FAMOX (JUNIOR RESEARCH GROUP)

ADAPTIVE NEURONAL SIGNAL PROCESSING METHODS FOR THE FACILITATION OF ATTEMPTED MOTOR EXECUTION


Relevant for Research Area

B - Core Technologies

C - Applications


Summary

FAMOX aims at developing multivariate machine learning methods, which can serve to decode predictive information about an upcoming motor task from pre-trial brain signals. The use of such data analysis methods is exemplified in a BCI-supported closed-loop scenario, where a brain-state decoding technique is utilized in the context of a hand motor training. This closed-loop training shall prospectively be applied with patients, that have experienced motor deficits by stroke. Besides the decoding task, important other practical data analysis problems are tackled in FAMOX, which are relevant to closed-loop applications of neurotechnological systems. Approaches to reduce the calibration and ramp-up time for brain signal decoding shall be tackled with methods, that allow to learn how to transfer hyperparameters, features or decoding models between users and sessions. Furthermore we develop a framework for data driven characterization of task-relevant brain states with clustering methods. In the context of adaptive deep brain stimulation (DBS), we explore machine learning algorithms, which help to select stimulation parameters optimally for the currently observed brain state.


Publications and Achievements

Journal publications (in chronological order):

[1]            I. Winkler, S. Brandl, F. Horn, E. Waldburger, C. Allefeld, M. Tangermann, “Robust artifactual independent component classification for BCI practitioners”, Journal of Neural Engineering 11 035013, 2014.

[2]            J. Höhne, E. Holz, P. Staiger-Sälzer, K.-R. Müller, A. Kübler, M. Tangermann, “Motor Imagery for Severely Motor-Impaired Patients: Evidence for Brain-Computer Interfacing as Superior Control Solution”, PLoS ONE 9 e104854, 2014.

[3]             P.-J. Kindermans, M. Schreuder, B. Schrauwen, K.-R. Müller and M. Tangermann, “True Zero-Training Brain-Computer Interfacing – An Online Study”, PLoS ONE 9 e102504, 2014.

[4]             J. E. Huggins, C. Guger, B. Allison, C. W. Anderson, A. Batista, A.-M. Brouwer, C. Brunner, R. Chavarriaga, M. Fried-Oken, A. Gunduz, D. Gupta, A. Kübler, R. Leeb, F. Lotte, L. E. Miller, G. Müller-Putz, T. Rutkowski, M. Tangermann, D. E. Thompson, “Workshops of the fifth international brain-computer interface meeting: defining the future.” Brain-Computer Interfaces, 1(1), 27-49, 2014.

[5]             Pieter-Jan Kindermans, Michael Tangermann, Klaus-Robert Müller and Benjamin Schrauwen, ”Integrating dynamic stopping, transfer learning and language models in an adaptive zero-training ERP speller”, Journal of Neural Engineering, Vol. 11 035005, 2014.

[6]             G. R. Müller-Putz, R. Leeb, M. Tangermann, J. Höhne, A. Kübler, F. Cincotti, D. Mattia, R. Rupp, K.-R. Müller, J. d. R. Millán, “Towards Non-Invasive Hybrid Brain-Computer Interfaces: Framework, Practice, Clinical Application and Beyond”, Proceedings of the IEEE 103 926-943, 2015.

[7]             S. Castaño-Candamil, J. Höhne, J.-D. Martinez-Vargas, X.-W. An, G. Castellanos-Dominguez, S. Haufe, “Solving the EEG inverse problem based on space-time- frequency structured sparsity constraints”, NeuroImage 118 598-612, 2015.

[8]             A. Meinel, S. Castaño-Candamil, J. Reis, M. Tangermann , “Pre-Trial EEG-Based Single-Trial Motor Performance Prediction to Enhance Neuroergonomics for a Hand Force Task.” Frontiers in Human Neuroscience, 2016.

[9]             R. T. Schirrmeister, J. T. Springenberg, L. D. J. Fiederer, M. Glasstetter, K. Eggensperger, M. Tangermann, F. Hutter, W. Burgard, T. Ball, “Deep Learning with Convolutional Neural Networks for EEG Decoding and Visualization”, Human Brain Mapping 38 5391-5420, 2017

[10]         D. Hübner, T. Verhoeven, K. Schmid, K.-R. Müller, M. Tangermann and P.-J. Kindermans, "Learning from label proportions in brain-computer interfaces: Online unsupervised learning with guarantees", PLOS ONE. Vol. 12(4), pp. e0175856. Public Library of Science. 2017.

