateral posterior middle temporal gyrus and middle frontal gyrus/superior frontal sulcus. Altered activation occurred in regions for tool-related cognition, corroborating known functions of the ventro-dorsal and ventral streams for praxis, and comprised domain-general areas, functionally related to cognitive control. The interaction analyses revealed different levels of activation in the left anterior middle temporal gyrus in the ventral stream in apraxic patients with frontal compared to parietal lesions, suggesting a modulation of network activation in relation to behavioral performance and lesion location as separate factors. By detecting apraxia-specific activation patterns modulated by lesion location, this study underlines the necessity to combine structural lesion information, behavioral parameters and functional activation to comprehensively examine cognitive functions in acute stroke patients.

Show abstract Background: Following an amputation, the human postural control system develops neuromuscular adaptations to regain an effective postural control. We investigated the compensatory mechanisms behind these adaptations and how sensorimotor integration is affected after a lower-limb transfemoral amputation. METHODS: Center of pressure (CoP) data of 12 unilateral transfemoral amputees and 12 age-matched able-bodied subjects were recorded during quiet standing with eyes open (EO) and closed (EC). CoP adjustments under each leg were recorded to study their contribution to posture control. The spatial structure of the CoP displacements was characterized by measuring the mean distance, the mean velocity of the CoP adjustments, and the sway area. The Entropic Half-Life (EnHL) quantifies the temporal structure of the CoP adjustments and was used to infer disrupted sensory feedback loops in amputees. We expanded the analysis with measures of weight-bearing imbalance and asymmetry, and with two standardized balance assessments, the Berg Balance Scale (BBS) and Timed Up-and-Go (TUG). RESULTS: There was no difference in the EnHL values of amputees and controls when combining the contributions of both limbs (p = 0.754). However, amputees presented significant differences between the EnHL values of the intact and prosthetic limb (p <  0.001). Suppressing vision reduced the EnHL values of the intact (p = 0.001) and both legs (p = 0.028), but not in controls. Vision feedback in amputees also had a significant effect (increase) on the mean CoP distance (p <  0.001), CoP velocity (p <  0.001) and sway area (p = 0.007). Amputees presented an asymmetrical stance. The EnHL values of the intact limb in amputees were positively correlated to the BBS scores (EO: ρ = 0.43, EC: ρ = 0.44) and negatively correlated to the TUG times (EO: ρ = - 0.59, EC: ρ = - 0.69). CONCLUSION: These results suggest that besides the asymmetry in load distribution, there exist neuromuscular adaptations after an amputation, possibly related to the loss of sensory feedback and an altered sensorimotor integration. The EnHL values suggest that the somatosensory system predominates in the control of the intact leg. Further, suppressing the visual system caused instability in amputees, but had a minimal impact on the CoP dynamics of controls. These findings points toward the importance of providing somatosensory feedback in lower-limb prosthesis to reestablish a normal postural control.

Show abstract Objective. Electrochemical microsensors based on noble metals can give essential information on their microenvironment with high spatio-temporal resolution. However, most advanced chemo- and biosensors lack the long-term stability for physiological monitoring of brain tissue beyond an acute application. Noble metal electrodes are widely used as neural interfaces, particularly for stimulating in the central nervous system. Our goal was to recruit already deployed, unmodified noble metal electrodes (Pt, Pt/Ir) as in situ chemical sensors. Approach. With advanced electrochemical sensor methods, we investigated electrode surface processes, oxidizable species and oxygen as an indicator for tissue mass transport. We developed a unique, multi-step, amperometric/potentiometric sensing procedure derived from the investigation of Pt surface processes by chronocoulometry providing fundamental characterization of the electrode itself. Main results. The resulting electrochemical protocol preconditions the electrode, measures oxidizable and reducible species, and the open circuit potential (OCP). A linear, stable sensor performance was demonstrated, also in the presence of proteins, validating signal stability of our cyclic protocol in complex environments. We investigated our sensor protocol with microelectrodes on custom Pt/Ir-wire tetrodes by in vivo measurements in the rat brain for up to four weeks. Results showed that catalytic activity of the electrode is lost over time, but our protocol is repeatedly able to both quantify and restore electrode sensitivity in vivo. Significance. Our approach is highly relevant because it can be applied to any existing Pt electrode. Current methods to assess the brain/electrode microenvironment mainly rely on imaging techniques, histology and analysis of explanted devices, which are often end-point methods. Our procedure delivers online and time-transient information on the chemical microenvironment directly at the electrode/tissue interface of neural implants, gives new insight into the charge transfer processes, and delivers information on the state of the electrode itself addressing long-term electrode degradation.

Show abstract Objective. Electrochemical microsensors based on noble metals can give essential information on their microenvironment with high spatio‐temporal resolution. However, most advanced chemo‐ and biosensors lack the long‐term stability for physiological monitoring of brain tissue beyond an acute application. Noble metal electrodes are widely used as neural interfaces, particularly for stimulating in the central nervous system. Our goal was to recruit already deployed, unmodified noble metal electrodes (Pt, Pt/Ir) as in situ chemical sensors. Approach. With advanced electrochemical sensor methods, we investigated electrode surface processes, oxidizable species and oxygen as an indicator for tissue mass transport. We developed a unique, multi‐step, amperometric/potentiometric sensing procedure derived from the investigation of Pt surface processes by chronocoulometry providing fundamental characterization of the electrode itself. Main results. The resulting electrochemical protocol preconditions the electrode, measures oxidizable and reducible species, and the open circuit potential. A linear, stable sensor performance was demonstrated, also in the presence of proteins, validating signal stability of our cyclic protocol in complex environments. We investigated our sensor protocol with microelectrodes on custom Pt/Ir‐wire tetrodes by in vivo measurements in the rat brain for up to four weeks. Results showed that catalytic activity of the electrode is lost over time, but our protocol is repeatedly able to both quantify and restore electrode sensitivity in vivo. Significance. Our approach is highly relevant because it can be applied to any existing Pt electrode. Current methods to assess the brain/electrode microenvironment mainly rely on imaging techniques, histology and analysis of explanted devices, which are often end‐point methods. Our procedure delivers online and time‐transient information on the chemical microenvironment directly at the electrode/tissue interface of neural implants, gives new insight into the charge transfer processes, and delivers information on the state of the electrode itself addressing long‐term electrode degradation.

Show abstract Objectives: Focal cortical dysplasias (FCDs) are local malformations of the human neocortex and a leading cause of medically intractable epilepsy. FCDs are characterized by local architectural disturbances of the neocortex and often by a blurred gray-white matter boundary indicating abnormal white matter myelination. We have recently shown that myelination is also compromised in the gray matter of dysplastic areas, since transcripts encoding factors for oligodendrocyte differentiation and myelination are downregulated and myelin fibers appear fractured and disorganized. Methods: Here, we characterized the gray matter-associated myelination pathology in detail by in situ hybridization, immunohistochemistry, and electron microscopy with markers for myelin, mature oligodendrocytes, and oligodendrocyte precursor cells in tissue sections of FCD IIa and control cortices. In addition, we isolated oligodendrocyte precursor cells from resected dysplastic tissue and performed proliferation assays. Results: We show that the proportion of myelinated gray matter is similar in the dysplastic cortex to that in controls and myelinated fibers extend up to layer III. On the ultrastructural level, however, we found that the myelin sheaths of layer V axons are thinner in dysplastic specimens than in controls. In addition, the density of oligodendrocyte precursor cells and of mature oligodendrocytes was reduced. Finally, we show for the first time that oligodendrocyte precursor cells isolated from resected dysplastic cortex have a reduced proliferation capacity in comparison to controls. Significance: These results indicate that proliferation and differentiation of oligodendrocyte precursor cells and the formation of myelin sheaths are compromised in FCD and might contribute to the epileptogenicity of this cortical malformation. Keywords: epilepsy; human neocortex; myelin; oligodendrocyte precursor; proliferation. © 2019 The Authors. Epilepsia published by Wiley Periodicals, Inc. on behalf of International League Against Epilepsy.

Show abstract Providing somatosensory feedback to amputees is a long-standing objective in prosthesis research. Recently, implantable neural interfaces have yielded promising results in this direction. There is now considerable evidence that the nervous system integrates redundant signals optimally, weighting each signal according to its reliability. One question of interest is whether artificial sensory feedback is combined with other sensory information in a natural manner. In this single-case study, we show that an amputee with a bidirectional prosthesis integrated artificial somatosensory feedback and blurred visual information in a statistically optimal fashion when estimating the size of a hand-held object. The patient controlled the opening and closing of the prosthetic hand through surface electromyography, and received intraneural stimulation proportional to the object’s size in the ulnar nerve when closing the robotic hand on the object. The intraneural stimulation elicited a vibration sensation in the phantom hand that substituted the missing haptic feedback. This result indicates that sensory substitution based on intraneural feedback can be integrated with visual feedback and make way for a promising method to investigate multimodal integration processes.

Show abstract Optogenetics has revolutionized neurobiological research by allowing to disentangle intricate neuronal circuits at a spatio-temporal precision unmatched by other techniques. Here, we review current advances of optogenetic applications in mammals, especially focusing on freely moving animals. State-of-the-art strategies allow the targeted expression of opsins in neuronal subpopulations, defined either by genetic cell type or neuronal projection pattern. Optogenetic manipulations of these subpopulations become particularly powerful when combined with behavioral paradigms and neurophysiological readout techniques. Thereby, specific roles can be assigned to identified cells. All-optical approaches with the opportunity to write complex three dimensional patterns into neuronal networks have recently emerged. While clinical implications of the new tool set seem tempting, we emphasize here the role of optogenetics for basic research.

Show abstract BACKGROUND: Intracranial electroencephalography (iEEG) is increasingly used in neuroscientific research. However, the position of the implanted electrodes varies greatly between patients, which makes group analyses particularly difficult. Therefore, an assignment procedure is needed that enables the neuroanatomical information to be obtained for each individual electrode contact. NEW METHOD: Here, we present a MATLAB-based electrode assignment approach for iEEG electrode contacts, implemented in the open-source toolbox ELAS, that allows a hierarchical probabilistic assignment of individual electrode contacts to cytoarchitectonically-defined brain areas. The here presented ELAS consists of two major steps: (I) a pre-assignment to the cerebral lobes and (II) a following probabilistic assignment based on lobe-specific probability maps of the SPM Anatomy Toolbox. RESULTS: We analyzed iEEG data obtained in 14 epilepsy patients with a total of 783 intracranial electrode contacts. The neuroanatomical assignment to cortical brain areas was possible in 72.5% of the electrode contacts that were located on the lateral cortical convexity. COMPARISON WITH EXISTING METHODS: This assignment procedure is to our knowledge the first approach that combines both individual macro-anatomical and cytoarchitectonic probabilistic information. Due to the integration of information about individual anatomical landmarks, incorrect assignments could be avoided in approx. 7% of electrode contacts. CONCLUSION: The present study demonstrates how probabilistic assignment procedures developed for the analysis of neuroimaging data can be adapted to iEEG, which is especially helpful for group analyses. The presented assignment approach is freely available via the open-source toolbox ELAS, including a 3D visualization and a file export for virtual reality setups.

Show abstract Neurons in different layers of sensory cortex generally have different functional properties. But what determines firing rates and tuning properties of neurons in different layers? Orientation selectivity in primary visual cortex (V1) is an interesting case to study these questions. Thalamic projections essentially determine the preferred orientation of neurons that receive direct input. But how is this tuning propagated though layers, and how can selective responses emerge in layers that do not have direct access to the thalamus? Here we combine numerical simulations with mathematical analyses to address this problem. We find that a large-scale network, which just accounts for experimentally measured layer and cell-type specific connection probabilities, yields firing rates and orientation selectivities matching electrophysiological recordings in rodent V1 surprisingly well. Further analysis, however, is complicated by the fact that neuronal responses emerge in a dynamic fashion and cannot be directly inferred from static neuroanatomy, as some connections tend to have unintuitive effects due to recurrent interactions and strong feedback loops. These emergent phenomena can be understood by linearizing and coarse-graining. In fact, we were able to derive a low-dimensional linear dynamical system effectively describing stimulus-driven activity layer by layer. This low-dimensional system explains layer-specific firing rates and orientation tuning by accounting for the different gain factors of the aggregate system. Our theory can also be used to design novel optogenetic stimulation experiments, thus facilitating further exploration of the interplay between connectivity and function.

Show abstract Controlling the growth of conductive polymers via electrolysis enables defined surface modifications and can be used as a rapid prototyping process. In this study, the controlled dendritic growth of poly(3,4-ethylenedioxythiophene) (PEDOT) in a two-electrode setup was investigated by pulsed voltage-driven electropolymerization of the precursor EDOT and a low concentration of tetrabutylammonium perchlorate dissolved in acetonitrile. Rapid growth of different polymeric shapes was reliably achieved by varying the reduction voltage and duty factor. The obtained structures were optically examined and quantified using fractal dimensions. Their shapes ranged from solid coatings over branched fractals to straight fibers without requiring any template. These rapid and controllable electropolymerization processes were further combined to increase conductor complexity.

Show abstract Neural oscillations are increasingly interpreted as transient bursts, yet a method to measure these short-lived events in real-time is missing. Here we present a real-time data analysis system, capable to detect short and narrowband bursts, and demonstrate its usefulness for volitional increase of beta-band burst-rate in rats. This neurofeedback-training induced changes in overall oscillatory power, and bursts could be decoded from the movement of the rats, thus enabling future investigation of the role of oscillatory bursts.

Show abstract Learning to reliably perceive and understand the scene is an integral enabler for robots to operate in the real-world. This problem is inherently challenging due to the multitude of object types as well as appearance changes caused by varying illumination and weather conditions. Leveraging complementary modalities can enable learning of semantically richer representations that are resilient to such perturbations. Despite the tremendous progress in recent years, most multimodal convolutional neural network approaches directly concatenate feature maps from individual modality streams rendering the model incapable of focusing only on the relevant complementary information for fusion. To address this limitation, we propose a mutimodal semantic segmentation framework that dynamically adapts the fusion of modality-specific features while being sensitive to the object category, spatial location and scene context in a self-supervised manner. Specifically, we propose an architecture consisting of two modality-specific encoder streams that fuse intermediate encoder representations into a single decoder using our proposed self-supervised model adaptation fusion mechanism which optimally combines complementary features. As intermediate representations are not aligned across modalities, we introduce an attention scheme for better correlation. In addition, we propose a computationally efficient unimodal segmentation architecture termed AdapNet++ that incorporates a new encoder with multiscale residual units and an efficient atrous spatial pyramid pooling that has a larger effective receptive field with more than 10× fewer parameters, complemented with a strong decoder with a multi-resolution supervision scheme that recovers high-resolution details. Comprehensive empirical evaluations on Cityscapes, Synthia, SUN RGB-D, ScanNet and Freiburg Forest benchmarks demonstrate that both our unimodal and multimodal architectures achieve state-of-the-art performance while simultaneously being efficient in terms of parameters and inference time as well as demonstrating substantial robustness in adverse perceptual conditions.

Show abstract Conventional leg prostheses do not convey sensory information about motion or interaction with the ground to above-knee amputees, thereby reducing confidence and walking speed in the users that is associated with high mental and physical fatigue1,2,3,4. The lack of physiological feedback from the remaining extremity to the brain also contributes to the generation of phantom limb pain from the missing leg5,6. To determine whether neural sensory feedback restoration addresses these issues, we conducted a study with two transfemoral amputees, implanted with four intraneural stimulation electrodes7 in the remaining tibial nerve (ClinicalTrials.gov identifier NCT03350061). Participants were evaluated while using a neuroprosthetic device consisting of a prosthetic leg equipped with foot and knee sensors. These sensors drive neural stimulation, which elicits sensations of knee motion and the sole of the foot touching the ground. We found that walking speed and self-reported confidence increased while mental and physical fatigue decreased for both participants during neural sensory feedback compared to the no stimulation trials. Furthermore, participants exhibited reduced phantom limb pain with neural sensory feedback. The results from these proof-of-concept cases provide the rationale for larger population studies investigating the clinical utility of neuroprostheses that restore sensory feedback.

Show abstract Conventional leg prostheses do not convey sensory information about motion or interaction with the ground to above-knee amputees, thereby reducing confidence and walking speed in the users that is associated with high mental and physical fatigue1,2,3,4. The lack of physiological feedback from the remaining extremity to the brain also contributes to the generation of phantom limb pain from the missing leg5,6. To determine whether neural sensory feedback restoration addresses these issues, we conducted a study with two transfemoral amputees, implanted with four intraneural stimulation electrodes7 in the remaining tibial nerve (ClinicalTrials.gov identifier NCT03350061). Participants were evaluated while using a neuroprosthetic device consisting of a prosthetic leg equipped with foot and knee sensors. These sensors drive neural stimulation, which elicits sensations of knee motion and the sole of the foot touching the ground. We found that walking speed and self-reported confidence increased while mental and physical fatigue decreased for both participants during neural sensory feedback compared to the no stimulation trials. Furthermore, participants exhibited reduced phantom limb pain with neural sensory feedback. The results from these proof-of-concept cases provide the rationale for larger population studies investigating the clinical utility of neuroprostheses that restore sensory feedback.

Show abstract OBJECTIVE: Hand amputation is a highly disabling event, which significantly affects quality of life. An effective hand replacement can be achieved if the user, in addition to motor functions, is provided with the sensations that are naturally perceived while grasping and moving. Intraneural peripheral electrodes have shown promising results toward the restoration of the sense of touch. However, the long-term usability and clinical relevance of intraneural sensory feedback have not yet been clearly demonstrated. METHODS: To this aim, we performed a 6-month clinical study with 3 transradial amputees who received implants of transverse intrafascicular multichannel electrodes (TIMEs) in their median and ulnar nerves. After calibration, electrical stimulation was delivered through the TIMEs connected to artificial sensors in the digits of a prosthesis to generate sensory feedback, which was then used by the subjects while performing different grasping tasks. RESULTS: All subjects, notwithstanding their important clinical differences, reported stimulation-induced sensations from the phantom hand for the whole duration of the trial. They also successfully integrated the sensory feedback into their motor control strategies while performing experimental tests simulating tasks of real life (with and without the support of vision). Finally, they reported a decrement of their phantom limb pain and a general improvement in mood state. INTERPRETATION: The promising results achieved with all subjects show the feasibility of the use of intraneural stimulation in clinical settings. ANN NEUROL 2019;85:137-154.

Show abstract Spatio-temporally controlled drug release based on conducting polymer films offers a powerful technology to improve the tissue integration for implantable neuroprobes. We here explore the release efficiency of such systems in order to improve the understanding of the release mechanism and allow for optimized implementation of this technology into future drug release applications. By exposing drug loaded PEDOT coatings of different thicknesses to a multitude of release signals, along with optimizing the steps during the polymer synthesis, we could identify a highly reproducible electrostatically controlled drug release next to a slow diffusion driven release component. The release efficiency was moreover observed to be higher for a cyclic voltammetry signal in comparison to release driven by a constant potential. Biphasic current pulses, as used during neural stimulation, did not allow for long enough diffusion times to yield efficient active drug expulsion from the polymer films. A quantitative analysis could confirm an overall linear dependency between drug release and film thickness. The amount of drug released in response to the trigger signals was however not linearly correlated with the amount of charge applied. By combining these findings we could develop a model which accurately describes the drug release mechanism from a PEDOT film. The proposed model thereby points the way for how actively controlled, and diffusion related, release can be tuned for obtaining delivery dynamics tailored to specific applications.

Show abstract In this letter, we deal with the reality gap from a novel perspective, targeting transferring deep reinforcement learning (DRL) policies learned in simulated environments to the real-world domain for visual control tasks. Instead of adopting the common solutions to the problem by increasing the visual fidelity of synthetic images output from simulators during the training phase, we seek to tackle the problem by translating the real-world image streams back to the synthetic domain during the deployment phase, to make the robot feel at home. We propose this as a lightweight, flexible, and efficient solution for visual control, as first, no extra transfer steps are required during the expensive training of DRL agents in simulation; second, the trained DRL agents will not be constrained to being deployable in only one specific real-world environment; and third, the policy training and the transfer operations are decoupled, and can be conducted in parallel. Besides this, we propose a simple yet effective shift loss that is agnostic to the downstream task, to constrain the consistency between subsequent frames which is important for consistent policy outputs. We validate the shift loss for artistic style transfer for videos and domain adaptation, and validate our visual control approach in indoor and outdoor robotics experiments.

Show abstract Future‐oriented directions in neural interface technologies point towards the development of multimodal devices that combine different functionalities such as neural stimulation, neurotransmitter sensing, and drug release within one platform. Conducting polymer hydrogels (CPHs) are suggested as materials for the coating of standard metal electrodes to add functionalities such as local delivery of therapeutic drugs. However, to make such coatings truly useful for multimodal devices, it is necessary to develop process technologies that allow the micropatterning of CPHs onto selected electrode sites. In this study, a wafer‐scale fabrication procedure is presented, which is used to coat the CPH, based on the hydrogel P(DMAA‐co‐5%MABP‐co‐2,5%SSNa) and the conducting polymer poly(3,4‐ethylenedioxythiophene) (PEDOT), onto flexible neural probes. The resulting material has favorable properties for the generation of recording electrodes and in addition offers a convenient platform for biofunctionalization. By controlling the PEDOT content within the hydrogel matrix, charge injection limits of up to 3.7 mC cm−2 are obtained. Long‐term stability is tested by immersing coated samples in phosphate‐buffered saline solution at 37 °C for 1 year. Non‐cytotoxicity of the coatings is confirmed with a direct cell culture test using a fluorescent neuroblastoma cell line.

