Document Actions

You are here: Home / The Network / NNCN / Bernstein Collaborations / Movement-Related Brain Signals

Movement-Related Brain Signals

Bernstein Collaboration for Computational Neuroscience (BCOL)

In the "Bernstein" collaboration "Decoding of movement-related brain signals: Improvements by brain and machine learning" between the universities of Freiburg and Tübingen it is intended to use theoretical and experimental know-how in a synergistic way to promote the understanding, application and optimization of on-line analysis of brain signals required for the development of brain-computer-interfaces. In particular patterns of cortical motor activity that are associated with specific movements should be reliably detected. The reliability of the categorization of brain signals will be improved by training subjects to generate specific patterns of brain activity (brain learning) and by using adaptive classification algorithms analyzing the recorded signals (machine learning). It is planned to focus on a well defined classification problem in order to systematically investigate and improve approaches suitable for the discrimination of specific brain activities. Therefore, brain activity will be studied in a so called center-out task, in which subjects move there hand outwards in different directions starting from a marked point.

Subproject 1: Estimation of voluntary hand movement directions from magnetic brain activity: comparison of algorithms and their improvement by learning

Subproject 2: Adaptive algorithm for MEG based brain computer interfaces