Signals from different brain regions can be used to control motor-prosthetic devices. Major limitations for such neuroprosthetic brain-machine interfaces result from the mixture of motor-related activity with other cognitive signals, especially during unconstraint (‘free’) movements, and insufficient long-term stability of the neural signals with the recording techniques available today. If brain signals can not be identified unambiguously, and recorded stably, this will lead to misinter­pretations of the motor commands and improper control to the prosthetic device. Signal loss could be prevented by continuous adaptive re-adjusting of microelectrode positions, but not in currently available chronic multi-electrode implants, since these are not adjustable.

The goal of this project is to achieve proper motor-goal identification from cortical activity, independent of other cog­nitive signals, and to develop an adaptive multi-electrode positioning (AMEP) system for semi-autonomous cortical multi-channel recordings. Both developments are important for improved brain machine interfacing in neuroprosthetic control and provide links to subprojects 3a and 3b.