Show abstract Neurotechnologies describe a field of science and engineering in which the nervous system is interfaced with technical devices. Fundamental research is conducted to explore functions of the brain, decipher the neural code and get a better understanding of diseases and disorders. Risk benefit assessment has been well established in all medical disciplines to treat patients best possible while minimizing jeopardizing their lives by the interventions. Is this set of assessment rules sufficient when the brain will be interfaced with a technical system and is this assessment enough? How will these new technologies change personality and society? This article will shortly review different stakeholders’ opinions and their expectation in the field, assembles information the state-of-the art in medical applications of neurotechnological implants and describes and assesses the fundamental technologies that are used to build up these implants starting with essential requirements of technical materials in contact with living tissue. The different paragraphs guide the reader through the main aspects of neurotechnologies and lay a foundation of knowledge to be able to contribute to the discussion in which cases implants will be beneficial and in which cases we should express serious concerns.

Show abstract The neuromodulator dopamine plays a key role in motivation, reward-related learning and normal motor function. The different affinity of striatal D1 and D2 dopamine receptor types has been argued to constrain the D1 and D2 signalling pathways to phasic and tonic dopamine signals, respectively. However, this view assumes that dopamine receptor kinetics are instantaneous so that the time courses of changes in dopamine concentration and changes in receptor occupation are basically identical. Here we developed a neurochemical model of dopamine receptor binding taking into account the different kinetics and abundance of D1 and D2 receptors in the striatum. Testing a large range of behaviorally-relevant dopamine signals, we found that the D1 and D2 dopamine receptor populations responded very similarly to tonic and phasic dopamine signals. Furthermore, due to slow unbinding rates, both receptor populations integrated dopamine signals over a timescale of minutes. Our model provides a description of how physiological dopamine signals translate into changes in dopamine receptor occupation in the striatum, and explains why dopamine ramps are an effective signal to occupy dopamine receptors. Overall, our model points to the importance of taking into account receptor kinetics for functional considerations of dopamine signalling. Significance statement Current models of basal ganglia function are often based on a distinction of two types of dopamine receptors, D1 and D2, with low and high affinity, respectively. Thereby, phasic dopamine signals are believed to mostly affect striatal neurons with D1 receptors, and tonic dopamine signals are believed to mostly affect striatal neurons with D2 receptors. This view does not take into account the rates for the binding and unbinding of dopamine to D1 and D2 receptors. By incorporating these kinetics into a computational model we show that D1 and D2 receptors both respond to phasic and tonic dopamine signals. This has implications for the processing of reward-related and motivational signals in the basal ganglia.

Show abstract Potentiometric oxygen monitoring using platinum as the electrode material was enabled by the combination of conventional potentiometry with active prepolarization protocols, what we call active potentiometry. The obtained logarithmic transfer function is well-suited for the measurement of dissolved oxygen in biomedical applications, as the physiological oxygen concentration typically varies over several decades. We describe the application of active potentiometry in phosphate buffered salt solution at different pH and ion strength. Sensitivity was in the range of 60 mV/dec oxygen concentration; the transfer function deviated from logarithmic behavior for smaller oxygen concentration and higher ion strength of the electrolyte. Long-term stability was demonstrated for 60 h. Based on these measurement results and additional cyclic voltammetry investigations a model is discussed to explain the potential forming mechanism. The described method of active potentiometry is applicable to many different potentiometric sensors possibly enhancing sensitivity or selectivity for a specific parameter.

Show abstract Modern deep learning methods are very sensitive to many hyperparameters, and, due to the long training times of state-of-the-art models, vanilla Bayesian hyperparameter optimization is typically computationally infeasible. On the other hand, bandit-based configuration evaluation approaches based on random search lack guidance and do not converge to the best configurations as quickly. Here, we propose to combine the benefits of both Bayesian optimization and bandit-based methods, in order to achieve the best of both worlds: strong anytime performance and fast convergence to optimal configurations. We propose a new practical state-of-the-art hyperparameter optimization method, which consistently outperforms both Bayesian optimization and Hyperband on a wide range of problem types, including high-dimensional toy functions, support vector machines, feed-forward neural networks, Bayesian neural networks, deep reinforcement learning, and convolutional neural networks. Our method is robust and versatile, while at the same time being conceptually simple and easy to implement.

Show abstract This paper reports on 1 the development, characterization, and validation of neural probes serving the growing need of neuroscience for miniaturized tools enabling simultaneous high-resolution recording of neural activity in multiple brain areas. The probes consist of a needle-shaped shaft with a crosssection of 100 × 50 μm2 and a length of 10 or 5 mm emerging from a base with dimensions of only 0.55×1.8 mm2. The shafts carry 1600 and 800 recording sites, respectively, grouped into 50 (respectively 25) blocks of 4 × 8 electrodes with an area of 17 × 17 μm2 each, a pitch of 20 μm, and an electrode-to electrode spacing of 3 μm. The probes are fabricated using a commercial 0.18 μm CMOS process followed by dedicated metallization, passivation, and microfabrication steps. Neural signals are accessible through 32 analog output channels via a hierarchical digital addressing scheme implementing an advanced electronic depth control concept giving the option of multiple scanning modes and offering a switching time of 416 μs at a clock frequency of 1 MHz. All output channels are shielded against each other, whereby crosstalk between neighboring channels is measured to be −58 dB at 1 kHz. Absolute impedance values at 1 kHz of single IrOx and Pt electrodes are 230 ± 38 kOhm and 2.2 ± 0.3 MOhm, respectively. In vivo recordings taking advantage of the new addressing concept for high-resolution recordings from multiple brain regions were successfully performed in anesthetized rats.

Show abstract OBJECTIVE: Rolandic epilepsy (RE) is characterized by typical interictal-electroencephalogram (EEG) patterns mainly localized in centrotemporal and parietooccipital areas. An aberrant intrinsic organization of the default mode network (DMN) due to repeated disturbances from spike-generating areas may be able to account for specific cognitive deficits and behavioral problems in RE. The aim of the present study was to investigate cognitive development (CD) and socioemotional development (SED) in patients with RE during active disease in relation to DMN connectivity and network topology. METHODS: In 10 children with RE and active EEG, CD was assessed using the Wechsler Intelligence Scale for Children-IV (WISC-IV); SED was assessed using the Funf-Faktoren-Fragebogen fur Kinder (FFFK), a Big-Five inventory for the assessment of personality traits in children. Functional connectivity (FC) in the DMN was determined from a 15-minute resting state functional magnetic resonance imaging (fMRI), and network properties were calculated using standard graph-theoretical measures. RESULTS: More severe deficits of verbal abilities tended to be associated with an earlier age at epilepsy onset, but were not directly related to the number of seizures and disease duration. Nonetheless, at the network level, disease duration was associated with alterations of the efficiency and centrality of parietal network nodes and midline structures. Particularly, centrality of the left inferior parietal lobe (IPL) was found to be linked with CD. Reduced centrality of the left IPL and alterations supporting a rather segregated processing within DMN's subsystems was associated with a more favorable CD. A more complicated SED was associated with high seizure frequency and long disease duration, and revealed links with a less favorable CD. SIGNIFICANCE: An impaired CD and - because of their interrelation - SED might be mediated by a common pathomechanism reflected in an aberrant organization, and thus, a potential functional deficit of the DMN. A functional segregation of (left) parietal network nodes from the DMN and a rather segregated processing mode within the DMN might have positive implications/protective value for CD in patients with RE.

Show abstract Recent studies have shown that direct nerve stimulation can be used to provide sensory feedback to hand amputees. The intensity of the elicited sensations can be modulated using the amplitude or frequency of the injected stimuli. However, a comprehensive comparison of the effects of these two encoding strategies on the amputees’ ability to control a prosthesis has not been performed. In this paper, we assessed the performance of two trans-radial amputees controlling a myoelectric hand prosthesis while receiving grip force sensory feedback encoded using either linear modulation of amplitude (LAM) or linear modulation of frequency (LFM) of direct nerve stimulation (namely, bidirectional prostheses). Both subjects achieved similar and significantly above-chance performance when they were asked to exploit LAM or LFM in different tasks. The feedbacks allowed them to discriminate, during manipulation through the robotic hand, objects of different compliances and shapes or different placements on the prosthesis. Similar high performances were obtained when they were asked to apply different levels of force in a random order on a dynamometer using LAM or LFM. In contrast, only the LAM strategy allowed the subjects to continuously modulate the grip pressure on the dynamometer. Furthermore, when long-lasting trains of stimulation were delivered, LFM strategy generated a very fast adaptation phenomenon in the subjects, which caused them to stop perceiving the restored sensations. Both encoding approaches were perceived as very different from the touch feelings of the healthy limb (natural). These results suggest that the choice of specific sensory feedback encodings can have an effect on user performance while grasping. In addition, our results invite the development of new approaches to provide more natural sensory feelings to the users, which could be addressed by a more biomimetic strategy in the future.

Show abstract Optogenetic tools have opened a rich experimental landscape for understanding neural function and disease. Here, we present the first validation of eight optogenetic constructs driven by recombinant adeno-associated virus (AAV) vectors and a WGA-Cre based dual injection strategy for projection targeting in a widely-used New World primate model, the common squirrel monkey Saimiri sciureus. We observed opsin expression around the local injection site and in axonal projections to downstream regions, as well as transduction to thalamic neurons, resembling expression patterns observed in macaques. Optical stimulation drove strong, reliable excitatory responses in local neural populations for two depolarizing opsins in anesthetized monkeys. Finally, we observed continued, healthy opsin expression for at least one year. These data suggest that optogenetic tools can be readily applied in squirrel monkeys, an important first step in enabling precise, targeted manipulation of neural circuits in these highly trainable, cognitively sophisticated animals. In conjunction with similar approaches in macaques and marmosets, optogenetic manipulation of neural circuits in squirrel monkeys will provide functional, comparative insights into neural circuits which subserve dextrous motor control as well as other adaptive behaviors across the primate lineage. Additionally, development of these tools in squirrel monkeys, a well-established model system for several human neurological diseases, can aid in identifying novel treatment strategies.

Show abstract Imitation of tool-use gestures (transitive; e.g., hammering) and communicative emblems (intransitive; e.g., waving goodbye) is frequently impaired after left-hemispheric lesions. We aimed 1) to identify lesions related to deficient transitive or intransitive gestures, 2) to delineate regions associated with distinct error types (e.g., hand configuration, kinematics), and 3) to compare imitation to previous data on pantomimed and actual tool use. Of note, 156 patients (64.3 +/- 14.6 years; 56 female) with first-ever left-hemispheric ischemic stroke were prospectively examined 4.8 +/- 2.0 days after symptom onset. Lesions were delineated on magnetic resonance imaging scans for voxel-based lesion-symptom mapping. First, while inferior-parietal lesions affected both gesture types, specific associations emerged between intransitive gesture deficits and anterior temporal damage and between transitive gesture deficits and premotor and occipito-parietal lesions. Second, impaired hand configurations were related to anterior intraparietal damage, hand/wrist-orientation errors to premotor lesions, and kinematic errors to inferior-parietal/occipito-temporal lesions. Third, premotor lesions impacted more on transitive imitation compared with actual tool use, pantomimed and actual tool use were more susceptible to lesioned insular cortex and subjacent white matter. In summary, transitive and intransitive gestures differentially rely on ventro-dorsal and ventral streams due to higher demands on temporo-spatial processing (transitive) or stronger reliance on semantic information (intransitive), respectively.

Show abstract Thin-film neural devices are an appealing alternative to traditional implants, although their chronic stability remains matter of investigation. In this study, a chronically stable class of thin-film devices for electrocorticography is manufactured incorporating silicon carbide and diamond-like carbon as adhesion promoters between glassy carbon (GC) electrodes and polyimide and between GC and platinum traces. The devices are aged in three solutions—phosphate-buffered saline (PBS), 30 × 10−3 and 150 × 10−3m H2O2/PBS—and stressed using cyclic voltammetry (2500 cycles) and 20 million biphasic pulses. Electrochemical impedance spectroscopy (EIS) and image analysis are performed to detect eventual changes of the electrodes morphology. Results demonstrate that the devices are able to undergo chemically induced oxidative stress and electrical stimulation without failing but actually improving their electrical performance until a steady state is reached. Additionally, cell viability tests are carried out to verify the noncytotoxicity of the materials, before chronically implanting them into rat models. The behavior of the GC electrodes in vivo is monitored through EIS and sensorimotor evoked potential recordings which confirm that, with GC being activated, impedance lowers and quality of recorded signal improves. Histological analysis of the brain tissue is performed and shows no sign of severe immune reaction to the implant.

Show abstract Microglia are yolk sac-derived macrophages residing in the parenchyma of brain and spinal cord, where they interact with neurons and other glial. After different conditioning paradigms and bone marrow (BM) or hematopoietic stem cell (HSC) transplantation, graft-derived cells seed the brain and persistently contribute to the parenchymal brain macrophage compartment. Here we establish that graft-derived macrophages acquire, over time, microglia characteristics, including ramified morphology, longevity, radio-resistance and clonal expansion. However, even after prolonged CNS residence, transcriptomes and chromatin accessibility landscapes of engrafted, BM-derived macrophages remain distinct from yolk sac-derived host microglia. Furthermore, engrafted BM-derived cells display discrete responses to peripheral endotoxin challenge, as compared to host microglia. In human HSC transplant recipients, engrafted cells also remain distinct from host microglia, extending our finding to clinical settings. Collectively, our data emphasize the molecular and functional heterogeneity of parenchymal brain macrophages and highlight potential clinical implications for HSC gene therapies aimed to ameliorate lysosomal storage disorders, microgliopathies or general monogenic immuno-deficiencies.

Show abstract The development of highly accelerated fMRI acquisition techniques has led to novel possibilities to monitor cerebral activity non-invasively and with unprecedented temporal resolutions. With the emergence of dynamic connectivity and its ability to provide a much richer characterization of brain function compared to static measures, fast fMRI may yet play a crucial role in tracking dynamically varying networks. In spite of the dominance of slow hemodynamic contributions to the BOLD signal, high temporal sampling rates nevertheless improve the measurement of physiological noise, yielding an exceptional sensitivity for the detection of periods of transient connectivity at time scales of a few tens of seconds. There is also evidence that relevant BOLD fluctuations are detectable at high frequencies, implying that the benefits of fast fMRI extend beyond the ability to sample nuisance confounds. Here we review the latest technological advancements that have established fast fMRI as an effective acquisition technique, as well as its current and future implications on the analysis of dynamic networks.

Show abstract Glassy carbon (GC) has high potential to serve as a biomaterial in neural applications because it is biocompatible, electrochemically inert and can be incorporated in polyimide-based implantable devices. Miniaturization and applicability of GC is, however, thought to be partially limited by its electrical conductivity. For this study, ultra-conformable polyimide-based electrocorticography (ECoG) devices with different-diameter GC electrodes were fabricated and tested in vitro and in rat models. For achieving conformability to the rat brain, polyimide was patterned in a finger-like shape and its thickness was set to 8 µm. To investigate different electrode sizes, each ECoG device was assigned electrodes with diameters of 50, 100, 200 and 300 µm. They were electrochemically characterized and subjected to 10 million biphasic pulses—for achieving a steady-state—and to X-ray photoelectron spectroscopy, for examining their elemental composition. The electrodes were then implanted epidurally to evaluate the ability of each diameter to detect neural activity. Results show that their performance at low frequencies (up to 300 Hz) depends on the distance from the signal source rather than on the electrode diameter, while at high frequencies (above 200 Hz) small electrodes have higher background noises than large ones, unless they are coated with poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS).

Show abstract Neural interfaces for neuroscientific research are nowadays mainly manufactured using standard microsystems engineering technologies which are incompatible with the integration of carbon as electrode material. In this work, we investigate a new method to fabricate graphitic carbon electrode arrays on flexible substrates. The devices were manufactured using infrared nanosecond laser technology for both patterning all components and carbonizing the electrode sites. Two laser pulse repetition frequencies were used for carbonization with the aim of finding the optimum. Prototypes of the devices were evaluated in vitro in 30 mM hydrogen peroxide to mimic the post-surgery oxidative environment. The electrodes were subjected to 10 million biphasic pulses (39.5 μC/cm2) to measure their stability under electrical stress. Their biosensing capabilities were evaluated in different concentrations of dopamine in phosphate buffered saline solution. Raman spectroscopy and x-ray photoelectron spectroscopy analysis show that the atomic percentage of graphitic carbon in the manufactured electrodes reaches the remarkable value of 75%. Results prove that the infrared nanosecond laser yields activated graphite electrodes that are conductive, non-cytotoxic and electrochemically inert. Their comprehensive assessment indicates that our laser-induced carbon electrodes are suitable for future transfer into in vivo studies, including neural recordings, stimulation and neurotransmitters detection.

Show abstract Abstract: Optogenetics opened not only new exciting opportunities to interrogate the nervous system but also requires adequate probes to facilitate these wishes. Therefore, a multidisciplinary effort is essential to match these technical opportunities with biological needs in order to establish a stable and functional material-tissue interface. This in turn can address an optical intervention of the genetically modified, light sensitive cells in the nervous system and recording of electrical signals from single cells and neuronal networks that result in behavioral changes. In this review, we present the state of the art of optoelectronic probes and assess advantages and challenges of the different design approaches. At first, we discuss mechanisms and processes at the material-tissue interface that influence the performance of optoelectronic probes in acute and chronic implantations. We classify optoelectronic probes by their property of delivering light to the tissue: by waveguides or by integrated light sources at the sites of intervention. Both approaches are discussed with respect to size, spatial resolution, opportunity to integrate electrodes for electrical recording and potential interactions with the target tissue. At last, we assess translational aspects of the state of the art. Long-term stability of probes and the opportunity to integrate them into fully implantable, wireless systems are a prerequisite for chronic applications and a transfer from fundamental neuroscientific studies into treatment options for diseases and clinical trials. Highlights: - Miniaturization technologies enable multichannel optoelectronic neural probes. - No external laser is needed when light sources are directly integrated on the probes. - Connectors to recording equipment still limit further miniaturization. - Wireless systems with integrated multiplexers and amplifiers can replace connectors. - Longevity of probes is mandatory for chronic implantation.

Show abstract Deep brain stimulation (DBS) is a successful medical therapy for many treatment resistant neuropsychiatric disorders such as movement disorders; e.g., Parkinson’s disease, Tremor, and dystonia. Moreover, DBS is becoming more and more appealing for a rapidly growing number of patients with other neuropsychiatric diseases such as depression and obsessive compulsive disorder. In spite of the promising outcomes, the current clinical hardware used in DBS does not match the technological standards of other medical applications and as a result could possibly lead to side effects such as high energy consumption and others. By implementing more advanced DBS devices, in fact, many of these limitations could be overcome. For example, a higher channels count and smaller electrode sites could allow more focal and tailored stimulation. In addition, new materials, like carbon for example, could be incorporated into the probes to enable adaptive stimulation protocols by biosensing neurotransmitters in the brain. Updating the current clinical DBS technology adequately requires combining the most recent technological advances in the field of neural engineering. Here, a novel hybrid multimodal DBS probe with glassy carbon microelectrodes on a polyimide thin-film device assembled on a silicon rubber tubing is introduced. The glassy carbon interface enables neurotransmitter detection using fast scan cyclic voltammetry and electrophysiological recordings while simultaneously performing electrical stimulation. Additionally, the presented DBS technology shows no imaging artefacts in magnetic resonance imaging. Thus, we present a promising new tool that might lead to a better fundamental understanding of the underlying mechanism of DBS while simultaneously paving our way towards better treatments.

Show abstract Stereo-electroencephalography depth electrodes, regularly implanted into drug-resistant patients with focal epilepsy to localize the epileptic focus, have a low channel count (6-12 macro- or microelectrodes), limited spatial resolution (0.5-1 cm) and large contact area of the recording sites (~mm2). Thus, they are not suited for high-density local field potential and multiunit recordings. In this paper, we evaluated the long-term electrophysiological recording performance and histocompatibility of a neural interface consisting of 32 microelectrodes providing a physical shape similar to clinical devices. The cylindrically-shaped depth probes made of polyimide (PI) were chronically implanted for 13 weeks into the brain of rats, while cortical or thalamic activity (local field potentials, single-unit and multi-unit activity) was recorded regularly to monitor the temporal change of several features of the electrophysiological performance. To examine the tissue reaction around the probe, neuron-selective and astroglia-selective immunostaining methods were applied. Stable single-unit and multi-unit activity were recorded for several weeks with the implanted depth probes and a weak or moderate tissue reaction was found around the probe track. Our data on biocompatibility presented here and in vivo experiments in non-human primates provide a strong indication that this type of neural probe can be applied in stereo-electroencephalography recordings of up to 2 weeks in humans targeting the localization of epileptic foci providing an increased spatial resolution and the ability to monitor local field potentials and neuronal spiking activity.