[11]         T. Verhoeven, D. Hübner, M. Tangermann, K.-R. Müller, J. Dambre and P.-J. Kindermans, "Improving zero-training brain-computer interfaces by mixing model estimators", Journal of Neural Engineering. Vol. 14(3), pp. 036021. IOP Publishing. 2017.

[12]         S. Castaño-Candamil, A. Meinel, M. Tangermann, “Post-hoc labeling of arbitrary EEG recordings for data-efficient evaluation of neural decoding methods”, ArXiv e-prints, 2017.

[13]         A. Meinel, S. Castaño-Candamil, B. Blankertz, F. Lotte, M. Tangermann, “Characterizing Regularization Techniques for Spatial Filter Optimization in Oscillatory EEG Regression Problems”, Springer Neuroinformatics, 2018.

[14]         D. Hübner, T. Verhoeven, K.-R. Müller, P.-J. Kindermans and M. Tangermann, "Unsupervised Learning for Brain-Computer Interfaces Based on Event-Related Potentials: Review and Online Comparison", IEEE Computational Intelligence Magazine. Vol. 13(2), pp. 66-77. IEEE. 2018.

Under revision:

[15]         A. Meinel, H. Kolkhorst, M. Tangermann, “Mining within-trial oscillatory brain dynamics to address the variability of optimized spatial filters”, submitted to IEEE TNSRE

Peer-reviewed conference publications (in chronological order):

[16]         P.-J. Kindermans, B. Schrauwen, B. Blankertz, K.-R. Müller, M. Tangermann, Transferring Unsupervised Adaptive Classifiers Between Users of a Spatial Auditory Brain-Computer Interface, Proc. 6th Int. BCI Conf., pp. 330-333, 2014.

[17]         I. Winkler, S. Debener, K.-R. Müller,M. Tangermann, “On the Influence of High-Pass Filtering on ICA-Based Artifact Reduction in EEG-ERP”, Proc. 37th Int. Conf. of the IEEE Eng. in Medicine and Biology Soc. (EMBC) 4101–4105, 2015

[18]         A. Meinel, S. Castaño-Candamil, S. Dähne, J. Reis, M. Tangermann, “EEG band power predicts single-trial reaction time in a hand motor task.” 7th Int. Conf. on Neural Engineering (NER), pp. 182–185, 2015.

[19]         S. Castaño-Candamil, A. Meinel, J. Reis, M. Tangermann, “Correlates to influence user performance in a hand motor rehabilitation task.” Clinical Neurophysiology, 126(8):e166–e167, 2015.

[20]         M. Tangermann, J. Reis, A. Meinel, “Commonalities of Motor Performance Metrics are Revealed by Predictive Oscillatory EEG Components.” Proc. Neurotechnix, 2015.

[21]         S. Castaño-Candamil, A. Meinel, S. Dähne, M. Tangermann, “Probing meaningfulness of oscillatory EEG components with bootstrapping, label noise and reduced training sets.”, 37th Int. Conf. IEEE EMBC, pp. 5159–5162, 2015.

[22]         A. Meinel, E. M. Schlichtmann, T. Koller, J. Reis, M. Tangermann, “Predicting Single-Trial Motor Performance from Oscillatory EEG in Chronic Stroke Patients.” In Proc. of the 6th Int. BCI meeting, 2016.

[23]         S. Castaño-Candamil, S. Dähne, M. Tangermann, “Relevant Frequency Estimation in EEG Recordings for Source Power Co-Modulation.” In Proc. of the 6th Int. BCI Meeting, 2016.

[24]         A. Meinel, K. Eggensperger, M. Tangermann, F. Hutter, “Hyperparameter Optimization for Machine Learning Problems in BCI.” In Proc. of the 6th Int. BCI meeting, 2016.

[25]         A. Meinel, F. Lotte, M. Tangermann, “Tikhonov Regularization Enhances EEG-Based Spatial Filtering For Single-Trial Regression”, Proc. of the 7th Graz BCI Conf., pp. 308- 313, 2017.

