RESEARCH AREA C: APPLICATIONS


In its clinical research area, BrainLinks-BrainTools explores new technological solutions and concepts to support neurological and psychiatric patients for whom currently no sufficient treatment is available. The clinical research of BrainLinks-BrainTools focuses on implants, therapy and rehabilitation of patients with epilepsy and stroke, and on neurorobotic as well as motion-capture-based assistive devices for patients with severe chronic paralysis and movement disorders, particularly Parkinson’s disease. Accompanying its clinical research, BrainLinks-BrainTools has a strong focus on neuroethics.

With regard to epilepsy, responsive neurostimulation is a novel concept that may reduce the frequency of epileptic seizures in cases when pharmacological therapy fails. This novel approach needs continuous recording and analysis of brain signals to detect time windows for effective intervention. BrainLinks-BrainTools investigates how feature extraction and machine-learning approaches can be developed and tested for the detection of specific patterns in interictal (i.e. occurring in-between seizures) long-term recordings of patients with epilepsy. The research by BrainLinks-BrainTools has already shown a high accuracy of pathological activity detection together with the ability to recognize changes in the patient’s state. This allows for fast intervention in the range of few seconds, resulting in high chances of seizure cessation or suppression of propagation of pathological activity to other brain areas. A first prototype was designed featuring a complete chain from signal amplification via signal digitization, feature computation to machine learning using a low-power design. For improved diagnostics and basic understanding of epilepsy, BrainLinks-BrainTools works on combined ultrafast fMRI and EEG to identify areas that are activated during interictal spikes and to investigate changes in functional network activity related to epileptic events.

Concerning assistive neurobotic applications for paralyzed patients, BrainLinks-BrainTools has already successfully developed a full demonstrator system. It integrates state-of-the-art online decoding of neuronal control signals using decoders such as deep neural networks, high-level hierarchical planning including a graphical user interface based on the planner’s world knowledge, and novel perception and control algorithms for mobile robots. This system has successfully been used in different scenarios, such as in a robot fetching objects and in an autonomous drinking assistant. Also utilizing novel BCI-based training, stroke-induced language deficits of chronic patients could be reduced significantly.

Addressing Parkinson’s disease, BrainLinks-BrainTools follows an innovative approach based on automatic analysis of motion capture data from a range of different everyday motor tasks. These were specifically chosen to represent the most relevant impairments affecting the quality of life in these patients. Since intended movement amplitudes or velocities of movements of patients suffering from Parkinson’s disease are not a priori known, BrainLinks-BrainTools came up with a new way to automatically quantify their abnormal motor behavior. This approach is based on joint contributions to a certain movement as a function of time rather than on absolute amplitudes and velocities. BrainLinks-BrainTools found that motor dysfunction could be characterized with a very high precision using this novel measure, thus opening up new diagnostic and therapeutic options.

Concerning philosophical and ethical investigations, the interdisciplinary research of BrainLinks-BrainTools has established a framework for ethical evaluation of neurotechnology. This integrates core concepts of action theory and action science (e.g., intention, guidance, affordance) in order to address the specific dimensions of clinical neurotechnology that are relevant to assess its ethical and societal acceptability and to ensure the safety of the user.