Show abstract Vor einigen Jahren waren bunte Aufnahmen unserer grauen Zellen überaus populär. Prominente Vertreter der Neurowissenschaften erweckten mit bildgebenden Verfahren den Eindruck, als ob unser Gehirn kurz vor der Enträtselung stünde. Auf einen solchen Durchbruch warten manche Veränderungsmanager schon sehnsüchtig. Inzwischen ist es um diesen Hype deutlich ruhiger geworden und von den einstigen Protagonisten ist nicht mehr viel zu hören. Es ist also Zeit für eine Bestandsaufnahme. Chefredakteur Martin Claßen hat mit einem der führenden deutschen Neurowissenschaftler, Stefan Rotter, über die Möglichkeiten und Grenzen seines Metiers und die Konsequenzen für das Change Management gesprochen.

Show abstract Patients suffering from neuronal degenerative diseases are increasingly being equipped with neural implants to treat symptoms or restore functions and increase their quality of life. Magnetic resonance imaging (MRI) would be the modality of choice for diagnosis and compulsory post-operative monitoring of such patients. However, interactions between the MR environment and implants pose severe health risks to the patient. Nevertheless, neural implant recipients regularly underwent MRI examinations, and adverse events were reported rarely. This should not imply that the procedures are safe. More than 300.000 cochlear implant recipients are excluded from MRI unless the indication outweighs excruciating pain. For 75.000 DBS recipients quite the opposite holds: MRI is considered essential part of the implantation procedure and some medical centres deliberately exceed safety regulations which they referred to as crucially impractical. MRI related permanent neurological dysfunctions in DBS recipients have occurred in the past when manufacturer recommendations were exceeded. Within the last decades extensive effort has been invested to identify, characterise, and quantify the occurring interactions. Today we are far from a satisfying solution to achieve a safe and beneficial MR procedure for all implant recipients. To contribute, we intend to raise awareness of a growing concern and want to summon the community to stop absurdities and instead improve the situation for the increasing number of patients. Therefore, we review implant safety in the MRI literature from an engineering point of view, with a focus on cochlear and DBS implants as success stories in clinical practice. We briefly explain fundamental phenomena which can lead to patient harm, and point out breakthroughs and errors made. We end with conclusions and strategies to avoid future implants from being contraindicated to MR examinations. We believe that implant recipients should enter MRI, but before doing so, we should make sure that the procedure is reasonable.

Show abstract Social robots can help meet the growing need for rehabilitation assistance; measures for creating and maintaining trust in human-robot interactions should be priorities when designing social robots for rehabilitation.

Show abstract In fMRI, functional connectivity is conventionally characterized by correlations between fMRI time series, which are intrinsically undirected measures of connectivity. Yet, some information about the directionality of network connections can nevertheless be extracted from the matrix of pairwise temporal correlations between all considered time series, when expressed in the frequency-domain as a cross-spectral density matrix. Using a sparsity prior, it then becomes possible to determine a unique directed network topology that best explains the observed undirected correlations, without having to rely on temporal precedence relationships that may not be valid in fMRI. Applying this method on simulated data with 100 nodes yielded excellent retrieval of the underlying directed networks under a wide variety of conditions. Importantly, the method did not depend on temporal precedence to establish directionality, thus reducing susceptibility to hemodynamic variability. The computational efficiency of the algorithm was sufficient to enable whole-brain estimations, thus circumventing the problem of missing nodes that otherwise occurs in partial-brain analyses. Applying the method to real resting-state fMRI data acquired with a high temporal resolution, the inferred networks showed good consistency with structural connectivity obtained from diffusion tractography in the same subjects. Interestingly, this agreement could also be seen when considering high-frequency rather than low-frequency connectivity (average correlation: r = 0.26 for f < 0.3 Hz, r = 0.43 for 0.3 < f < 5 Hz). Moreover, this concordance was significantly better (p<0.05) than for networks obtained with conventional functional connectivity based on correlations (average correlation r = 0.18). The presented methodology thus appears to be well-suited for fMRI, particularly given its lack of explicit dependence on temporal lag structure, and is readily applicable to whole-brain effective connectivity estimation.

Show abstract Synchronous spiking of multiple neurons is a key phenomenon in normal brain function and pathologies. Recently, approaches to record spikes from the intact cortical surface using small high-density arrays of microelectrodes have been reported. It remained unaddressed how epicortical spiking relates to intracortical unit activity. We introduced a mesoscale approach using an array of 64 electrodes with intermediate diameter (250 μm) and combined large-coverage epicortical recordings in ferrets with intracortical recordings via laminar probes. Empirical data and modelling strongly suggest that our epicortical electrodes selectively captured synchronized spiking of neurons in the cortex beneath. As a result, responses to sensory stimulation were more robust and less noisy compared to intracortical activity, and receptive field properties were well preserved in epicortical recordings. This should promote insights into assembly-coding beyond the informative value of subdural EEG or single-unit spiking, and be advantageous to real-time applications in brain-machine interfacing.

Show abstract Basal ganglia output neurons transmit motor signals by decreasing their firing rate during movement. This decrease can lead to post-inhibitory rebound spikes in thalamocortical neurons in motor thalamus (Mthal). While in healthy animals neural activity in the basal ganglia is markedly uncorrelated, in Parkinson{\textquoteright}s disease neural activity becomes pathologically correlated. Here we investigated the impact of correlations in the basal ganglia output on the transmission of motor signals to Mthal using a Hodgkin-Huxley model of a thalamocortical neuron. We found that correlations in the basal ganglia output disrupt the transmission of motor signals via rebound spikes by increasing the signal-to-noise ratio and trial-to-trial variability. We further examined the role of brief sensory responses in basal ganglia output neurons and the effect of cortical excitation of Mthal in modulating rebound spiking. Interestingly, both the sensory responses and cortical inputs could either promote or suppress the generation of rebound spikes depending on their timing relative to the motor signal. Finally, in the model rebound spiking occurred despite the presence of moderate levels of excitation, indicating that rebound spiking might be feasible in a parameter regime relevant also in vivo. Overall, our model provides novel insights into the transmission of motor signals from the basal ganglia to Mthal by suggesting new functional roles for active decorrelation and sensory responses in the basal ganglia, as well as cortical excitation of Mthal.Author summary The output of the basal ganglia might act like a brake on our brain{\textquoteright}s motor circuits such as motor thalamus. When we move, this brake is released, letting motor thalamus execute the selected movement. However, the neural processes that underlie the communication of the basal ganglia with the motor thalamus during movement are unclear. We utilise a computational model of a neuron in motor thalamus to investigate how this transmission might work, how it can be modulated by sensory and cortical inputs, and how it is compromised in Parkinson{\textquoteright}s disease. Our results explain how pathological correlations in the neural activity in Parkinson{\textquoteright}s disease disturb the transmission of motor signals, which might underlie some of the motor symptoms.

Show abstract The finding that very large networks can be trained efficiently and reliably has led to a paradigm shift in computer vision from engineered solutions to learning formulations. As a result, the research challenge shifts from devising algorithms to creating suitable and abundant training data for supervised learning. How to efficiently create such training data? The dominant data acquisition method in visual recognition is based on web data and manual annotation. Yet, for many computer vision problems, such as stereo or optical flow estimation, this approach is not feasible because humans cannot manually enter a pixel-accurate flow field. In this paper, we promote the use of synthetically generated data for the purpose of training deep networks on such tasks. We suggest multiple ways to generate such data and evaluate the influence of dataset properties on the performance and generalization properties of the resulting networks. We also demonstrate the benefit of learning schedules that use different types of data at selected stages of the training process.

Show abstract Background Studies on sleep-spindles are typically based on visual-marks performed by experts, however this process is time consuming and presents a low inter-expert agreement, causing the data to be limited in quantity and prone to bias. An automatic detector would tackle these issues by generating large amounts of objectively marked data. New Method Our goal was to develop a sensitive, precise and robust sleep-spindle detection method. Emphasis has been placed on achieving a consistent performance across heterogeneous recordings and without the need for further parameter fine tuning. The developed detector runs on a single channel and is based on multivariate classification using a support vector machine. Scalp-electroencephalogram recordings were segmented into epochs which were then characterized by a selection of relevant and non-redundant features. The training and validation data came from the Medical Center-University of Freiburg, the test data consisted of 27 records coming from 2 public databases. Results Using a sample based assessment, 53% sensitivity, 37% precision and 96% specificity was achieved on the DREAMS database. On the MASS database, 77% sensitivity, 46% precision and 96% specificity was achieved. The developed detector performed favorably when compared to previous detectors. The classification of normalized EEG epochs in a multidimensional space, as well as the use of a validation set, allowed to objectively define a single detection threshold for all databases and participants. Conclusions The use of the developed tool will allow increasing the data-size and statistical significance of research studies on the role of sleep-spindles.

Show abstract Simultaneous recordings and manipulations of neural circuits that control the behavior of animals is one of the key techniques in modern neuroscience. Rapid advances in optogenetics have led to a variety of probes combining multichannel readout and optogenetic write in. Given the complexity of the brain, it comes as no surprise that the choice of the device is constrained by several factors such as the animal model, the structure and location of the brain area of interest, as well as the behavioral read out. Here we provide an overview of available devices for chronic simultaneous neural recordings and optogenetic manipulation in awake behaving rats. We focus on two fixed arrays and two moveable drives. For both options, we present data from one custom-made (in house) and one commercially available device. Here we provide evidence that simultaneous neural recordings and optogenetic manipulations are feasible with all four tested devices. Further we give detailed information about the recording quality, and also contrast the different features of the probes. As we provide detailed information about equipment and building procedures for combined chronic multichannel readout and optogenetic control with maximum performance at minimized costs, this overview might help especially researchers who want to enter the field of in vivo optophysiology.

Show abstract Stable interconnection to neurons in vivo over long time-periods is critical for the success of future advanced neuroelectronic applications. The inevitable foreign body reaction towards implanted materials challenges the stability and an active intervention strategy would be desirable to treat inflammation locally. Here, we investigate whether controlled release of the anti-inflammatory drug Dexamethasone from flexible neural microelectrodes in the rat hippocampus has an impact on probe-tissue integration over 12 weeks of implantation. The drug was stored in a conducting polymer coating (PEDOT/Dex), selectively deposited on the electrode sites of neural probes, and released on weekly basis by applying a cyclic voltammetry signal in three electrode configuration in fully awake animals. Dex-functionalized probes provided stable recordings and impedance characteristics over the entire chronic study. Histological evaluation after 12 weeks of implantation revealed an overall low degree of inflammation around all flexible probes whereas electrodes exposed to active drug release protocols did have neurons closer to the electrode sites compared to controls. The combination of flexible probe technology with anti-inflammatory coatings accordingly offers a promising approach for enabling long-term stable neural interfaces.

Show abstract This study presents a new conducting polymer hydrogel (CPH) system, consisting of the synthetic hydrogel P(DMAA-co-5%MABP-co-2,5%SSNa) and the conducting polymer (CP) poly(3,4-ethylenedioxythiophene) (PEDOT), intended as coating material for neural interfaces. The composite material can be covalently attached to the surface electrode, can be patterned by a photolithographic process to influence selected electrode sites only and forms an interpenetrating network. The hybrid material was characterized using cyclic voltammetry (CV), impedance spectroscopy (EIS) and X-ray photoelectron spectroscopy (XPS), which confirmed a homogeneous distribution of PEDOT throughout all CPH layers. The CPH exhibited a 2,5 times higher charge storage capacity (CSC) and a reduced impedance when compared to the bare hydrogel. Electrochemical stability was proven over at least 1000 redox cycles. Non-toxicity was confirmed using an elution toxicity test together with a neuroblastoma cell-line. The described material shows great promise for surface modification of neural probes making it possible to combine the beneficial properties of the hydrogel with the excellent electronic properties necessary for high quality neural microelectrodes. STATEMENT OF SIGNIFICANCE: Conductive polymer hydrogels have emerged as a promising new class of materials to functionalize electrode surfaces for enhanced neural interfaces and drug delivery. Common weaknesses of such systems are delamination from the connection surface, and the lack of suitable patterning methods for confining the gel to the selected electrode site. Various studies have reported on conductive polymer hydrogels addressing one of these challenges. In this study we present a new composite material which offers, for the first time, the unique combination of properties: it can be covalently attached to the substrate, forms an interpenetrating network, shows excellent electrical properties and can be patterned via UV-irradiation through a structured mask.

Show abstract We present a combined report on the results of three editions of the Cell Tracking Challenge, an ongoing initiative aimed at promoting the development and objective evaluation of cell segmentation and tracking algorithms. With 21 participating algorithms and a data repository consisting of 13 data sets from various microscopy modalities, the challenge displays today's state-of-the-art methodology in the field. We analyzed the challenge results using performance measures for segmentation and tracking that rank all participating methods. We also analyzed the performance of all of the algorithms in terms of biological measures and practical usability. Although some methods scored high in all technical aspects, none obtained fully correct solutions. We found that methods that either take prior information into account using learning strategies or analyze cells in a global spatiotemporal video context performed better than other methods under the segmentation and tracking scenarios included in the challenge.

Show abstract Hebbian and homeostatic plasticity have been studied extensively in the past, both experimentally and theoretically, but many aspects of their interaction remain to be elucidated. Hebbian plasticity is thought to shape neuronal connectivity during development and learning, whereas homeostatic plasticity would stabilize network activity. Here we investigate another aspect of this interaction, which is whether Hebbian associative properties can also emerge as a network effect from a plasticity rule based on homeostatic principles on the neuronal level. The maturation of cortical networks during sensory experience is an ideal case to explore this question. Excitatory neurons in the visual cortex of rodents have been shown to connect preferentially to neurons that respond to similar visual features. Since this connectivity bias is not existent at the time of eye opening, but only after some weeks of visual experience, it has been suggested that plastic mechanisms are responsible for the changes taking place during sensory stimulation. We consider a structural plasticity rule driven by a homeostasis of firing rate in a recurrent network of leaky integrate-and-fire (LIF) neurons exposed to external input that is modulated by the orientation of a visual stimulus. Our results show that feature specific connectivity, similar to what has been experimentally observed in rodent visual cortex, can emerge out of a random balanced network of LIF neurons with a plasticity rule that is not explicitly dependent on correlations between pre- and postsynaptic neuronal activity. The synaptic association of neurons responding to similar stimulus features occurs as a side-effect of controlling the activity of individual neurons. The experience dependent structural changes that are triggered by simulation are long lasting and decay only slowly when the neurons are exposed again to non modulated external input. arXiv 1706.02912 [q-bio.NC], 2017 (pdf)

Show abstract Medical brain implants for closed-loop interaction with the cerebral cortex promise new treatment options for brain disorders, and thus great efforts are being made to develop devices for long-term application. Closed-loop interaction can be implemented using electrophysiological recording techniques, and can be used to modulate local cortical activity or long-range functional connectivity. In a case study performed in sheep chronically implanted with a novel micro-electrocorticography-based device, we show that (1) open-loop single-pulse electrical stimulation (SPES) elicited the well-known cortico-cortical evoked potentials (CCEPs), and (2) closed-loop repetitive-pulse electrical stimulation (RPES) elicited specific cortico-cortical spectral responses (CCSRs). CCSRs were spatially focalized in the gamma band, compared with beta band independent of RPES frequency. The topography of CCSRs was different compared with CCEPs, suggesting that CCEPs and CCSRs capture different aspects of cortico-cortical connectivity. We propose that CCSRs provide new useful measures of functional connectivity, and that in particular gamma-band CCSRs may be an optimal choice if spatially precise closed-loop interaction is desired. However, the parameter space of micro-electrocorticography stimulation patterns and associated changes in μECoG frequency bands needs to be further explored and many questions remain before closed-loop brain implants can be used in clinical applications.

Show abstract Wireless implants for interaction with the cortex have developed rapidly over the last decade and increasingly meet demands of clinical brain–computer interfaces. For such applications, well-established technologies are available, suitable for recording of neural activity at different spatial scales and adequate for modulating brain activity by cortical electrical stimulation. The incorporation of recording and stimulation into closed-loop systems is a major aim in active, fully implantable medical device design. To reduce clinical long-term implantation risk and to increase the spatial specificity of epicortical recordings and stimulation, micro-electrocorticography is a promising technology. However, currently there is a lack of implants suitable for chronic human clinical applications that utilize micro-electrocorticography and possess closed-loop functionality. Here, we describe the clinical importance of cortical stimulation, give an overview of existing implants that use mainly epicortical recording methods, and present results of a closed-loop micro-electrocorticography system developed for clinical application within a collaborative framework. Finally, we conclude with our vision of future design options in the field of neuroprosthetic devices.

Show abstract Background: Objective assessments of Parkinson’s disease (PD) patients’ motor state using motion capture techniques are still rarely used in clinical practice, even though they may improve clinical management. One major obstacle relates to the large dimensionality of motor abnormalities in PD. We aimed to extract global motor performance measures covering different everyday motor tasks, as a function of a clinical intervention, i.e., deep brain stimulation (DBS) of the subthalamic nucleus. Methods: We followed a data-driven, machine-learning approach and propose performance measures that employ Random Forests with probability distributions. We applied this method to 14 PD patients with DBS switched-off or -on, and 26 healthy control subjects performing the Timed Up and Go Test (TUG), the Functional Reach Test (FRT), a hand coordination task, walking 10-m straight, and a 90° curve. Results: For each motor task, a Random Forest identified a specific set of metrics that optimally separated PD off DBS from healthy subjects. We noted the highest accuracy (94.6%) for standing up. This corresponded to a sensitivity of 91.5% to detect a PD patient off DBS, and a specificity of 97.2% representing the rate of correctly identified healthy subjects. We then calculated performance measures based on these sets of metrics and applied those results to characterize symptom severity in different motor tasks. Task-specific symptom severity measures correlated significantly with each other and with the Unified Parkinson’s Disease Rating Scale (UPDRS, part III, correlation of r2 = 0.79). Agreement rates between different measures ranged from 79.8 to 89.3%. Conclusion: The close correlation of PD patients’ various motor abnormalities quantified by different, task-specific severity measures suggests that these abnormalities are only facets of the underlying one-dimensional severity of motor deficits. The identification and characterization of this underlying motor deficit may help to optimize therapeutic interventions, e.g., to “automatically” adapt DBS settings in PD patients.

Show abstract Deep learning with convolutional neural networks (deep ConvNets) has revolutionized computer vision through end-to-end learning, i.e. learning from the raw data. Now, there is increasing interest in using deep ConvNets for end-to-end EEG analysis. However, little is known about many important aspects of how to design and train ConvNets for end-to-end EEG decoding, and there is still a lack of techniques to visualize the informative EEG features the ConvNets learn. Here, we studied deep ConvNets with a range of different architectures, designed for decoding imagined or executed movements from raw EEG. Our results show that recent advances from the machine learning field, including batch normalization and exponential linear units, together with a cropped training strategy, boosted the deep ConvNets decoding performance, reaching or surpassing that of the widely-used filter bank common spatial patterns (FBCSP) decoding algorithm. While FBCSP is designed to use spectral power modulations, the features used by ConvNets are not fixed a priori. Our novel methods for visualizing the learned features demonstrated that ConvNets indeed learned to use spectral power modulations in the alpha, beta and high gamma frequencies. These methods also proved useful as a technique for spatially mapping the learned features, revealing the topography of the causal contributions of features in different frequency bands to decoding the movement classes. Our study thus shows how to design and train ConvNets to decode movement-related information from the raw EEG without handcrafted features and highlights the potential of deep ConvNets combined with advanced visualization techniques for EEG-based brain mapping.

Show abstract Gamma oscillations (30–150 Hz) in neuronal networks are associated with the processing and recall of information. We measured local field potentials in the dentate gyrus of freely moving mice and found that gamma activity occurs in bursts, which are highly heterogeneous in their spatial extensions, ranging from focal to global coherent events. Synaptic communication among perisomatic-inhibitory interneurons (PIIs) is thought to play an important role in the generation of hippocampal gamma patterns. However, how neuronal circuits can generate synchronous oscillations at different spatial scales is unknown. We analyzed paired recordings in dentate gyrus slices and show that synaptic signaling at interneuron-interneuron synapses is distance dependent. Synaptic strength declines whereas the duration of inhibitory signals increases with axonal distance among interconnected PIIs. Using neuronal network modeling, we show that distance-dependent inhibition generates multiple highly synchronous focal gamma bursts allowing the network to process complex inputs in parallel in flexibly organized neuronal centers.

Show abstract Throughout each day, the brain displays transient changes in state, as evidenced by shifts in behavior and vigilance. While the electrophysiological correlates of brain states have been studied for some time, it remains unclear how large-scale cortico-cortical functional connectivity systematically reconfigures across states. Here, we investigate state-dependent shifts in cortical functional connectivity by recording local field potentials (LFPs) during spontaneous behavioral transitions in the ferret using chronically implanted micro-electrocorticographic (µECoG) arrays positioned over occipital, parietal, and temporal cortical regions. To objectively classify brain state, we describe a data-driven approach that projects time-varying LFP spectral properties into brain state space. Distinct brain states displayed markedly different patterns of cross-frequency phase-amplitude coupling and inter-electrode phase synchronization across several LFP frequency bands. The largest across-state differences in functional connectivity were observed between periods of presumed slow-wave and rapid-eye-movement-sleep/active-state, which were characterized by the contrasting phenomena of cortical network fragmentation and global synchronization, respectively. Collectively, our data provide strong evidence that large-scale functional interactions in the brain dynamically reconfigure across behavioral states.

