Gabriel Wagner vom Berg: Riemannian Manifold Learning Beyond EEG

BCCN Berlin / Technische Universität Berlin

 

Abstract

 

Riemannian manifold learning using covariance matrices has gained popularity in neuroscientific research. It has shown to outperform variance-based methods in EEG-based BCIs. In principal, all oscillatory multivariate electrophysiological data can be adapted to apply Riemannian methods. This thesis analyses the application of Riemannian manifold learning to ECoG and LFP data. It was found that Riemannian methods outperform Euclidean methods for ECoG data while LFP-based analysis favors Euclidean methods.

 

 

Additional Information

Master Thesis Defense

 

Presently, the audience is limited to 10.

If you would like to attend you must show either proof of full vaccination (2 weeks post-final vaccination) or a negative corona test (no self-tests).

Please send a short email to gabriel AT bccn-berlin DOT de & graduateprograms AT bccn-berlin DOT de if you plan on joining.

 

Organized by

Prof. Benjamin Blankertz   & Prof. Wolf-Julian Neumann   / Lisa Velenosi