Characterization of Occipital Alpha Rhythm Towards a Brain Activated Motor Neuroprosthesis

A procedure for EEG acquisition and processing of occipital alpha rhythm signals is presented, in order to characterize and define useful parameters to develop an algorithm for identifying open or closed eyes conditions with a minimum number of electrodes and signal processing. Time and frequency domain features are explored, as well as signal epoch length used in acquisition and processing. A simulation of online operation of a classification algorithm based on a threshold of RMS values is developed and tested, obtaining 90% classification accuracy. This simple approach for EEG signal processing can be adapted as a robust activation and command signal for FES-based Neuroprostheses, with minimal requirements for EEG recording preparation and processing algorithms, suitable for clinical applications.