Show abstract We present a variational approach for surface reconstruction from a set of oriented points with scale information. We focus particularly on scenarios with non-uniform point densities due to images taken from different distances. In contrast to previous methods, we integrate the scale information in the objective and globally optimize the signed distance function of the surface on a balanced octree grid. We use a finite element discretization on the dual structure of the octree minimizing the number of variables. The tetrahedral mesh is generated efficiently from the dual structure, and also memory efficiency is optimized, such that robust data terms can be used even on very large scenes. The surface normals are explicitly optimized and used for surface extraction to improve the reconstruction at edges and corners.

Show abstract OBJECTIVE: Technology for localizing epileptogenic brain regions plays a central role in surgical planning. Recent improvements in acquisition and electrode technology have revealed that high-frequency oscillations (HFOs) within the 80-500 Hz frequency range provide the neurophysiologist with new information about the extent of the epileptogenic tissue in addition to ictal and interictal lower frequency events. Nevertheless, two decades after their discovery there remain questions about HFOs as biomarkers of epileptogenic brain and there use in clinical practice. METHODS: In this review, we provide practical, technical guidance for epileptologists and clinical researchers on recording, evaluation, and interpretation of ripples, fast ripples, and very high-frequency oscillations. RESULTS: We emphasize the importance of low noise recording to minimize artifacts. HFO analysis, either visual or with automatic detection methods, of high fidelity recordings can still be challenging because of various artifacts including muscle, movement, and filtering. Magnetoencephalography and intracranial electroencephalography (iEEG) recordings are subject to the same artifacts. SIGNIFICANCE: High-frequency oscillations are promising new biomarkers in epilepsy. This review provides interested researchers and clinicians with a review of current state of the art of recording and identification and potential challenges to clinical translation.

Show abstract Objective. Brain-computer interfaces (BCI) based on event-related potentials (ERP) incorporate a decoder to classify recorded brain signals and subsequently select a control signal that drives a computer application. Standard supervised BCI decoders require a tedious calibration procedure prior to every session. Several unsupervised classification methods have been proposed that tune the decoder during actual use and as such omit this calibration. Each of these methods has its own strengths and weaknesses. Our aim is to improve overall accuracy of ERP-based BCIs without calibration.Approach. We consider two approaches for unsupervised classification of ERP signals. Learning from label proportions (LLP) was recently shown to be guaranteed to converge to a supervised decoder when enough data is available. In contrast, the formerly proposed expectation maximization (EM) based decoding for ERP-BCI does not have this guarantee. However, while this decoder has high variance due to random initialization of its parameters, it obtains a higher accuracy faster than LLP when the initialization is good.We introduce a method to optimally combine these two unsupervised decoding methods, letting one method’s strengths compensate for the weaknesses of the other and vice versa. The new method is compared to the aforementioned methods in a resimulation of an experiment with a visual speller.Main Results. Analysis of the experimental results shows that the new method exceeds the performance of the previous unsupervised classification approaches in terms of ERP classification accuracy and symbol selection accuracy during the spelling experiment. Furthermore, the method shows less dependency on random initialization of model parameters and is consequently more reliable.Significance. Improving the accuracy and subsequent reliability of calibrationless BCIs makes these systems more appealing for frequent use.

Show abstract The relative timing of action potentials in neurons recorded from local cortical networks often shows a non-trivial dependence, which is then quantified by cross-correlation functions. Theoretical models emphasize that such spike train correlations are an inevitable consequence of two neurons being part of the same network and sharing some synaptic input. For non-linear neuron models, however, explicit correlation functions are difficult to compute analytically, and perturbative methods work only for weak shared input. In order to treat strong correlations, we suggest here an alternative non-perturbative method. Specifically, we study the case of two leaky integrate-and-fire neurons with strong shared input. Correlation functions derived from simulated spike trains fit our theoretical predictions very accurately. Using our method, we computed the non-linear correlation transfer as well as correlation functions that are asymmetric due to inhomogeneous intrinsic parameters or unequal input.

Show abstract We train generative 'up-convolutional' neural networks which are able to generate images of objects given object style, viewpoint, and color. We train the networks on rendered 3D models of chairs, tables, and cars. Our experiments show that the networks do not merely learn all images by heart, but rather find a meaningful representation of 3D models allowing them to assess the similarity of different models, interpolate between given views to generate the missing ones, extrapolate views, and invent new objects not present in the training set by recombining training instances, or even two different object classes. Moreover, we show that such generative networks can be used to find correspondences between different objects from the dataset, outperforming existing approaches on this task.

Show abstract Shakey the robot was one of the first autonomous robots that showed impressive capabilities of navigation and mobile manipulation. Since then, robotics research has made great progress, showing more and more capable robotic systems for a large variety of application domains and tasks. In this letter, we look back on decades of research by rebuilding Shakey with modern robotics technology in the open-source Shakey 2016 system. Hereby, we demonstrate the impact of research by showing that ideas from the original Shakey are still alive in state-of-the-art systems, while robotics in general has improved to deliver more robust and more capable software and hardware. Our Shakey 2016 system has been implemented on real robots and leverages mostly open-source software. We experimentally evaluate the system in real-world scenarios on a PR2 robot and a Turtlebot-based robot and particularly investigate the development effort. The experiments documented in this letter demonstrate that results from robotics research are readily available for building complex robots such as Shakey within a short amount of time and little effort.

Show abstract Pairs of neurons in brain networks often share much of the input they receive from other neurons. Due to essential nonlinearities of the neuronal dynamics, the consequences for the correlation of the output spike trains are generally not well understood. Here we analyze the case of two leaky integrate-and-fire neurons using an approach which is nonperturbative with respect to the degree of input correlation. Our treatment covers both weakly and strongly correlated dynamics, generalizing previous results based on linear response theory.

Show abstract Somatostatin-expressing-interneurons (SOMIs) in the dentate gyrus (DG) control formation of granule cell (GC) assemblies during memory acquisition. Hilar-perforant-path-associated interneurons (HIPP cells) have been considered to be synonymous for DG-SOMIs. Deviating from this assumption, we show two functionally contrasting DG-SOMI-types. The classical feedback-inhibitory HIPPs distribute axon fibers in the molecular layer. They are engaged by converging GC-inputs and provide dendritic inhibition to the DG circuitry. In contrast, SOMIs with axon in the hilus, termed hilar interneurons (HILs), provide perisomatic inhibition onto GABAergic cells in the DG and project to the medial septum. Repetitive activation of glutamatergic inputs onto HIPP cells induces long-lasting-depression (LTD) of synaptic transmission but long-term-potentiation (LTP) of synaptic signals in HIL cells. Thus, LTD in HIPPs may assist flow of spatial information from the entorhinal cortex to the DG, whereas LTP in HILs may facilitate the temporal coordination of GCs with activity patterns governed by the medial septum.

Show abstract Accurate simulations of peripheral nerve recordings are needed to develop improved neuroprostheses. Previous models of peripheral nerves contained simplifications whose effects have not been investigated. We created a novel detailed finite element (FE) model of a peripheral nerve, and used it to carry out a sensitivity analysis of several model parameters. To construct the model, in vivo recordings were obtained in a rat sciatic nerve using an 8-channel nerve cuff electrode, after which the nerve was imaged using magnetic resonance imaging (MRI). The FE model was constructed based on the MRI data, and included progressive branching of the fascicles. Neural pathways were defined in the model for the tibial, peroneal and sural fascicles. The locations of these pathways were selected so as to maximize the correlations between the simulated and in vivo recordings. The sensitivity analysis showed that varying the conductivities of neural tissues had little influence on the ability of the model to reproduce the recording patterns obtained experimentally. On the other hand, the increased anatomical detail did substantially alter the recording patterns observed, demonstrating that incorporating fascicular branching is an important consideration in models of nerve cuff recordings. The model used in this study constitutes an improved simulation tool and can be used in the design of neural interfaces.

Show abstract With the continued miniaturisation of portable embedded systems, wireless EEG recording techniques are becoming increasingly prevalent in animal behavioural research. However, in spite of their versatility and portability, they have seldom been used inside water-maze tasks designed for rats. As such, a novel 3D printed implant and waterproof connector is presented, which can facilitate wireless water-maze EEG recordings in freely-moving rats, using a commercial wireless recording system (W32; Multichannel Systems). As well as waterproofing the wireless system, battery, and electrode connector, the implant serves to reduce movement-related artefacts by redistributing movement-related forces away from the electrode connector. This implant/connector was able to successfully record high-qual- ity LFP in the hippocampo-striatal brain regions of rats as they undertook a procedural- learning variant of the double-H water-maze task. Notably, there were no significant perfor- mance deficits through its use when compared with a control group across a number of met- rics including number of errors and speed of task completion. Taken together, this method can expand the range of measurements that are currently possible in this diverse area of behavioural neuroscience, whilst paving the way for integration with more complex behaviours.

Show abstract Direct current (DC) stimulation can be used to influence the orientation and migratory behavior of cells and results in cellular electrotaxis. Experimental work on such phenomena commonly relies on electrochemical dissolution of silver:silver–chloride (Ag:AgCl) electrodes to provide the stimulation via salt bridges. The strong ionic flow can be expected to influence the cell culture environment. In order to shed more light on which effects that must be considered, and possibly counter balanced, we here characterize a typical DC stimulation system. Silver migration speed was determined by stripping voltammetry. pH variability with stimulation was measured by ratiometric image analysis and conductivity alterations were quantified via two electrode impedance spectroscopy. It could be concluded that pH shifts towards more acidic values, in a linear manner with applied charge, after the buffering capability of the culture medium is exceeded. In contrast, the influence on conductivity was of negligible magnitude. Silver ions could enter the culture chamber at low concentrations long before a clear effect on the viability of the cultured cells could be observed. A design rule of 1 cm salt bridge per C of stimulation charge transferred was however sufficient to ensure separation between cells and silver at all times.

Show abstract Abstract There has been substantial progress over the last decade towards miniaturizing implantable microelectrodes for use in Active Implantable Medical Devices (AIMD). Compared to the rapid development and complexity of electrode miniaturization, methods to monitor and assess functional integrity and electrical functionality of these electrodes, particularly during long term stimulation, have not progressed to the same extent. Evaluation methods that form the gold standard, such as stimulus pulse testing, cyclic voltammetry and electrochemical impedance spectroscopy, are either still bound to laboratory infrastructure (impractical for long term in vivo experiments) or deliver no comprehensive insight into the material’s behaviour. As there is a lack of cost effective and practical predictive measures to understand long term electrode behaviour in vivo, material investigations need to be performed after explantation of the electrodes. We propose the analysis of the electrode and its environment in situ, to better understand and correlate the effects leading to electrode failure. The derived knowledge shall eventually lead to improved electrode designs, increased electrode functionality and safety in clinical applications. In this paper, the concept, design and prototyping of a sensor framework used to analyse the electrode’s behaviour and to monitor diverse electrode failure mechanisms, even during stimulation pulses, is presented. We focused on the electronic circuitry and data acquisition techniques required for a conceptual multi-sensor system. Functionality of single modules and a prototype framework have been demonstrated, but further work is needed to convert the prototype system into an implantable device. In vitro studies will be conducted first to verify sensor performance and reliability.

Show abstract Connectome genetics seeks to uncover how genetic factors shape brain functional connectivity; however, the causal impact of a single gene's activity on whole-brain networks remains unknown. We tested whether the sole targeted deletion of the mu opioid receptor gene (Oprm1) alters the brain connectome in living mice. Hypothesis-free analysis of combined resting-state fMRI diffusion tractography showed pronounced modifications of functional connectivity with only minor changes in structural pathways. Fine-grained resting-state fMRI mapping, graph theory, and intergroup comparison revealed Oprm1-specific hubs and captured a unique Oprm1 gene-to-network signature. Strongest perturbations occurred in connectional patterns of pain/aversion-related nodes, including the mu receptor-enriched habenula node. Our data demonstrate that the main receptor for morphine predominantly shapes the so-called reward/aversion circuitry, with major influence on negative affect centers.

Show abstract OBJECTIVES: In simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI), artifacts on the EEG arise from the switching of magnetic field gradients in the MR scanner. These artifacts depend on head position, and are, therefore, difficult to remove in the presence of subject motion. In this study, gradient artifacts are modeled by multiple templates extracted from externally recorded motion information.METHODS: Gradient artifact correction was performed in EEG-fMRI recordings by estimating artifactual templates modulated by slowly varying splines, as well as head position information. The EEG signal quality was then compared following two common methods: averaged artifact subtraction (AAS) and optimal basis sets (OBS).RESULTS: Artifact correction using multiple templates estimated from splines or motion time courses outperformed the existing AAS and OBS approaches, as quantified by root-mean-square power across gradient epochs. Improvements were mostly seen in posterior EEG channels, where most of the residual artifacts are seen following the AAS and OBS methods. Residual spectral power was comparable to that of EEG signals recorded without fMRI scanning.CONCLUSION: Gradient artifacts can be well modeled by multiple templates estimated from head position information, resulting in an effective artifact removal.SIGNIFICANCE: This method can facilitate EEG-fMRI of uncooperative subjects in whom motion is inevitable, for example, to investigate high-frequency EEG activity in which gradient artifacts are particularly prominent.

Show abstract This paper describes the complete mathematical optimization process of an inductive powering system suitable for the application within implanted biomedical systems. The optimization objectives are thereby size, energy efficiency, and tissue absorption. Within the first step, the influence of the operational frequency on the given quantities is computed by means of finite element simulations, yielding a compromise of power transfer efficiency of the wireless link and acceptable tissue heating in terms of the specific absorption rate. All simulations account for the layered structure of the human head, modeling the dielectric properties with Cole-Cole dispersion effects. In the second step, the relevant coupling and loss effects of the transmission coils are modeled as a function of the geometrical design parameters, enabling a noniterative and comprehensible mathematical derivation of the optimum coil geometry given an external size constraint. Further investigations of the optimum link design also consider high-permeability structures being applied to the primary coil, enhancing the efficiency by means of an increased mutual inductance. Thereby, a final link efficiency of 80% at a coil separation distance of 5 mm and 20% at 20 mm using a 10-mm planar receiving coil can be achieved, contributing to a higher integration density of multichannel brain implanted sensors. Moreover, the given procedure does not only give insight into the optimization of the coil design, but also provides a minimized set of mathematical expressions for designing a highly efficient primary side coil driver and for selecting the components of the secondary side impedance matching. All mathematical models and descriptions have been verified by simulation and concluding measurements.

Show abstract Abstract Resting-state networks have become an important tool for the study of brain function. An ultra-fast imaging technique that allows to measure brain function, called Magnetic Resonance Encephalography (MREG), achieves an order of magnitude higher temporal resolution than standard echo-planar imaging (EPI). This new sequence helps to correct physiological artifacts and improves the sensitivity of the fMRI analysis. In this study, EPI is compared with MREG in terms of capability to extract resting-state networks. Healthy controls underwent two consecutive resting-state scans, one with EPI and the other with MREG. Subject-level independent component analyses (ICA) were performed separately for each of the two datasets. Using Stanford FIND atlas parcels as network templates, the presence of ICA maps corresponding to each network was quantified in each subject. The number of detected individual networks was significantly higher in the MREG data set than for EPI. Moreover, using short time segments of MREG data, such as 50 seconds, one can still detect and track consistent networks. Fast fMRI thus results in an increased capability to extract distinct functional regions at the individual subject level for the same scan times, and also allow the extraction of consistent networks within shorter time intervals than when using EPI, which is notably relevant for the analysis of dynamic functional connectivity fluctuations. Hum Brain Mapp 38:817–830, 2017. © 2016 Wiley Periodicals, Inc.

Show abstract An innovative micro thermoelectric generator (lTEG) fabrication process has been developed. Two selectively dissolvable photoresists and galvanostatic electrodeposition are used to grow p- and n-type thermoelectric materials as well as the upper and lower contacts of the µTEGs onto a single substrate. Two particular features ofthe process are the usage ofa multilamination technique to create structures for legs and contacts, as well as an industrial pick and placer (P&P), which allows dispensing of a second, selectively dissolvable, photoresist to protect certain areas during material deposition. This allows sequential electrochemical deposition oftwo different thermoelectric materials on a single substrate, without further costly and time-consuming process steps. The process therefore provides a highly flexible fabrication platform for research and development.

Show abstract Polyimide based shaft electrodes were coated with a bioresorbable layer to stiffen them for intracortical insertion and to reduce the mechanical mismatch between the target tissue and the implanted device after degradation of the coating. Molten saccharose was used as coating material. In a proof-of-concept study, the electrodes were implanted into the cortex of Wistar rats and the insertion forces during implantation were recorded. Electrochemical impedance spectroscopy was performed immediately after implantation and up to 13 weeks after implantation to monitor the tissue responsetotheimplantedelectrodes.Therecordedspectrawere modeled with an equivalent circuit to differentiate the influence of the single components. In one rat, a peak in the encapsulation resistance was observable after two weeks of implantation, indicating the peak of the acute inflammatory response. In another rat, the lowest resistances were observed after four weeks, indicating the termination of the acute inflammatory response. Multiunitactivity was recorded with an adequatesignaltonoiseratiotoallowspikesorting.Histology was performed after 7, 45 and 201 days of implantation. The results showed the highest tissue reaction after 45 days and confirmed impedance data that acute inflammatory reactions terminate over time.

Show abstract The objective of this work is to produce a laser- fabricated polymer-metal-polymer electrode with the merit of a carbon-based coating as the active site. A 10 μm-thick layer of parylene-C is used serving as the insulation layer in which the active site is locally laser-pyrolyzed. Our preliminary results show that the proposed method is promising in terms of fabrication feasibility and desired electrochemical capabilities.

Show abstract Over the past decades, optical technologies have entered neural implant technologies. Applications such as optogenetics, near-infrared spectroscopy (NIRS), and direct-near-infrared stimulation (NIS) request technical devices that combine electrical and optical recording as well as stimulation capabilities using light sources and/or optical sensors. Optoprobes are the technical devices that meet these requirements. This paper provides basic insights into optogenetic mechanisms, the background of NIRS and NIS, and focuses on fundamental requirements of technical systems from a biological background. The state of the art of optoprobes is reviewed and attention is drawn on the potential long-term stability of these technical devices for chronic neural implants. Further, material selection for stiff and flexible devices, applicable light sources, waveguide and coupling concepts, packaging paradigms as well as system assembly and integration aspects are discussed in view of biocompatible and biostable devices. This paper also considers the physical background of light scattering and heat generation when light sources are implanted into biological tissue.

Show abstract Conducting polymers (CPs) have frequently been described as outstanding coating materials for neural microelectrodes, providing significantly reduced impedance or higher charge injection compared to pure metals. Usability has until now, however, been limited by poor adhesion of polymers like poly(3,4-ethylenedioxythiophene) (PEDOT) to metallic substrates, ultimately precluding long-term applications. The aim of this study was to overcome this weakness of CPs by introducing two novel adhesion improvement strategies that can easily be integrated with standard microelectrode fabrication processes. Iridium Oxide (IrOx) demonstrated exceptional stability for PEDOT coatings, resulting in polymer survival over 10 000 redox cycles and 110 days under accelerated aging conditions at 60 °C. Nanostructured Pt was furthermore introduced as a purely mechanical adhesion promoter providing 10-fold adhesion improvement compared to smooth Pt substrates by simply altering the morphology of Pt. This layer can be realized in a very simple process that is compatible with any electrode design, turning nanostructured Pt into a universal adhesion layer for CP coatings. By the introduction of these adhesion-promoting strategies, the weakness of CP-based neural probes can ultimately be eliminated and true long-term stable use of PEDOT on neural probes will be possible in future electrode generations.

Show abstract Magnetic resonance imaging (MRI) was used to gain in vivo insight into load-induced displacements of inner plant tissues making a non-invasive and non-destructive stress and strain analysis possible. The central aim of this study was the identification of a possible load-adapted orientation of the vascular bundles and their fibre caps as the mechanically relevant tissue in branch-stem-attachments of Dracaena marginata. The complex three-dimensional deformations that occur during mechanical loading can be analysed on the basis of quasi-three-dimensional data representations of the outer surface, the inner tissue arrangement (meristem and vascular system), and the course of single vascular bundles within the branch-stem-attachment region. In addition, deformations of vascular bundles could be quantified manually and by using digital image correlation software. This combination of qualitative and quantitative stress and strain analysis leads to an improved understanding of the functional morphology and biomechanics of D. marginata, a plant that is used as a model organism for optimizing branched technical fibre-reinforced lightweight trusses in order to increase their load bearing capacity.