[26]         S. Castaño-Candamil, M. Tangermann, “Subspace Decomposition in the Frequency Domain”,Proc. of the 7th Graz BCI Conf., pp. 64-69, 2017.

[27]         D. Hübner, T. Verhoeven, P.-J. Kindermans and M. Tangermann, "Mixing two unsupervised estimators for event-related potential decoding: An online evaluation", Proc. of the 7th Graz BCI Conf., pp. 198-203. 2017.

[28]         D. Hübner, P.-J. Kindermans, T. Verhoeven and M. Tangermann, "Improving learning from label proportions by reducing the feature dimensionality", Proc. of the 7th Graz BCI Conf., pp. 186-191. 2017.

[29]         D. Hübner, T. Verhoeven, K. Schmid, K.-R. Müller, M. Tangermann and P.-J. Kindermans (2017), "Learning from label proportions in BCI -- A symbiotic design for stimulus presentation and signal decoding", In The First Biannual Neuroadaptive Technology Conference. pp. 27-29.

[30]         A. Meinel, T. Koller, M. Tangermann, "Time-Frequency Sensitivity Characterization of Single-Trial Oscillatory EEG Components", In The First Biannual Neuroadaptive Technology Conference., pp. 36-37, 2017.

[31]         M. Tangermann, A. Meinel, "Informative Oscillatory EEG Components and their Persistence in Time and Frequency", Neurotechnix, Extended Abstracts Volume 1: CogNeuroEng 17-21, 2017.

[32]         S. Castaño-Candamil, A. Meinel, M. Tangermann, “Post-hoc labeling of arbitrary EEG recordings for data-efficient evaluation of neural decoding methods”, Proc. of the 7th Graz BCI Conf., pp. 58-59, 2018.

[33]         J. Sosulski, D. Hübner, M. Tangermann, “Closed-loop stimulus parameter optimization framework for event-related potential paradigms”, Proc. Seventh Int. BCI Meeting, 2018.

Book articles:

[34]        P.-J. Kindermans, D. Hübner, T. Verhoeven, K. Schmid, K.-R. Müller and M. Tangermann, “Chapter 6: Unsupervised learning for brain–computer interfaces based on event-related potentials”, Signal Processing and Machine Learning for Brain- Machine Interfaces, 2018.

[35]        D. Hübner, P.-J. Kindermans, T. Verhoeven, K.-R. Müller, M. Tangermann, “Rethinking BCI Paradigm and Machine Learning Algorithm as a Symbiosis: Zero Calibration, Guaranteed Convergence and High Decoding Performance”. In Brain-Computer Interface Research: A State-of-the-Art Summary 7, Springer. In press.

Others (workshops, demos, etc.)

[36]         May 2018: We held a workshop about unsupervised learning for Seventh Int. BCI Meeting in Pacific Grove, USA with over 70 attendees.

[37]         Jan 2018: We presented a workshop about unsupervised learning for BCIs at the applied machine learning days in Lausanne with live demo.

[38]         Dec 2017: Schüler-Ingenieur-Akademie: Pupils learn about perspectives in science and could experience a live BCI demo with a visual speller.

[39]         July 2017: Members from the general audience could participate in a live demo at the Wissenschaftsmarkt Freiburg on the central market in Freiburg.

[40]         Mar 2017: Students could participate in a BCI experiment at the central library in Freiburg during the Brain Awareness Week.

[41]         Nov 2016: Pupils could try out an online EEG-based BCI system during the "Tag der offenen Tür".

[42]         Nov 2016: NIPS Workshop, "Making Brain-Computer Interfaces robust, reliable and adaptive with Learning from Label Proportions".

[43]         June 2016: We held a workshop on “Improving BCI Usability through Transfer Learning Methods” at the Seventh Int. BCI Meeting in Pacific Grove, USA with over 40 attendees.

[44]         June 2016: We held a workshop on “Novel application fields for auditory BCIs” at the Seventh Int. BCI Meeting in Pacific Grove, USA with over 30 attendees.

[45]         May 2016: BCI demo for pupils in Straßburg

[46]         Mar 2016: BCI demos at different schools in the Freiburg region during the “Brain Awareness Week”

[47]         Nov 2015: BCI demo during the "Tag der offenen Tür" of the Technical Faculty.