Show abstract Goal: To identify and overcome barriers to creating new neurotechnologies capable of restoring both motor and sensory function in individuals with neurological conditions. Methods: This report builds upon the outcomes of a joint workshop between the US National Science Foundation and the German Research Foundation on New Perspectives in Neuroengineering and Neurotechnology convened in Arlington, VA, USA, November 13–14, 2014. Results: The participants identified key technological challenges for recording and manipulating neural activity, decoding, and interpreting brain data in the presence of plasticity, and early considerations of ethical and social issues pertinent to the adoption of neurotechnologies. Conclusions: The envisaged progress in neuroengineering requires tightly integrated hardware and signal processing efforts, advances in understanding of physiological adaptations to closed-loop interactions with neural devices, and an open dialog with stakeholders and potential end-users of neurotechnology. Significance: The development of new neurotechnologies (e.g., bidirectional brain–computer interfaces) could significantly improve the quality of life of people living with the effects of brain or spinal cord injury, or other neurodegenerative diseases. Focused efforts aimed at overcoming the remaining barriers at the electrode tissue interface, developing implantable hardware with on-board computation, and refining stimulation methods to precisely activate neural tissue will advance both our understanding of brain function and our ability to treat currently intractable disorders of the nervous system.

Show abstract As service robots become more and more capable of performing useful tasks for us, there is a growing need to teach robots how we expect them to carry out these tasks. However, different users typically have their own preferences, for example with respect to arranging objects on different shelves. As many of these preferences depend on a variety of factors including personal taste, cultural background, or common sense, it is challenging for an expert to pre-program a robot in order to accommodate all potential users. At the same time, it is impractical for robots to constantly query users about how they should perform individual tasks. In this work, we present an approach to learn patterns in user preferences for the task of tidying up objects in containers, e.g. shelves or boxes. Our method builds upon the paradigm of collaborative filtering for making personalized recommendations and relies on data from different users which we gather using crowdsourcing. To deal with novel objects for which we have no data, we propose a method that compliments standard collaborative filtering by leveraging information mined from the Web. When solving a tidy-up task, we first predict pairwise object preferences of the user. Then, we subdivide the objects in containers by modeling a spectral clustering problem. Our solution is easy to update, does not require complex modeling, and improves with the amount of user data. We evaluate our approach using crowdsourcing data from over 1200 users and demonstrate its effectiveness for two tidy-up scenarios. Additionally, we show that a real robot can reliably predict user preferences using our approach.

Show abstract We propose a framework for building electrophysiological predictors of single-trial motor performance variations, exemplified for SVIPT, a sequential isometric force control task suitable for hand motor rehabilitation after stroke. Electroencephalogram (EEG) data of 20 subjects with mean age of 53 years was recorded prior to and during 400 trials of SVIPT. They were executed within a single session with the non-dominant left hand, while receiving continuous visual feedback of the produced force trajectories. The behavioral data showed strong trial-by-trial performance variations for five clinically relevant metrics, which accounted for reaction time as well as for the smoothness and precision of the produced force trajectory. 18 out of 20 tested subjects remained after preprocessing and entered offline analysis. Source Power Comodulation (SPoC) was applied on EEG data of a short time interval prior to the start of each SVIPT trial. For 11 subjects, SPoC revealed robust oscillatory EEG subspace components, whose bandpower activity are predictive for the performance of the upcoming trial. Since SPoC may overfit to non-informative subspaces, we propose to apply three selection criteria accounting for the meaningfulness of the features. Across all subjects, the obtained components were spread along the frequency spectrum and showed a variety of spatial activity patterns. Those containing the highest level of predictive information resided in and close to the alpha band. Their spatial patterns resemble topologies reported for visual attention processes as well as those of imagined or executed hand motor tasks. In summary, we identified subject-specific single predictors that explain up to 36% of the performance fluctuations and may serve for enhancing neuroergonomics of motor rehabilitation scenarios.

Show abstract The theory on the glymphatic convection mechanism of cerebrospinal fluid holds that cardiac pulsations in part pump cerebrospinal fluid from the peri-arterial spaces through the extracellular tissue into the peri-venous spaces facilitated by aquaporin water channels. Since cardiac pulses cannot be the sole mechanism of glymphatic propulsion, we searched for additional cerebrospinal fluid pulsations in the human brain with ultra-fast magnetic resonance encephalography. We detected three types of physiological mechanisms affecting cerebral cerebrospinal fluid pulsations: cardiac, respiratory, and very low frequency pulsations. The cardiac pulsations induce a negative magnetic resonance encephalography signal change in peri-arterial regions that extends centrifugally and covers the brain in ≈1 Hz cycles. The respiratory ≈0.3 Hz pulsations are centripetal periodical pulses that occur dominantly in peri-venous areas. The third type of pulsation was very low frequency (VLF 0.001-0.023 Hz) and low frequency (LF 0.023-0.73 Hz) waves that both propagate with unique spatiotemporal patterns. Our findings using critically sampled magnetic resonance encephalography open a new view into cerebral fluid dynamics. Since glymphatic system failure may precede protein accumulations in diseases such as Alzheimer's dementia, this methodological advance offers a novel approach to image brain fluid dynamics that potentially can enable early detection and intervention in neurodegenerative diseases.

Show abstract Highlights • A novel 3D-printed headstage was developed for protecting skull-mounted implants in rodents. • The socket allowed for successful chronic pair-housing of rats following stereotaxic surgery. • Rats were able to carry out a range of normal behaviours, with no significant implant damage observed. • This implant can help to improve the well-being of post-surgical rats, whilst reducing the cost of rodent upkeep

Show abstract The possibility to release drugs from conducting polymers, like polypyrrole or poly(3,4‐ethylenedioxythiophene) (PEDOT), has been described and investigated for a variety of different substances during the last years, showing a wide interest in these release systems. A point that has not been looked at so far however is the possibility of other substances, next to the intended ones, leaving the polymer film under the high voltage excursions during redox sweeping. In this study we target this weakness of commonly used detection methods by implementing a high precision analytical method (high‐performance liquid chromatography) that allows a separation and subsequently a detailed quantification of all possible release products. We could identify a significantly more complex release behavior for a PEDOT:Dex system than has been assumed so far, revealing the active release of the monomer upon redox activation. The released EDOT could thereby be shown to result from the bulk material, causing a considerable loss of polymer (>10% during six release events) that could partly account for the observed degradation or delamination effects of drug‐eluting coatings. The monomer leakage was found to be substantially higher for a PEDOT:Dex film compared to a PEDOT:PSS sample. This finding indicates an overestimation of drug release if side products are mistaken for the actual drug mass. Moreover the full picture of released substances implements the need for further studies to reduce the monomer leakage and identify possible adverse effects, especially in the perspective of releasing an anti‐inflammatory substance for attenuation of the foreign body reaction toward implanted electrodes.

Show abstract A typical Go/No-Go decision is suggested to be implemented in the brain via the activation of the direct or indirect pathway in the basal ganglia. Medium spiny neurons (MSNs) in the striatum, receiving input from cortex and projecting to the direct and indirect pathways express D1 and D2 type dopamine receptors, respectively. Recently, it has become clear that the two types of MSNs markedly differ in their mutual and recurrent connectivities as well as feedforward inhibition from FSIs. Therefore, to understand striatal function in action selection, it is of key importance to identify the role of the distinct connectivities within and between the two types of MSNs on the balance of their activity. Here, we used both a reduced firing rate model and numerical simulations of a spiking network model of the striatum to analyze the dynamic balance of spiking activities in D1 and D2 MSNs. We show that the asymmetric connectivity of the two types of MSNs renders the striatum into a threshold device, indicating the state of cortical input rates and correlations by the relative activity rates of D1 and D2 MSNs. Next, we describe how this striatal threshold can be effectively modulated by the activity of fast spiking interneurons, by the dopamine level, and by the activity of the GPe via pallidostriatal backprojections. We show that multiple mechanisms exist in the basal ganglia for biasing striatal output in favour of either the `Go' or the `No-Go' pathway. This new understanding of striatal network dynamics provides novel insights into the putative role of the striatum in various behavioral deficits in patients with Parkinson's disease, including increased reaction times, L-Dopa-induced dyskinesia, and deep brain stimulation-induced impulsivity.

Show abstract In the absence of sensory stimulation or motor output, the brain exhibits complex spatiotemporal patterns of intrinsically generated neural activity. Analysis of ongoing brain dynamics has identified the prevailing modes of cortico-cortical interaction; however, little is known about how such patterns of intrinsically generated activity are correlated between cortical and subcortical brain areas. We investigate the correlation structure of ongoing cortical and superior colliculus (SC) activity across multiple spatial and temporal scales. Ongoing cortico-tectal interaction was characterized by correlated fluctuations in the amplitude of delta, spindle, low gamma, and high-frequency oscillations (>100 Hz). Of these identified coupling modes, topographical patterns of high-frequency coupling were the most consistent with patterns of anatomical connectivity, reflecting synchronized spiking within cortico-tectal networks. Cortico-tectal coupling at high frequencies was temporally parcellated by the phase of slow cortical oscillations and was strongest for SC-cortex channel pairs that displayed overlapping visual spatial receptive fields. Despite displaying a high degree of spatial specificity, cortico-tectal coupling in lower-frequency bands did not match patterns of cortex-to-SC anatomical connectivity. Collectively, our findings demonstrate that neural activity is spontaneously coupled between cortex and SC, with high- and low-frequency modes of coupling reflecting direct and indirect cortico-tectal interactions, respectively.

Show abstract Extracellular, large scale in vivo recording of neural activity is mandatory for elucidating the interaction of neurons within large neural networks at the level of their single unit activity. Technological achievements in MEMS-based multichannel electrode arrays offer electrophysiological recording capabilities that go far beyond those of classical wire electrodes. Despite their impressive channel counts, recording systems with modest interconnection overhead have been demonstrated thanks to the hybrid integration of CMOS circuitry for signal preprocessing and data handling. The number of addressable channels is increased even further by a switch matrix for electrode selection co-integrated along the slender probe shafts. When realized by IC fabrication technologies, these probes offer highest recording site densities along the entire shaft length.

Show abstract A cost effective bottom-up process for the fabrication of micro thermoelectric generators (μTEGs) was developed. It is based on a novel fabrication method involving a selectively sacrificial photoresist for the sequential galvanostatic electrodeposition of thermoelectric materials. The use of an industrial pick and placer (P&P) for dispensing the second photoresist allows for accurate and flexible μTEG designs. The process makes use of Ordyl® as a negative dry film photoresist template and sequential lamination steps for shaping all thermoelectric legs and contacts. All structures of the μTEG are generated in one photoresist multi-layer - this represents the most significant advantage of the process. The process uses a minimum of clean room processing for the preparation of pre-structured substrates for electrodeposition and therefore provides a cost-effective, highly flexible fabrication platform for research and development.

Show abstract As service robots become more and more capable of performing useful tasks for us, there is a growing need to teach robots how we expect them to carry out these tasks. However, learning our preferences is a nontrivial problem, as many of them stem from a variety of factors including personal taste, cultural background, or common sense. Obviously, such factors are hard to formulate or model a priori. In this paper, we present a solution for tidying up objects in containers, e.g., shelves or boxes, by following user preferences. We learn the user preferences using collaborative filtering based on crowdsourced and mined data. First, we predict pairwise object preferences of the user. Then, we subdivide the objects in containers by modeling a spectral clustering problem. Our solution is easy to update, does not require complex modeling, and improves with the amount of user data. We evaluate our approach using crowdsoucing data from over 1,200 users and demonstrate its effectiveness for two tidy-up scenarios. Additionally, we show that a real robot can reliably predict user preferences using our approach.

Show abstract We are interested in the analysis of very large continuous-time Markov chains (CTMCs) with many distinct rates. Such models arise naturally in the context of the reliability analysis, e.g., of computer networks performability analysis, of power grids, of computer virus vulnerability, and in the study of crowd dynamics. We use abstraction techniques together with novel algorithms for the computation of bounds on the expected final and accumulated rewards in continuous-time Markov decision processes (CTMDPs). These ingredients are combined in a partly symbolic and partly explicit (symblicit) analysis approach. In particular, we circumvent the use of multi-terminal decision diagrams, because the latter do not work well if facing a large number of different rates. We demonstrate the practical applicability and efficiency of the approach on two case studies.

Show abstract Animal models are an essential testbed for new devices on their path from the bench to the patient. Po- tential impairments by brain stimulation are often investi- gated in water mazes to study spatial memory and learn- ing. Video camera based tracking systems exist to quan- tify rodent behaviour, but reflections of ambient light- ing on the water surface and artefacts due to the waves caused by the swimming animal cause errors. This often requires tweaking of algorithms and parameters, or even potentially modifying the lab setup. In the following, we provide a simple solution to alleviate these problem using a combination of region based tracking and independent multimodal background subtraction (IMBS) without having to tweak a plethora of parameters.

Show abstract Spatial learning requires the hippocampus, and the replay of spatial sequences during hippocampal sharp wave-ripple (SPW-R) events of quiet wakefulness and sleep is believed to play a crucial role. To test whether the coordination of VTA reward prediction error signals with these replayed spatial sequences could contribute to this process, we recorded from neuronal ensembles of the hippocampus and VTA as rats performed appetitive spatial tasks and subsequently slept. We found that many reward responsive (RR) VTA neurons coordinated with quiet wakefulness-associated hippocampal SPW-R events that replayed recent experience. In contrast, coordination between RR neurons and SPW-R events in subsequent slow wave sleep was diminished. Together, these results indicate distinct contributions of VTA reinforcement activity associated with hippocampal spatial replay to the processing of wake and SWS-associated spatial memory.

Show abstract In this paper, we present a novel mapping system that robustly generates highly accurate 3-D maps using an RGB-D camera. Our approach requires no further sensors or odometry. With the availability of low-cost and light-weight RGB-D sensors such as the Microsoft Kinect, our approach applies to small domestic robots such as vacuum cleaners, as well as flying robots such as quadrocopters. Furthermore, our system can also be used for free-hand reconstruction of detailed 3-D models. In addition to the system itself, we present a thorough experimental evaluation on a publicly available benchmark dataset. We analyze and discuss the influence of several parameters such as the choice of the feature descriptor, the number of visual features, and validation methods. The results of the experiments demonstrate that our system can robustly deal with challenging scenarios such as fast camera motions and feature-poor environments while being fast enough for online operation. Our system is fully available as open source and has already been widely adopted by the robotics community.

Show abstract Conducting polymer films offer a convenient route for the functionalization of implantable microelectrodes without compromising their performance as excellent recording units. A micron thick coating, deposited on the surface of a regular metallic electrode, can elute anti-inflammatory drugs for the treatment of glial scarring as well as growth factors for the support of surrounding neurons. Electro-activation of the polymer drives the release of the substance and should ideally provide a reliable method for controlling quantity and timing of release. Driving signals in the form of a constant potential (CP), a slow redox sweep or a fast pulse are all represented in literature. Few studies present such release in vivo from actual recording and stimulating microelectronic devices. It is essential to bridge the gap between studies based on release in vitro, and the intended application, which would mean release into living and highly delicate tissue. In the biological setting, signals are limited both by available electronics and by the biological safety. Driving signals must not be harmful to tissue and also not activate the tissue in an uncontrolled manner. This review aims at shedding more light on how to select appropriate driving parameters for the polymer electrodes for the in vivo setting. It brings together information regarding activation thresholds for neurons, as well as injury thresholds, and puts this into context with what is known about efficient driving of release from conducting polymer films.

Show abstract The cortex processes stimuli through a distributed network of specialized brain areas. This processing requires mechanisms that can route neuronal activity across weakly connected cortical regions. Routing models proposed thus far are either limited to propagation of spiking activity across strongly connected networks or require distinct mechanisms that create local oscillations and establish their coherence between distant cortical areas. Here, we propose a novel mechanism which explains how synchronous spiking activity propagates across weakly connected brain areas supported by oscillations. In our model, oscillatory activity unleashes network resonance that amplifies feeble synchronous signals and promotes their propagation along weak connections ("communication through resonance"). The emergence of coherent oscillations is a natural consequence of synchronous activity propagation and therefore the assumption of different mechanisms that create oscillations and provide coherence is not necessary. Moreover, the phase-locking of oscillations is a side effect of communication rather than its requirement. Finally, we show how the state of ongoing activity could affect the communication through resonance and propose that modulations of the ongoing activity state could influence information processing in distributed cortical networks.

Show abstract We demonstrate and discuss the fabrication of cross-plane microthermoelectric generators with electrochemically deposited thermoelectric materials. A new process based on two layers of photoresist is presented to deposit the legs on the final substrate. The n-type Bi 2 Te 3 , Cu, and p-type Sb x Te y are integrated into the generator. The deposition of antimony telluride is performed to the largest thickness reported to date. The influence of thermal annealing on the material properties is studied. A new flip-chip reflow soldering process with Bi 57 Sn 42 Ag 1 soldering paste is presented that allows for enhanced thermal coupling of the generators. The manufactured generators are electrically and thermally fully characterized. They generate up to 1.63 μW cm -2 K -2 , which corresponds to a maximum power density of 2434.4 μW cm -2 .

Show abstract EEG-fMRI is a unique method to combine the high temporal resolution of EEG with the high spatial resolution of MRI to study generators of intrinsic brain signals such as sleep grapho-elements or epileptic spikes. While the standard EPI sequence in fMRI experiments has a temporal resolution of around 2.5-3s a newly established fast fMRI sequence called MREG (Magnetic-Resonance-Encephalography) provides a temporal resolution of around 100ms. This technical novelty promises to improve statistics, facilitate correction of physiological artifacts and improve the understanding of epileptic networks in fMRI. The present study compares simultaneous EEG-EPI and EEG-MREG analyzing epileptic spikes to determine the yield of fast MRI in the analysis of intrinsic brain signals. Patients with frequent interictal spikes (>3/20min) underwent EEG-MREG and EEG-EPI (3T, 20min each, voxel size 3×3×3mm, EPI TR=2.61s, MREG TR=0.1s). Timings of the spikes were used in an event-related analysis to generate activation maps of t-statistics. (FMRISTAT, |t|>3.5, cluster size: 7 voxels, p<0.05 corrected). For both sequences, the amplitude and location of significant BOLD activations were compared with the spike topography. 13 patients were recorded and 33 different spike types could be analyzed. Peak T-values were significantly higher in MREG than in EPI (p<0.0001). Positive BOLD effects correlating with the spike topography were found in 8/29 spike types using the EPI and in 22/33 spikes types using the MREG sequence. Negative BOLD responses in the default mode network could be observed in 3/29 spike types with the EPI and in 19/33 with the MREG sequence. With the latter method, BOLD changes were observed even when few spikes occurred during the investigation. Simultaneous EEG-MREG thus is possible with good EEG quality and shows higher sensitivity in regard to the localization of spike-related BOLD responses than EEG-EPI. The development of new methods of analysis for this sequence such as modeling of physiological noise, temporal analysis of the BOLD signal and defining appropriate thresholds is required to fully profit from its high temporal resolution.

Show abstract Understanding the intrinsic circuit-level functional organization of the brain has benefited tremendously from the advent of resting-state fMRI (rsfMRI). In humans, resting-state functional network has been consistently mapped and its alterations have been shown to correlate with symptomatology of various neurological or psychiatric disorders. To date, deciphering the mouse brain functional connectivity (MBFC) with rsfMRI remains a largely underexplored research area, despite the plethora of human brain disorders that can be modeled in this specie. To pave the way from pre-clinical to clinical investigations we characterized here the intrinsic architecture of mouse brain functional circuitry, based on rsfMRI data acquired at 7T using the Cryoprobe technology. High-dimensional spatial group independent component analysis demonstrated fine-grained segregation of cortical and subcortical networks into functional clusters, overlapping with high specificity onto anatomical structures, down to single gray matter nuclei. These clusters, showing a high level of stability and reliability in their patterning, formed the input elements for computing the MBFC network using partial correlation and graph theory. Its topological architecture conserved the fundamental characteristics described for the human and rat brain, such as small-worldness and partitioning into functional modules. Our results additionally showed inter-modular interactions via "network hubs". Each major functional system (motor, somatosensory, limbic, visual, autonomic) was found to have representative hubs that might play an important input/output role and form a functional core for information integration. Moreover, the rostro-dorsal hippocampus formed the highest number of relevant connections with other brain areas, highlighting its importance as core structure for MBFC.

Show abstract Exchange of thoughts by means of expressive speech is fundamental to human communication. However, the neuronal basis of real-life communication in general, and of verbal exchange of ideas in particular, has rarely been studied until now. Here, our aim was to establish an approach for exploring the neuronal processes related to cognitive “idea” units (IUs) in conditions of non-experimental speech production. We investigated whether such units corresponding to single, coherent chunks of speech with syntactically-defined borders, are useful to unravel the neuronal mechanisms underlying real-world human cognition. To this aim, we employed simultaneous electrocorticography (ECoG) and video recordings obtained in pre-neurosurgical diagnostics of epilepsy patients. We transcribed non-experimental, daily hospital conversations, identified IUs in transcriptions of the patients' speech, classified the obtained IUs according to a previously-proposed taxonomy focusing on memory content, and investigated the underlying neuronal activity. In each of our three subjects, we were able to collect a large number of IUs which could be assigned to different functional IU subclasses with a high inter-rater agreement. Robust IU-onset-related changes in spectral magnitude could be observed in high gamma frequencies (70–150 Hz) on the inferior lateral convexity and in the superior temporal cortex regardless of the IU content. A comparison of the topography of these responses with mouth motor and speech areas identified by electrocortical stimulation showed that IUs might be of use for extraoperative mapping of eloquent cortex (average sensitivity: 44.4%, average specificity: 91.1%). High gamma responses specific to memory-related IU subclasses were observed in the inferior parietal and prefrontal regions. IU-based analysis of ECoG recordings during non-experimental communication thus elicits topographically- and functionally-specific effects. We conclude that segmentation of spontaneous real-world speech in linguistically-motivated units is a promising strategy for elucidating the neuronal basis of mental processing during non-experimental communication.

Show abstract Activation detection in functional Magnetic Resonance Imaging (fMRI) typically assumes the hemodynamic response to neuronal activity to be invariant across brain regions and subjects. Reports of substantial variability of the morphology of blood-oxygenation-level-dependent (BOLD) responses are accumulating, suggesting that the use of a single generic model of the expected response in general linear model (GLM) analyses does not provide optimal sensitivity due to model misspecification. Relaxing assumptions of the model can limit the impact of hemodynamic response function (HRF) variability, but at a cost on model parsimony. Alternatively, better specification of the model could be obtained from a priori knowledge of the HRF of a given subject, but the effectiveness of this approach has only been tested on simulation data. Using fast BOLD fMRI, we characterized the variability of hemodynamic responses to a simple event-related auditory-motor task, as well as its effect on activation detection with GLM analyses. We show the variability to be higher between subjects than between regions and variation in different regions to correlate from one subject to the other. Accounting for subject-related variability by deriving subject-specific models from responses to the task in some regions lead to more sensitive detection of responses in other regions. We applied the approach to epilepsy patients, where task-derived patient-specific models provided additional information compared to the use of a generic model for the detection of BOLD responses to epileptiform activity identified on scalp electro-encephalogram (EEG). This work highlights the importance of improving the accuracy of the model for detecting neuronal activation with fMRI, and the fact that it can be done at no cost to model parsimony through the acquisition of independent a priori information about the hemodynamic response.

Show abstract Objective. Most BCIs have to undergo a calibration session in which data is recorded to train decoders with machine learning. Only recently zero-training methods have become a subject of study. This work proposes a probabilistic framework for BCI applications which exploit event-related potentials (ERPs). For the example of a visual P300 speller we show how the framework harvests the structure suitable to solve the decoding task by (a) transfer learning, (b) unsupervised adaptation, (c) language model and (d) dynamic stopping. Approach. A simulation study compares the proposed probabilistic zero framework (using transfer learning and task structure) to a state-of-the-art supervised model on n = 22 subjects. The individual influence of the involved components (a)–(d) are investigated. Main results. Without any need for a calibration session, the probabilistic zero-training framework with inter-subject transfer learning shows excellent performance—competitive to a state-of-the-art supervised method using calibration. Its decoding quality is carried mainly by the effect of transfer learning in combination with continuous unsupervised adaptation. Significance. A high-performing zero-training BCI is within reach for one of the most popular BCI paradigms: ERP spelling. Recording calibration data for a supervised BCI would require valuable time which is lost for spelling. The time spent on calibration would allow a novel user to spell 29 symbols with our unsupervised approach. It could be of use for various clinical and non-clinical ERP-applications of BCI.

Show abstract Brain-Computer Interfaces (BCIs) strive to decode brain signals into control commands for severely handicapped people with no means of muscular control. These potential users of noninvasive BCIs display a large range of physical and mental conditions. Prior studies have shown the general applicability of BCI with patients, with the conflict of either using many training sessions or studying only moderately restricted patients. We present a BCI system designed to establish external control for severely motor-impaired patients within a very short time. Within only six experimental sessions, three out of four patients were able to gain significant control over the BCI, which was based on motor imagery or attempted execution. For the most affected patient, we found evidence that the BCI could outperform the best assistive technology (AT) of the patient in terms of control accuracy, reaction time and information transfer rate. We credit this success to the applied user-centered design approach and to a highly flexible technical setup. State-of-the art machine learning methods allowed the exploitation and combination of multiple relevant features contained in the EEG, which rapidly enabled the patients to gain substantial BCI control. Thus, we could show the feasibility of a flexible and tailorable BCI application in severely disabled users. This can be considered a significant success for two reasons: Firstly, the results were obtained within a short period of time, matching the tight clinical requirements. Secondly, the participating patients showed, compared to most other studies, very severe communication deficits. They were dependent on everyday use of AT and two patients were in a locked-in state. For the most affected patient a reliable communication was rarely possible with existing AT.

Show abstract Hand loss is a highly disabling event that markedly affects the quality of life. To achieve a close to natural replacement for the lost hand, the user should be provided with the rich sensations that we naturally perceive when grasping or manipulating an object. Ideal bidirectional hand prostheses should involve both a reliable decoding of the user’s intentions and the delivery of nearly “natural” sensory feedback through remnant afferent pathways, simultaneously and in real time. However, current hand prostheses fail to achieve these requirements, particularly because they lack any sensory feedback. We show that by stimulating the median and ulnar nerve fascicles using transversal multichannel intrafascicular electrodes, according to the information provided by the artificial sensors from a hand prosthesis, physiologically appropriate (near-natural) sensory information can be provided to an amputee during the real-time decoding of different grasping tasks to control a dexterous hand prosthesis. This feedback enabled the participant to effectively modulate the grasping force of the prosthesis with no visual or auditory feedback. Three different force levels were distinguished and consistently used by the subject. The results also demonstrate that a high complexity of perception can be obtained, allowing the subject to identify the stiffness and shape of three different objects by exploiting different characteristics of the elicited sensations. This approach could improve the efficacy and “life-like” quality of hand prostheses, resulting in a keystone strategy for the near-natural replacement of missing hands.

Show abstract Objective. EEG artifacts of non-neural origin can be separated from neural signals by independent component analysis (ICA). It is unclear (1) how robustly recently proposed artifact classifiers transfer to novel users, novel paradigms or changed electrode setups, and (2) how artifact cleaning by a machine learning classifier impacts the performance of brain–computer interfaces (BCIs). Approach . Addressing (1), the robustness of different strategies with respect to the transfer between paradigms and electrode setups of a recently proposed classifier is investigated on offline data from 35 users and 3 EEG paradigms, which contain 6303 expert-labeled components from two ICA and preprocessing variants. Addressing (2), the effect of artifact removal on single-trial BCI classification is estimated on BCI trials from 101 users and 3 paradigms. Main results . We show that (1) the proposed artifact classifier generalizes to completely different EEG paradigms. To obtain similar results under massively reduced electrode setups, a proposed novel str

Show abstract Despite several approaches to realize subject-to-subject transfer of pre-trained classifiers, the full performance of a Brain-Computer Interface (BCI) for a novel user can only be reached by presenting the BCI system with data from the novel user. In typical state-of-the-art BCI systems with a supervised classifier, the labeled data is collected during a calibration recording, in which the user is asked to perform a specific task. Based on the known labels of this recording, the BCI’s classifier can learn to decode the individual’s brain signals. Unfortunately, this calibration recording consumes valuable time. Furthermore, it is unproductive with respect to the final BCI application, e.g. text entry. Therefore, the calibration period must be reduced to a minimum, which is especially important for patients with a limited concentration ability. The main contribution of this manuscript is an online study on unsupervised learning in an auditory event-related potential (ERP) paradigm. Our results demonstrate that the calibration recording can be bypassed by utilizing an unsupervised trained classifier, that is initialized randomly and updated during usage. Initially, the unsupervised classifier tends to make decoding mistakes, as the classifier might not have seen enough data to build a reliable model. Using a constant re-analysis of the previously spelled symbols, these initially misspelled symbols can be rectified posthoc when the classifier has learned to decode the signals. We compare the spelling performance of our unsupervised approach and of the unsupervised posthoc approach to the standard supervised calibration-based dogma for n = 10 healthy users. To assess the learning behavior of our approach, it is unsupervised trained from scratch three times per user. Even with the relatively low SNR of an auditory ERP paradigm, the results show that after a limited number of trials (30 trials), the unsupervised approach performs comparably to a classic supervised model.

Show abstract Mit bildhaften Vergleichen versuchen Philosophen und Wissenschaftler seit der Antike, die Arbeits­weise des menschlichen Gehirns zu beschreiben. Diese Metaphern sind Kinder ihrer jeweiligen Zeit. Sie spiegeln den aktuellen Stand der Technik wider und prägen somit die Vorstellung vom menschlichen Geist. Die Begriffsschablonen können helfen, die Komplexität des Gehirns besser zu verstehen. Indem sie eine bestimmte Eigenschaft hervorheben, unter­schlagen sie allerdings andere Aspekte, die für das Verständnis ebenso wichtig sein könnten.

Show abstract In optogenetics, neurons are genetically modified to become sensitive to light and thus, they can be stimulated or inhibited by light of certain wavelengths. In this work, we describe the fabrication of a polymer-based shaft electrode as a tool for optogenetics. This device can conduct light as well as fluids to a target brain region and record electrical neural signals from the same part of the tissue simultaneously. It is intended to facilitate optogenetic in vivo experiments with those novel multimodal neural probes or polymer optrodes. We used microsystems technology to integrate an SU-8 based waveguide and fluidic channel into a polyimide-based electrode shaft to allow simultaneous optical stimulation, fluid delivery, and electrophysiological recording in awake behaving animals. In a first acute proof-of-concept experiment in genetically modified mice, our device recorded single unit activity that was modulated by laser light transmitted into the tissue via the integrated waveguide.

Show abstract Neuronal networks confront researchers with an overwhelming complexity of interactions between their elements. A common approach to understanding neuronal processing is to reduce complexity by defining subunits and infer their functional role by selectively modulating them. However, this seemingly straightforward approach may lead to confusing results if the network exhibits parallel pathways leading to recurrent connectivity. We demonstrate limits of the selective modulation approach and argue that, even though highly successful in some instances, the approach fails in networks with complex connectivity. We argue to refine experimental techniques by carefully considering the structural features of the neuronal networks involved. Such methods could dramatically increase the effectiveness of selective modulation and may lead to a mechanistic understanding of principles underlying brain function.

Show abstract A major challenge in neuroscience is to accurately decipher in vivo the entire brain circuitry (connectome) at a microscopic level. Currently, the only methodology providing a global noninvasive window into structural brain connectivity is diffusion tractography. The extent to which the reconstructed pathways reflect realistic neuronal networks depends, however, on data acquisition and postprocessing factors. Through a unique combination of approaches, we designed and evaluated herein a framework for reliable fiber tracking and mapping of the living mouse brain connectome. One important wiring scheme, connecting gray matter regions and passing fiber-crossing areas, was closely examined: the lemniscal thalamocortical (TC) pathway. We quantitatively validated the TC projections inferred from in vivo tractography with correlative histological axonal tracing in the same wild-type and reeler mutant mice. We demonstrated noninvasively that changes in patterning of the cortical sheet, such as highly disorganized cortical lamination in reeler, led to spectacular compensatory remodeling of the TC pathway.

Show abstract In many species, spatial navigation is supported by a network of place cells that exhibit increased firing whenever an animal is in a certain region of an environment. Does this neural representation of location form part of the spatiotemporal context into which episodic memories are encoded? We recorded medial temporal lobe neuronal activity as epilepsy patients performed a hybrid spatial and episodic memory task. We identified place-responsive cells active during virtual navigation and then asked whether the same cells activated during the subsequent recall of navigation-related memories without actual navigation. Place-responsive cell activity was reinstated during episodic memory retrieval. Neuronal firing during the retrieval of each memory was similar to the activity that represented the locations in the environment where the memory was initially encoded.

Show abstract Can we infer the function of a biological neural network (BNN) if we know the connectivity and activity of all its constituent neurons?This question is at the core of neuroscience and, accordingly, various methods have been developed to record the activity and connectivity of as many neurons as possible. Surprisingly, there is no theoretical or computational demonstration that neuronal activity and connectivity are indeed sufficient to infer the function of a BNN. Therefore, we pose the Neural Systems Identification and Prediction Challenge (nuSPIC). We provide the connectivity and activity of all neurons and invite participants (1) to infer the functions implemented (hard-wired) in spiking neural networks (SNNs) by stimulating and recording the activity of neurons and, (2) to implement predefined mathematical/biological functions using SNNs. The nuSPICs can be accessed via a web-interface to the NEST simulator and the user is not required to know any specific programming language. Furthermore, the nuSPICs can be used as a teaching tool. Finally, nuSPICs use the crowd-sourcing model to address scientific issues. With this computational approach we aim to identify which functions can be inferred by systematic recordings of neuronal activity and connectivity. In addition, nuSPICs will help the design and application of new experimental paradigms based on the structure of the SNN and the presumed function which is to be discovered.

Show abstract Variable responses of neuronal networks to repeated sensory or electrical stimuli reflect the interaction of the stimulus' response with ongoing activity in the brain and its modulation by adaptive mechanisms, such as cognitive context, network state, or cellular excitability and synaptic transmission capability. Here, we focus on reliability, length, delays, and variability of evoked responses with respect to their spatial distribution, interaction with spontaneous activity in the networks, and the contribution of GABAergic inhibition. We identified network-intrinsic principles that underlie the formation and modulation of spontaneous activity and stimulus-response relations with the use of state-dependent stimulation in generic neuronal networks in vitro. The duration of spontaneously recurring network-wide bursts of spikes was best predicted by the length of the preceding interval. Length, delay, and structure of responses to identical stimuli systematically depended on stimulus timing and distance to the stimulation site, which were described by a set of simple functions of spontaneous activity. Response length at proximal recording sites increased with the duration of prestimulus inactivity and was best described by a saturation function y(t) = A(1 − e−αt). Concomitantly, the delays of polysynaptic late responses at distant sites followed an exponential decay y(t) = Be−βt + C. In addition, the speed of propagation was determined by the overall state of the network at the moment of stimulation. Disinhibition increased the number of spikes/network burst and interburst interval length at unchanged gross firing rate, whereas the response modulation by the duration of prestimulus inactivity was preserved. Our data suggest a process of network depression during bursts and subsequent recovery that limit evoked responses following distinct rules. We discuss short-term synaptic depression due to depletion of neurotransmitter vesicles as an underlying mechanism. The seemingly unreliable patterns of spontaneous activity and stimulus-response relations thus follow a predictable structure determined by the interdependencies of network structures and activity states. electrical stimulation of nervous tissue is used increasingly in the treatment of central nervous system disorders, e.g., by deep brain stimulation, in neuroprosthetic devices aiding sensory perception, as well as in examining the biophysiological properties of single cells and the function of neuronal networks. Whereas the reproducible responses of directly stimulated individual neurons consist of precisely timed single action potentials (APs) or trains of APs under specific conditions (Bryant and Segundo 1976; Mainen and Sejnowski 1995; but see Gal et al. 2010), stimulation in recurrent networks elicits multiphasic responses. These typically consist of: 1) a fast excitatory component of precise and reliable firing by antidromic or monosynaptic activation of local neurons with delays between 2 and 20 ms, 2) a transition phase with low activity thought to be mediated by inhibitory neurons, and 3) a delayed excitatory component driven by recurrent polysynaptic activation (Butovas and Schwarz 2003; Eccles et al. 1974; Fanselow and Nicolelis 1999; Rowland and Jaeger 2008; Wagenaar et al. 2004). More physiological sensory responses induced by, e.g., foot tapping, whisker deflection, or air puffs unfold comparable dynamics in various brain regions of awake and anesthetized animals (Cody et al. 1981; Fanselow and Nicolelis 1999; Rowland and Jaeger 2005). Across stimulation trials, however, the variability of timing and duration of the response, as well as the number and distribution of neurons involved, is typically high (Azouz and Gray 1999; Jones et al. 2007). On short time scales, temporal nonstationarities by modulation of activity and excitability tend to prevail and modify response amplitude, latency, and spatial spread (Kisley and Gerstein 1999; Petersen et al. 2003). In addition, the activity state of the neocortex at stimulus onset may dominate trial-by-trial variability (Arieli et al. 1996; Hasenstaub et al. 2007). Although these influences are known in a general sense, they have not been assessed quantitatively, and the prediction of stimulation outcomes for individual stimuli given during autonomous network dynamics is unreliable. Predictable stimulus-response relations, however, become increasingly important to control adequate functionality of neurotechnological devices, which seems incompatible with trial-by-trial variability. To identify the general rules of the interaction between ongoing activity and stimulus-response relations, independent of a specific tissue architecture, function, or sensory stimulus, we analyzed spontaneous and evoked activity dynamics in generic neuronal networks in vitro. These networks exhibit spontaneous spiking with typical patterns and long-term modulation (Wagenaar et al. 2006) that defines the network state with which different stimuli may interact. Here, we focus on reliability, length, delays, and variability of evoked responses with respect to their spatial distribution, interaction with spontaneous activity in the networks, and the contribution of GABAergic inhibition. We asked which interactions arise between weak, local electrical pulses and network state and how they influence evoked responses. Specifically, we developed quantitative models that show how the state of the network at the moment of stimulation determines response length and delay. Closed-loop stimulation relative to ongoing activity significantly reduced trial-by-trial variability and enabled us to examine systematically the influence of network inhibition and short-term plasticity on stimulus-response relations.

Show abstract Theoretische Neurowissenschaftler versuchen, Hirnerkrankungen wie Morbus Parkinson am Computer zu modellieren. Ein von Freiburger Ingenieuren und Mathematikern ersonnenes Computermodell kann die bei Parkinsonpatienten auftretenden Aktivitätsschwankungen in bestimmten Hirnarealen nachbilden. Das Modell sagt vorher, dass eine tiefe Hirnstimulation die krankhaften Oszillationen auch mit deutlich weniger Reizen unterdrücken könnte.

Show abstract Choosing a suitable classifier for a given dataset is an important part of developing a pattern recognition system. Since a large variety of classification algorithms are proposed in literature, non-experts do not know which method should be used in order to obtain good classification results on their data. Meta-learning tries to address this problem by recommending promising classifiers based on meta-features computed from a given dataset. In this paper, we empirically evaluate five different categories of state-of-the-art meta-features for their suitability in predicting classification accuracies of several widely used classifiers (including Support Vector Machines, Neural Networks, Random Forests, Decision Trees, and Logistic Regression). Based on the evaluation results, we have developed the first open source meta-learning system that is capable of accurately predicting accuracies of target classifiers. The user provides a dataset as input and gets an automatically created high-performance ready-to-use pattern recognition system in a few simple steps. A user study of the system with non-experts showed that the users were able to develop more accurate pattern recognition systems in significantly less development time when using our system as compared to using a state-of-the-art data mining software.

Show abstract A central motif in neuronal networks is convergence, linking several input neurons to one target neuron. In visual cortex, convergence renders target neurons responsive to complex stimuli. Yet, convergence typically sends multiple stimuli to a target, and the behaviorally relevant stimulus must be selected. We used two stimuli, activating separate electrocorticographic V1 sites, and both activating an electrocorticographic V4 site equally strongly. When one of those stimuli activated one V1 site, it gamma synchronized (60-80 Hz) to V4. When the two stimuli activated two V1 sites, primarily the relevant one gamma synchronized to V4. Frequency bands of gamma activities showed substantial overlap containing the band of interareal coherence. The relevant V1 site had its gamma peak frequency 2-3 Hz higher than the irrelevant V1 site and 4-6 Hz higher than V4. Gamma-mediated interareal influences were predominantly directed from V1 to V4. We propose that selective synchronization renders relevant input effective, thereby modulating effective connectivity.

Show abstract The vertebrate brain comprises a plethora of cell types connected by intertwined pathways. Optogenetics enriches the neuroscientific tool set for disentangling these neuronal circuits in a manner which exceeds the spatio-temporal precision of previously existing techniques. Technically, optogenetics can be divided in three types of optical and genetic combinations: (1) it is primarily understood as the manipulation of the activity of genetically modified cells (typically neurons) with light, i.e. optical actuators. (2) A second combination refers to visualizing the activity of genetically modified cells (again typically neurons), i.e. optical sensors. (3) A completely different interpretation of optogenetics refers to the light activated expression of a genetically induced construct. Here, we focus on the first two types of optogenetics, i.e. the optical actuators and sensors in an attempt to give an overview into the topic. We first cover methods to express opsins into neurons and introduce strategies of targeting specific neuronal populations in different animal species. We then summarize combinations of optogenetics with behavioral read out and neuronal imaging. Finally, we give an overview of the current state-of-the-art and an outlook on future perspectives.

Show abstract Emerging technological developments in nano- and microsystems engineering have delivered powerful tools for neuroscience research over the last 50 years. However, only a few neural implants have been transferred into clinical practice. Challenges and opportunities for translational research are discussed herein.

Show abstract The widespread use of conducting polymers, especially poly(3,4-ethylene dioxythiophene) (PEDOT), within the space of bioelectronics has enabled improvements, both in terms of electrochemistry and functional versatility, of conventional metallic electrodes. This short review aims to provide an overview of how PEDOT coatings have contributed to functionalizing existing bioelectronics, the challenges which meet conducting polymer coatings from a regulatory and stability point of view and the possibilities to bring PEDOT-based coatings into large-scale clinical applications. Finally, their potential use for enabling new technologies for the field of bioelectronics as biodegradable, stretchable and slow-stimulation materials will be discussed.

Show abstract Over the past decades, optical technologies have entered neural implant technologies. Applications such as optogenetics, near-infrared spectroscopy (NIRS), and direct-near-infrared stimulation (NIS) request technical devices that combine electrical and optical recording as well as stimulation capabilities using light sources and/or optical sensors. Optoprobes are the technical devices that meet these requirements. This paper provides basic insights into optogenetic mechanisms, the background of NIRS and NIS, and focuses on fundamental requirements of technical systems from a biological background. The state of the art of optoprobes is reviewed and attention is drawn on the potential long-term stability of these technical devices for chronic neural implants. Further, material selection for stiff and flexible devices, applicable light sources, waveguide and coupling concepts, packaging paradigms as well as system assembly and integration aspects are discussed in view of biocompatible and biostable devices. This paper also considers the physical background of light scattering and heat generation when light sources are implanted into biological tissue.

Show abstract Neural prostheses are technical systems that interface nerves to treat the symptoms of neurological diseases and to restore sensory of motor functions of the body. Success stories have been written with the cochlear implant to restore hearing, with spinal cord stimulators to treat chronic pain as well as urge incontinence, and with deep brain stimulators in patients suffering from Parkinson's disease. Highly complex neural implants for novel medical applications can be miniaturized either by means of precision mechanics technologies using known and established materials for electrodes, cables, and hermetic packages or by applying microsystems technologies. Examples for both approaches will be introduced and discussed. Electrode arrays for recording of electrocorticograms during presurgical epilepsy diagnosis have been manufactured using approved materials and a marking laser to achieve an integration density that is adequate in the context of brain machine interfaces, e.g. on the motor cortex. Microtechnologies have to be used for further miniaturization to develop polymer-based flexible and light weighted electrode arrays to interface the peripheral and central nervous system. Polyimide as substrate and insulation material will be discussed as well as several application examples for nerve interfaces like cuffs, filament like electrodes and large arrays for subdural implantation.

Show abstract Two big trends are leading the way to a new generation of thin-film electrocorticography (ECoG) micro electrode arrays (MEAs): miniaturization, which combines higher electrode densities with thinner substrates for conformability purposes; and the pursuit to extend the recording frequency band to 1 kHz and beyond (recording of spikes). When combining these two trends, however, the frequency-dependent phenomenon of crosstalk emerges as a possible setback to the so desired spatial selectivity. In this work, high in vivo coherences at 1 kHz between electrodes with neighboring tracks are reported when using the MuSA (Multi-Species Array) as recording ECoG MEA on rats. These results suggest a high degree of crosstalk between closely routed electrodes, even if placed far apart on the array, and are corroborated by coherence plots of control recordings in vitro in phosphate buffered saline (PBS). As means to estimate the combined leakage resistance and capacitance between the signal lines and the targeted brain tissue, an impedance spatial sweep over the 32 tracks routing the MuSA electrodes is performed in PBS in a two-electrode electrochemical impedance spectroscopy setup. This study should raise awareness of crosstalk as an important aspect to consider when aiming for high-quality, high-density and high-frequency neural recordings.

Show abstract Micro-electrode arrays for electrocorticography (ECoG) represent the best compromise between invasiveness and signal quality, as they are surface devices that still allow high sensitivity recordings. In this work, an assessment of different technical aspects determining the ultimate performance of ultra-conformable polyimide-based μECoG arrays is conducted via a finite element model, impedance spectroscopy measurements and recordings of sensorimotor evoked potentials (SEPs) in rats. The finite element model proves that conformability of thin-film arrays can be achieved with polyimide, a non-stretchable material, by adjusting its thickness according to the curvature of the targeted anatomical area. From the electrochemical characterization of the devices, intrinsic thermal noise of platinum and gold electrodes is estimated to be 3–5 μV. Results show that electrode size and in vitro impedance do not influence the amplitude of the recorded SEPs. However, the use of a reference on-skull (a metal screw), as compared to reference on-array (a metal electrode surrounding the recording area), provides higher-amplitude SEPs. Additionally, the incorporation of a grounded metal shield in the thin-film devices limits crosstalk between tracks and does not compromise the recording capabilities of the arrays.

Show abstract Intercranial EEG (icEEG) electrodes are implanted for pre-surgical assessment of cortical electrical activity in patients with epilepsy. MRI helps to localize the implant with respect to the individual’s anatomy and fMRI can improve understanding of the neuropathology. However, the magnetic susceptibility artifacts caused by the metal components of commercially available implants produce MRI artifacts that compromise the results especially in the direct vicinity of the implants. Next generation “thin-film” implants which feature 100x less metal thickness can mitigate these artifacts, but thin-film implants produce inconspicuous MRI signal voids in many clinical MRI sequences. Imperatively, physicians need to know the implant position and the value of EEG increases with the precision of electrode localization. Therefore, we investigated various concentrations of super paramagnetic iron oxide (SPIO) nanoparticles (NP) to label thin-film implants for localization in MRI using a range of sequences. In particular, we aim to create a marker that conspicuously renders the implant, enables spatial localization of the individual electrodes, and keeps disruptive imaging artifacts small.

Show abstract The lack of sensory feedback during grasping is a very important limitation of current hand prostheses, which affects their everyday usability. In the last years several research groups have demonstrated that nerve stimulation by implantable peripheral nerve interfaces can be reliably used to restore sensory feedback to upper limb amputees. They have shown that direct neural stimulation of peripheral nerves can effectively provide tactile information to the amputees, controlling the sensation intensity by modulating either the amplitude or the frequency of the injected stimuli. However, efforts are still necessary to identify encoding strategies converting tactile information into neural stimulation patterns capable of eliciting percepts that are both felt as natural and effective for prosthesis control. In this study, we compared the naturalness and efficacy of a set of encoding strategies based on biomimetic (model-driven) frequency modulation, amplitude modulation, or combinations of both. Such strategies were used to deliver neural stimulation to a trans-radial amputee implanted with intraneural electrodes (TIMEs). Frequency modulation was based on a biomimetic model (TouchSim) able to reproduce nerve activation patterns of the multifaceted mechanics of the skin and mechano-transduction. It was perceived as more natural, while amplitude modulation enabled better performance in tasks requiring fine identification of the applied force. Notably, hybrid encoding strategies involving both amplitude and frequency modulation were able to convey at least as much information as the amplitude modulation (for the completion of tasks), and were perceived at least as natural as the frequency modulation. The hybrid strategies improved the gross manual dexterity of the subjects during functional tasks while maintaining high manual accuracy. They also improved the level of prosthesis embodiment and reduced abnormal body perceptions of the phantom limb (“telescoping”). Encoding strategies based on the combination of biomimetic frequency modulation and amplitude modulation are able to provide highly sensitive and natural percepts and should be preferred in bidirectional prosthesis use.

Show abstract Treatments of neurological disorders utilizing active implantable devices which interact with the activity of the brain are demonstrating increasingly promising advances. Next to the continuous improvement of established therapies for movement disorders, Epilepsy and chronic pain, new therapies for depression, paralysis and many more are under investigation. The current technology available for the development of these treatments is derived from the first active implants, the cardiac pacemakers, developed in mid to late 20th century: battery powered devices with few channels and limited intelligence. The Brain Interchange technology, developed in a joint effort by the University of Freiburg and CorTec, is a new system, enabling battery-free, intelligent closed-loop applications with up to 32 channels in its first version. The implantable part, including a novel hermetic encapsulation, custom electronics and firmware, were presented last year. The electrode technology will be presented in a companion poster. Here we present the progress in the development of the external parts and the software of the system and discuss potential applications. The first version of the external parts of the system manages power and communication with the implant, as well as the software for controlling the system. Physically, these parts consist of a head piece, a relay- and a controller unit. The main functions, accessible via a graphical user interface, are: managing recording and stimulation, measuring impedance and reading out humidity, temperature, supply voltage and unique ID of the implant. A filter pipeline for signal processing and feature extraction, suitable for computationally demanding closed-loop algorithms, and the control of external devices with minimal latency has also been implemented. For future collaboration partners programmable interfaces for C++, Matlab and Python are available. The development has been done under ISO-13485 and according to EN 62304. Potential applications range from closed-loop stimulation for the treatment of Parkinson’s disease or Epilepsy to the control of assistive technology in chronic paralysis or for rehabilitation purposes. Further applications could lie in closed-loop treatments in the peripheral nervous system.

Show abstract Increasing the accessibility of autonomous robots also for inexperienced users requires user-friendly and highlevel control opportunities of robotic systems. While automated planning is able to decompose a complex task into a sequence of steps which reaches an intended goal, it is difficult to formulate such a goal without knowing the internals of the planning system and the exact capabilities of the robot. This becomes even more important in dynamic environments in which manipulable objects are subject to change. In this paper, we present an adaptive control interface which allows users to specify goals based on an internal world model by incrementally building referring expressions to the objects in the world. We consider fetch-and-carry tasks and automatically deduce potential high-level goals from the world model to make them available to the user. Based on its perceptions our system can react to changes in the environment by adapting the goal formulation within the domain-independent planning system.

Show abstract Abstract-Implantable, closed-loop devices for automated early detection and stimulation of epileptic seizures are promising treatment options for patients with severe epilepsy that cannot be treated with traditional means. Most approaches for early seizure detection in the literature are, however, not optimized for implementation on ultra-low power microcontrollers required for long-term implantation. In this paper we present a convolutional neural network for the early detection of seizures from intracranial EEG signals, designed specifically for this purpose. In addition, we investigate approximations to comply with hardware limits while preserving accuracy. We compare our approach to three previously proposed convolutional neural networks and a feature-based SVM classifier with respect to detection accuracy, latency and computational needs. Evaluation is based on a comprehensive database with long-term EEG recordings. Key ResultThe proposed method outperforms the other detectors with a median sensitivity of 0.96, false detection rate of 10.1 per hour and median detection delay of 3.7 seconds, while being the only approach suited to be realized on a low power microcontroller due to its parsimonious use of computational and memory resources.

Show abstract An innovative fabrication process of glass waveguides on silicon substrates for miniaturized implants is presented. Thin glass was bonded on oxidized silicon wafers and patterned using wet etching. Multimode waveguides with different shapes and a low surface roughness as well as low scattering of light were successfully fabricated. For efficient coupling of light and accurate alignment, KOH-grooves were etched in the silicon with respect to the glass waveguides to attach optical fibers from external light sources. Towards higher biostability, several coating and cladding materials were evaluated in accelerated in vitro tests in 60°C PBS. TiO2, SiC polyimide, Parylene C and SU-8 showed a very stable optical transmittance after 320 days in accelerated aging while PECVD Si3N4 showed significant changes within the first days.

Show abstract The high complexity of the biological response to implanted materials builds a serious barrier against implanted recording and stimulation electrode arrays to succeed in clini- cally relevant chronic studies. Some of the cell and molecular inte ractions and the ir contribution to inflammation and de vice failure are s till uncle ar. The inte rre late d me chanis ms le ading to tissue damage and electrode array failure during simultaneous faradaic, electrochemical reactions and biological response under electrical stimulation are not understood sufficiently. One variable, with which inflammatory and electrode surface processes can be analyzed and assessed, is the pH change in the immediate environment of the material-tissue interface. Here, the greatest challenges are in the biocompatibility and in-vivo long-term stability of selected sensor materials, the measure- ment of small transient pH oscillations and positioning of the sensor at a defined and nearest possible distance in the mi- crometer range, to the site of activity without the pH sensing being affected by the material-tissue interactions itself. This work represents the in-situ measurement of local and transient pH changes at a pulsed electrode with an embedded in-vivo compatible pH sensor and therein differentiating from current approaches of pH sensing during electrical stimulation.

Show abstract In dynamic environments, learned controllers are supposed to take motion into account when selecting the action to be taken. However, in existing reinforcement learning works motion is rarely treated explicitly; it is rather assumed that the controller learns the necessary motion representation from temporal stacks of frames implicitly. In this paper, we show that for continuous control tasks learning an explicit representation of motion clearly improves the quality of the learned controller in dynamic scenarios. We demonstrate this on common benchmark tasks (Walker, Swimmer, Hopper), on target reaching and ball catching tasks with simulated robotic arms, and on a dynamic single ball juggling task. Moreover, we find that when equipped with an appropriate network architecture, the agent can, on some tasks, learn motion features also with pure reinforcement learning, without additional supervision.

Show abstract Cervical spinal cord injury (SCI) and stroke severely impact grasping movements required for activities of daily living. Intraneural peripheral nerve stimulation enables specific activation of passing fibers. This paradigm has restored precise leg movements in animal models of SCI and selective sensation in human amputees. Intraneural peripheral nerve stimulation may also restore fine grasping in paralyzed hands, but this possibility has not been investigated. Here, we assess the feasibility of using intrafascicular electrical stimulation of peripheral nerves to produce precise hand movements in the non-human primate (NHP). We first extensively characterized the branching points of the median, radial, and ulnar nerves to their target muscles in the adult macaca fascicularis in order to identify the optimal implantation site for intraneural electrodes. We then reconstructed the tridimensional structure of the identified portion of each nerve in order to analyze the fascicular organization within the nerve at this level. Additionally, we assessed the distribution of motor fibers within the fascicles using immunohistochemistry. The obtained data was used to build realistic computational models of intraneural peripheral nerve stimulation for each nerve. The simulations confirmed the advantages of using intrafascicular electrodes to induce precise hand movements, demonstrating the possibility to selectively recruit patches of motor fibers without a priori knowledge of the electrode placement. We validated these results during electrophysiology experiments using transverse intrafascicular multichannel electrodes (TIMEs) implanted in the nerves of anesthetized NHPs. The stimulation achieved a selective recruitment of wrist and finger flexors and extensors in a reproducible manner across animals. We exploited these results to determine stimulation sequences that aimed at reproducing the muscle activation patterns underlying different grasping movements. For this, we mapped the obtained muscle recruitment maps to continuous EMG recordings during behavioral experiments and automatically optimized the selection of stimulation channels for each phase of the movement such as to reproduce the desired EMG pattern. Taken together, these results suggest encouraging evidences for the usability of the TIMEs to restore fine hand control after paralysis.

Show abstract Brain implants are increasingly used in neuroscientific research and medical applications. The requirements for such implants are diverse due to different experimental paradigms, scientific problem to address and demands by the researcher or clinician. To overcome these requirements, brain implants that can be built in a customized fashion might be beneficial. We previously introduced such a research grade wireless brain implant, developed exclusively using off-the-shelf components, which allows for quick and customized assembly. To verify the operability of the device during recording and stimulation, we present neurophysiological data obtained in an ovine animal model. During general anesthesia, the 63-channel µECoG electrode array was placed on the cortex of the sheep brain and both auditory and electrical stimuli were used to evoke neurophysiological responses which were recorded at a sampling rate of 4 kHz. Experiments performed in this study were conducted according to EU Directive 2010/63/EU and approved by the Animal Committee of the University of Freiburg and the Regierungspraesidium Freiburg, Germany. We show that our off-the-shelf research grade brain implant is capable to reliably record neurophysiological brain activity and electrically stimulate, evoking cortico-cortical responses, under in vivo conditions. The obtained neurophysiological activity showed clear responses with distinct spatio-temporal patterns to both auditory stimuli and cortical electrical stimulation, the latter with response patterns systematically depending on the exact stimulation site. In addition, spectral analysis revealed a neurophysiological frequency profile of the recorded activity which is in agreement with the well-known frequency power-law, i.e., the frequency-dependent linear decrease of log-log absolute spectral power. The presented neurophysiological data are in agreement with previously published recording and stimulation results obtained in sheep using different implantable devices that were not off-the-shelf. The results and their neurobiological implications as presented here highlight that off-the-shelf component brain interfacing devices are feasible and thus open up new avenues for implant-based research, especially when flexible requirements have to be addressed, as it is often the case in basic neuroscientific research. In the future we want to expand our approach to higher channel counts (128-channel high-resolution µECoG electrode arrays), higher recording sampling rate, and functional systems beyond the auditory system, as further use cases of our implant concept in neurotechnological research.

Show abstract As research becomes more and more data intensive, manag-ing this data becomes a major challenge in any organization.At university level there is seldom a unified data managementsystem in place. The general approach to storing data in suchenvironments is to deploy network storage. Each member canstore their data organized to their own likings in their ded-icated location on the network. Additionally, users tend tostore data in distributed manner such as on private devices,portable storage, or public and private repositories. Addingto this complexity, it is common for university departmentsto have high fluctuation of staff, resulting in major loss ofinformation and data on an employee’s departure. A com-mon scenario then is that it is known that certain data hasalready been created via experiments or simulation. However,it can not be retrieved, resulting in a repetition of generation,which is costly and time-consuming. Additionally, as of re-cent years, publishers and funding agencies insist on storing,sharing, and reusing existing research data. We show howdigital preservation can help group leaders and their employ-ees cope with these issues, by introducing our own archivalsystem OntoRAIS.

Show abstract This work presents reliability investigations of silicone gasket as solid underfill for interconnection interfaces in hybrid implant systems with high channel count flexible electrode arrays and hermetically packed electronics. The gasket is fabricated by laser structuring thin sheet of silicone rubber. The surface activation of silicone sheet ensures mechanical bonds with the mating surfaces thereby improving the mechanical stability of the assembly and the insulation of the interconnects. The gasket samples with 10 × 10 openings for interconnect pads, each with diameter of 270 μm and a center to center pitch size of 490 μm, were sandwiched between a polyimide array and a metallized ceramic substrate. The gasket maintained high insulation impedance of 15 ± 0.30 MΩ between the adjacent interconnects with markedly capacitive behavior (phase angle, -89 °) after 17 weeks in soaked conditions under accelerated aging at 60 °C. The gasket also survived electrical stresses and sustained high impedance (10.93 MΩ with phase angle of -88 °) when subjected to constant 3 VDC for 100 days.

Show abstract Alternating current stimulation (ACS) provides a versatile tool for modulating brain activity and presents a promising strategy for the treatment of neurological disorders like Parkinson’s disease or epilepsy. Stimulation of neural tissue at low-frequency however poses new challenges on conventional electrode materials which support limited charge transfer in the desired frequency range, from less than 0.1 Hz to several tens of Hz. In our study we address this challenge by investigating the charge transfer properties of PEDOT/PSS coatings for low-frequency applications, focusing on the impact of the polymer bulk. PEDOT films of various thicknesses were exposed to low-frequency as well as DC stimulation in vitro and compared to Pt and IrOx electrodes as controls. The charge injection performance of the metallic substrates could be substantially improved already by a thin PEDOT coating. Additionally a linear dependency between charge injection and polymer thickness suggests that PEDOT coatings are promising as materials for future ACS applications.

Show abstract Leg amputation destroys the communication between brain and environment during walking. Leg amputees rely on practically inexistent and often uncomfortable haptic feedback from the stump-socket interaction to monitor ground and obstacles contact, climb stairs, or walk in challenging environments. The lack of sensory feedback causes specific impairments to subjects that do not perceive the prosthesis as part of their body and risk falls, have decreased mobility, increased cognitive burden during walking resulting in prosthesis abandonment. In hand amputees, to restore the bidirectional communication, nerve interfaces have directly linked sensors readout from robotic hands to direct stimulation of nerves above injury. We believe that this strategy could also restore sensations from missing legs, with many scientific and technological barriers to overcome. It has never been proved that the electrical stimulation of the leg nerves by implantable neural interfaces can induce reliable sensations from missing leg and foot. In this work we developed a leg prosthesis restoring sensory feedback by means of direct nerve stimulation injected through transversal intraneural electrodes (TIME) implanted in the sciatic nerve. The stimulation was driven by the readout of pressure sensors placed under the prosthetic foot, and an encoder embedded in the prosthetic knee. We assessed the capability of 3 transfemoral amputees to recognize, blindfolded and acoustically insulated, the location where the prosthetic foot was touched and the degree of flexion of the prosthetic knee. The subjects were asked to recognize only touch, only flexion and then both conditions at the same time. We compared the performance when sensory feedback was restored and when no nerve stimulation was delivered. We found that single and double condition-tasks were executed on average respectively with a success rate of about 85% and 75%, when sensory feedback was provided. Without nerve stimulation the average success rate dropped to 25%.

Show abstract The design of microelectrodes for electrical stimulation or recording of neuronal activity increasingly focuses on biocompatibility. Still, implantation of a cortical probe evokes a sustained sterile inflammatory response (SIR) within brain tissue. This manifests in activation of microglia, astrocytic scarring, neurodegeneration, and other molecular and morphological changes that may ultimately lead to failure of the device. To visualize the extent of the spatially limited tissue response, immunohistochemistry (IHC) is the well-established method of choice. But while qualitative assessment of resulting microscopic images can be easily accomplished, quantitative analysis poses a larger challenge to the investigator. This is partly due to the large number of images required to adequately represent the SIR, but also accounted for by high variability within and across tissue sections and animals. To this point there have been several publications that employed custom-built codes for quantification of IHC signals at the implantation sites of cortical electrodes. However, these differ greatly with respect to their evaluation methods and are not provided for open access. Hence, it is difficult to compare the extent of the SIR across different or even identical electrode types utilized by different groups. We therefore aimed at creating a semi-automated quantification method that allows fast and flexible delineation of the signal-intensity distribution from the site of lesion towards healthy brain tissue, based on the widely used software tools ImageJ and MATLAB. In this study cortical implantation of lithographically fabricated, flexible polyimide probes (10µm thickness with iridium oxide electrodes) was performed on adult Sprague Dawley rats. After different survival times transversal tissue sections were immunohistochemically stained and imaged with an epifluorescence microscope. Subsequent signal quantification was performed in concentric distance bins with flexibly adaptable size (the shape of the latter is determined by the respective cavity shape of the individual image). Processing artifacts can optionally be selected and excluded from the analysis. Apart from that normalization parameters can be modified towards the needs of the study design. This also yields the possibility of using differently processed and/or imaged tissue sections within one experimental population. Our approach provides a novel, straightforward tool for quantification of IHC-images that is versatile and adaptable for various purposes.

Show abstract In 1997, deep brain stimulation (DBS) was approved by the FDA for treatment of essential tremor. In the following decades neuromodulation of the CNS became a active field and was applied for treating different conditions. Similar to the technological progress of cardiac pacemakers, concepts were developed to adapt the stimulation to the patient's need, making the devices responsive. Today, two of these closed-loop devices are approved for clinical use, Medtronic Activa PC+S and Neuropace RNS. Both devices work with eight electrode contacts on the surface or deep inside the brain and permit delivery of electrical stimuli initiated, or modified in intensity, based on neural recordings. Here, we present a closed-loop device that overcomes current application limitation by increasing the electrode contact number, minimizing the closed-loop response time and transferring the closed-loop algorithms to a device outside the body, allowing maximum freedom for clinical research. The design is inspired by today's cochlear implants: The implant is wirelessly powered by a body-external transceiver. Cortical electrode arrays and DBS electrodes can be connected to the hermetically packaged implanted electronics. The device records synchronously from 32 electrode contacts at 1kS/s (16bit) at a pass band of 0.5 to 450Hz. Data are wirelessly streamed to the body-external transceiver, which is connected to a laptop-PC, running the control software. The software can send instructions to the implant to generate electrical stimuli of up to 6mA on each of the 32 electrode contacts. Typically, it takes some 10ms for closing the loop of recording and recording-based stimulation, strongly depended on the signal analysis and decision-taking algorithms used. The system was implanted in sheep (approved by the Regierungspraesidium Freiburg, Germany and the Animal Ethics Committee of the University of Freiburg) to investigate long-term functionality and biological acceptance. Excellent robustness of the implanted hardware, good biological acceptance and stable recording signal quality could be demonstrated. We present the latest results from the animal studies and technical improvements developed based on prior results. In conclusion, the implant system presented has the potential for researching closed-loop therapies for the central nervous system. The validations towards clearance for clinical studies are currently on the way.

Show abstract The variety of "`ready-to-use'" implantable recording and stimulation systems commercially available for neuroscience is very limited and fabrication of custom made implants is commonly considered expensive and time consuming. We present a circuit design that allows cost efficient and fast translation of available components into fully wireless implants. As demonstration fully wireless implantable bidirectional neural interfaces are presented which are made of commercial off-the-shelf components (COTS) only. It is demonstrated that they are competitive to currently available state-of-the-art systems regarding size and performance.

Show abstract Decoding the human brain state with BCI methods can be seen as a building block for human-machine interaction, providing a noisy but objective, low-latency information channel including human reactions to the environment. Specifically in the context of autonomous driving, human judgement is relevant in high-level scene understanding. Despite advances in computer vision and scene understanding, it is still challenging to go from the detection of traffic events to the detection of hazards. We present a preliminary study on hazard perception, implemented in the context of natural driving videos. These have been augmented with artificial events to create potentially hazardous driving situations. We decode brain signals from electroencephalography (EEG) in order to classify single events into hazardous and non-hazardous ones. We find that event-related responses can be discriminated and the classification of events yields an AUC of 0.79. We see these results as a step towards incorporating EEG feedback into more complex, real-world tasks.

Show abstract We apply convolutional neural networks (ConvNets) to the task of distinguishing pathological from normal EEG recordings in the Temple University Hospital EEG Abnormal Corpus. We use two basic, shallow and deep ConvNet architectures recently shown to decode task-related information from EEG at least as well as established algorithms designed for this purpose. In decoding EEG pathology, both ConvNets reached substantially better accuracies (about 6% better, ~85% vs. ~79%) than the only published result for this dataset, and were still better when using only 1 minute of each recording for training and only six seconds of each recording for testing. We used automated methods to optimize architectural hyperparameters and found intriguingly different ConvNet architectures, e.g., with max pooling as the only nonlinearity. Visualizations of the ConvNet decoding behavior showed that they used spectral power changes in the delta (0-4 Hz) and theta (4-8 Hz) frequency range, possibly alongside other features, consistent with expectations derived from spectral analysis of the EEG data and from the textual medical reports. Analysis of the textual medical reports also highlighted the potential for accuracy increases by integrating contextual information, such as the age of subjects. In summary, the ConvNets and visualization techniques used in this study constitute a next step towards clinically useful automated EEG diagnosis and establish a new baseline for future work on this topic.

Show abstract In this paper we consider the problem of robot navigation in simple maze-like environments where the robot has to rely on its onboard sensors to perform the navigation task. In particular, we are interested in solutions to this problem that do not require localization, mapping or planning. Additionally, we require that our solution can quickly adapt to new situations (e.g., changing navigation goals and environments). To meet these criteria we frame this problem as a sequence of related reinforcement learning tasks. We propose a successor-feature-based deep reinforcement learning algorithm that can learn to transfer knowledge from previously mastered navigation tasks to new problem instances. Our algorithm substantially decreases the required learning time after the first task instance has been solved, which makes it easily adaptable to changing environments. We validate our method in both simulated and real robot experiments with a Robotino and compare it to a set of baseline methods including classical planning-based navigation.

Show abstract Maintaining the insulation between adjacent electrical interconnections is critical for the success of active implantable medical device. Underfilling and globtop coating of dense arrays of microfabricated interconnects pose a big reliability risk. Contamination, voids and the difficulty to deposit liquid underfillers with the appropriate adhesive behavior remains a major hurdle when developing miniaturized high-channel neural interfaces. We approach a bottom up process to fabricate adhesion promoting, graded interfaces on the bottom and top side of polyimide-based microelectrode arrays. This allows the use of pre-molded silicone rubber gaskets as dry underfill material and silicone rubber as globtop material. In this work we present the layer deposition approach to solve the difficulties of providing a double-sided, inversely oriented layer stack for an adhering silicon-oxide termination on polyimide substrates. By introducing a sacrificial polyimide layer, we permit high temperature depositions of the required layers allowing a release of the fabricated stack at the desired interface. Long term stable silicone rubber underfill and overcoat is thus achievable despite the use of polyimide substrates. The fabricated samples showed better adhesion to silicone rubber even after storing in phosphate buffered saline (PBS) at 85 °C for 18 hours and at 60°C for 72 hours. The Fourier Transform infrared (FTIR) spectrum also revealed the integrity of the structural stack after detachment from the release layer. The fabrication of double side layer stacks increases the confidence in long term stability of interconnects in polyimide electrodes.

Show abstract Long-term stability of neural interfaces is a challenge that has still to be overcome. In this study, we manufactured a highly stable multi-layer thin-film class of carbon-based devices for electrocorticography (ECoG) incorporating silicon carbide (SiC) and amorphous carbon (DLC) as adhesion promoters between glassy carbon (GC) electrodes and polyimide (PI) substrate and between PI and platinum (Pt) traces. We aged the thin-film electrodes in 30 mM H2O2 at 39 °C for one week - to mimic the effects of post-surgery inflammatory reaction - and subsequently stressed them with 2500 CV cycles. We additionally performed stability tests stimulating the electrodes with 15 million biphasic pulses. Finally, we implanted the electrodes for 6 weeks into rat models and optically characterized the explanted devices. Results show that the fabricated ECoG devices were able to withstand the in vitro and in vivo tests without significant change in impedance and morphology.

Show abstract In the field of neural prostheses, much attention has lately been given to the long-term performance not only of the electronic components but also of the parts directly interfacing with the nervous system. Neural interfaces have, in fact, a critical role in chronic applications, where they have to outlast the highly humid and oxidative body environment without undergoing delamination, corroding and without losing their functionality over time. Among all, carbon has been proved to be the material with the highest potential to contemporary serve as biomaterial for recording nerve cells activity, electrically stimulating them and, in addition, for selectively detecting the presence of neurotransmitters or other electrically active biomolecules. However, the feasibility of the fabrication method - with respect to process complexity and cost - is a factor of great importance and it is not always easy to accomplish with carbon electrodes. In this work, we present a new method to manufacture thin-film microelectrode arrays (MEAs) with laser-induced carbon active sites made from parylene c coatings on platinum iridium tracks. Such MEAs are manufactured without the need for cleanroom and MEMS processes. Prototypes of these carbon electrodes were evaluated first in vitro in hydrogen peroxide to mimic the post-surgery oxidative environment due to the acute inflammatory reaction to the implant. Electrodes were stimulated using biphasic pulses to prove their stability under electrical stress and testes with respect to their biosensing capabilities on different concentrations of dopamine in PBS. Results show that our laser-induced carbon electrodes do not deteriorate under chemical and electrochemical loads. They were able to detect different dopamine levels in vitro. These new laser-induced carbon electrodes show promising potential to successfully be implanted in vivo and be used for long-term neural applications for recording, stimulation and biochemical sensing.

Show abstract The technology for miniaturization of bioelectronics is making great progress, and the interest for high density electrode recordings in the neural systems is continually increasing. High density recording of, e.g., the cortical activity could help scientists to elucidate the language of the brain and further increase our understanding of the behavior of the cells in the neural system. Furthermore, a high density electrode implant would also increase the possibility to choose specific electrodes, e.g. in the proximity of the desired neural target, or active sites instead of silent. Also, the ability to map larger surface areas with high density arrays could help growing the understanding for the connectivity between different regions in the brain. In our group we have developed several structures based on the polymer polyimide, which is a flexible, stable and biocompatible polymer, therefore excellent for neural probes, especially for long-term applications with high demands on reliability. Designs have been targeted to animal models of turtles, rats, ferrets, cats and macaque monkeys. Modular, finger-based designs adapted well to the two-dimensionally curved structure of the brain surface even though the substrate material itself is not stretchable. Electrode size and pitch have been adapted to the size of the target structures. Array variations comprised 36, 64, 96, 192 and 252 electrode sites. Several high density ECoG arrays have been fabricated and implanted into primates. These showed good long-term stability and both single-unit activity, as well as multi-unit activity and local field potentials could be recorded via platinum and iridium oxide thin-film metal sites. Signal-to-noise ratios were sufficiently high over months and degraded only slowly. Results on stability and functionality are promising and consistent with other translational studies on peripheral nerves.

Show abstract Miniaturization of electrodes is a prerequisite of selective and targeted interaction with single neurons, enabling more applications in the continuously growing field of neuroprostheses. Miniaturization in all three dimensions of the electrical contact sites should maintain or increase longevity and electrical functionality. The thin-film metallization of the electrode site, which is only a couple of hundreds of nanometers thick, has to withstand high chemical load through the corrosive environment in the body and the electrochemical processes during electrical stimulation in vivo. Platinum (Pt), which is known to be chemically inert and mechanical stable as bulk material shows a lack of chemical and mechanical integrity applied in thin-film microelectrodes. In our study we investigated failure mechanisms of thin-film Pt electrodes under conditions of electrode aging and electrical stimulation in different physiological media. To understand and eventually overcome stability loss, we investigated the intrinsic structural stress and deformations that arose from mechanical loading through chemical impact and electrical stimulation using optical microscopy and white-light interferometry. Electrochemical measurements indicated oxidation and surface roughening as two of the degradation processes in thin-film electrodes. From the results presumptions about the underlying microstructural changes were made.

Show abstract We study motion planning problems where agents move inside environments that are not fully observable and subject to uncertainties. The goal is to compute a strategy for an agent that is guaranteed to satisfy certain safety and performance specifications. Such problems are naturally modeled by partially observable Markov decision processes (POMDPs). Because of the potentially huge or even infinite belief space of POMDPs, verification and strategy synthesis is in general computationally intractable. We tackle this difficulty by exploiting typical structural properties of such scenarios; for instance, we assume that agents have the ability to observe their own positions inside an evironment. Ambiguity in the state of the environment is abstracted into non-deterministic choices over the possible states of the environment. Technically, this abstraction transforms POMDPs into probabilistic two-player games (PGs). For these PGs, efficient verification tools are able to determine strategies that approximate certain measures on the POMDP. If an approximation is too coarse to provide guarantees, an abstraction refinement scheme further resolves the belief space of the POMDP. We demonstrate that our method improves the state of the art by orders of magnitude compared to a direct solution of the POMDP.

Show abstract Nanostructured materials exhibit large electrochemical surface areas and are thus of high interest for neural interfaces where low impedance and high charge transfer characteristics are desired. While progress in nanotechnology successively enabled smaller feature sizes and thus improved electrochemical properties, concerns were raised with respect to the mechanical stability of such nano structures for use in neural applications. In our study we address these concerns by investigating the mechanical and electrochemical stability of nanostructured platinum. Neural probes with nano-Pt were exposed to exaggerated stress tests resembling insertion into neural tissue over 60 mm distance or long-term stimulation over 240 M biphasic current pulses. Thereby only insignificant changes in electrochemical properties and morphological appearance could be observed in response to the test, proving that nanostructured platinum exhibits outstanding stability. With this finding, a major concern in using nanostructured materials for interfacing neural tissue could be eliminated, demonstrating the high potential of nanostructured platinum for neuroprosthetic devices.

Show abstract The combination of implantable neural electrodes and fMRI holds great potential for better understanding the human brain. However, the image acquisition - especially in the vicinity of the implants - is compromised by artifacts caused by metal components. In this work we address this issue by studying different types of devices in terms of designs and materials, and by quantifying their MRI artifacts. Doing so we demonstrate the quasi artifact-free behavior of a hybrid probe combining surface and penetrating carbon electrodes into a single sheet of polyimide, after comparing it with conventional implants in high field MRI and clinical fMRI.

Show abstract The loss of a limb permanently disrupts daily living activities. Prosthetic devices are an alternative to partially circumvent this disability. The lack of sensory feedback of current prosthetic options limits their acceptance and usability rate. In the upper limb, somatosensory percepts are essential for proper object manipulation, while in the lower counterparts proprioceptive and cutaneous sensations are required to maintain balance and stable gait. Restoral of sensory feedback also improves the sense of embodiment of prosthetic limbs, which positively impacts user satisfaction. The introduction of surgically targeted muscle reinnervation (TMR) led to promising outcomes in terms of controllability of prosthetic devices, and in providing a novel channel to restore sensory information. TMR consists on re-routing the remaining peripheral nerves from the amputee’s stump to the chest area. After transferring the nerves into the chest, afferent and efferent fibers reinnervate the hosting muscles, amplifying the signals from the efferent pathways, and providing a more selective channel for activating afferent fibers. We have previously demonstrated that electrical stimulation of peripheral nerves using implanted intrafascicular electrodes elicits natural sensory feedback during real-time, bidirectional control of a prosthetic hand. In this study, we propose that after TMR, the afferent reinnervated fibers can be electrically stimulated to restore natural somatosensations. We further propose that this stimulation can be provided wirelessly by capacitive coupling through the chest skin, eliminating the need for percutaneous cables and implanted electronics. An electrode implanted in the inner side of the skin picks up a portion of the current flowing inside bodily tissue when two surface electrodes located in the vicinity of the reinnervated muscles are electrically stimulated. Practically, this implanted electrode is electrically connected to the fibers reinnervated in the chest muscles such to transfer the collected current and depolarize the surrounding nerve axons. We simulate different stimulation paradigms of electrical currents travelling through the skin to the sensory fibers. Activation outputs of reinnervated afferent fibers of the peripheral nerves are then analyzed using a hybrid model of the electrical field generated by the stimulation. The outcomes of this proof-of-concept prototype as well as the implication of this novel technique for restoring natural sensory feedback in the amputee are discussed.

Show abstract Continuous control of high-dimensional systems can be achieved by current state-of-the-art reinforcement learning methods such as the Deep Deterministic Policy Gradient algorithm, but needs a significant amount of data samples. For real-world systems, this can be an obstacle since excessive data collection can be expensive, tedious or lead to physical damage. The main incentive of this work is to keep the advantages of model-free Q-learning while minimizing real-world interaction by the employment of a dynamics model learned in parallel. To counteract adverse effects of imaginary rollouts with an inaccurate model, a notion of uncertainty is introduced, to make use of artificial data only in cases of high uncertainty. We evaluate our approach on three simulated robot tasks and achieve faster learning by at least 40 per cent in comparison to vanilla DDPG with multiple updates.

Show abstract After the development of a single-sided fabrication process for intrafascicular parylene C based electrode arrays tests showed that an increase in integration density can only be achieved by a double-side process. The process uses 25 μm thick platinum iridium foil, which is thinned down with the laser and sandwiched between two 10 μm thick parylene C layers. Utilizing a picosecond laser (355 nm Nd:YVO4) it was possible to fabricate 40 μm thick electrodes that can be implanted directly in the nerve without relying on additional support layers like chitosan or silk. The fabricated samples feature three 80 μm diameter electrodes on each side and a large ground electrode that is opened to both sides. Impedance mismatches from front to back side as a result of the fabrication process are compensated by electrochemical deposition of nanostructured platinum. This step makes it possible to bring the impedances of the small electrodes down to the range of just a few kΩ at 1 kHz and illustrate the additionally gained surface due to the picosecond laser ablation on the front side electrodes. The safely injectable charge per pulse was found to be 635.75 μC/cm2 for such coated electrodes. Optical investigations show that this fabrication process offers an alternative to established lithographic processes for thin and flexible electrode arrays in neural implants.

Show abstract The objective characterization of human motion is required in a variety of fields including competitive sports, rehabilitation and the detection of motor deficits. Nowadays, typically human experts evaluate the motor behavior. These evaluations are based on their individual experience which leads to a low inter- and intra-expert reliability. Standardized tests improve on the reliability but are still prone to subjective ratings and require human expert knowledge. This paper presents a novel method to characterize the motor state of Parkinson patients using full body motion capturing data based on a combination of multiple metrics. Our approach merges various metrics with a Random Forest and uses a probabilistic formulation to compute a one-dimensional measure for the performed motion. We present an application of our approach to the problem of relating subject motion to different classes like healthy subjects and Parkinson disease patients with deep brain stimulation switched on or off. In the experimental session we show that our measure leads to high classification rates and high entropy values for real-world data. Besides, we show that our method discriminates between Parkinson’s subjects (with and without stimulation) and healthy persons as good as the Unified Parkinson’s Disease Rating Scale (UPDRS).

Show abstract Robotic assistants have the potential to greatly improve our quality of life by supporting us in our daily activities. A service robot acting autonomously in an indoor environment is faced with very complex tasks. Consider the problem of pouring a liquid into a cup, the robot should first determine if the cup is empty or partially filled. RGB-D cameras provide noisy depth measurements which depend on the opaqueness and refraction index of the liquid. In this paper, we present a novel probabilistic approach for estimating the fill-level of a liquid in a cup using an RGB-D camera. Our approach does not make any assumptions about the properties of the liquid like its opaqueness or its refraction index. We develop a probabilistic model using features extracted from RGB and depth data. Our experiments demonstrate the robustness of our method and an improvement over the state of the art.

Show abstract Motion analysis is important in a broad range of contexts, including animation, bio-mechanics, robotics and experiments investigating animal behavior. For applications, in which tracking accuracy is one of the main require- ments, passive optical motion capture systems are widely used. Many skeleton tracking methods based on such systems use a predefined skeleton model, which is scaled once in the initialization step to the individual size of the character to be tracked. However, there are remarkable differences in the bone length relations across gender and even more across mammal races. In practice, the optimal skeleton model has to be determined in a manual and time-consuming process. In this paper, we reformulate this task as an optimization problem aiming to rescale a rough hierarchical skeleton structure to optimize probabilistic skeleton tracking performance. We solve this optimization problem by means of state-of-the-art black- box optimization methods based on sequential model-based Bayesian optimization (SMBO). We compare different SMBO methods on three real-world datasets with an animal and humans, demonstrating that we can automatically find skeleton structures for previously unseen mammals. The same methods also allow an automated choice of a suitable starting frame for initializing tracking.

Show abstract State-of-the-art neural microprobes contain hundreds of electrodes within a single shaft. Due to hardware and wiring restrictions, it is usually only possible to measure a small subset of the available electrodes simultaneously. The selection of the best channels is typically performed offline either manually or automatically. However, having a fixed selection for long-term observation does not allow the system to react to changes in the neural activity, and may therefore lead to the loss of important information. In this paper, we formulate the process of autonomously selecting the best subset of electrodes as a combinatorial multi-armed bandit problem with non-stationary rewards, thus allowing the probe to adapt its selection policies online. In order to minimize exploratory actions of the probe, we furthermore take advantage of the existing dependencies between neighboring channels. Our approach is an adaptation of the discounted upper confidence bounds (D-UCB) algorithm, and identifies the electrodes providing the largest amount of non-redundant information. To the best of our knowledge, this is the first online approach for the problem of electrode selection. In extensive experiments, we demonstrate that our solution is not only able to converge towards an average optimal selection policy, but it is also able to react to changes in the neural activity or to damages of the recording electrodes.