Master Thesis Proposal: Comparison between ERD/ERS and MRCP based movement prediction on EEG-data

Man Machine Interfaces (MMIs) are getting more and more beneficial and applicable to many different fields to support men in both scientific and everyday life. Their usage is however still limited and they need to be further developed and improved to become really efficient. A potentially efficacious approach is movement prediction. It could improve different MMIs systems by providing an interface independent from constant attentional control, reducing the perceived command-response time-lag, or giving an insight into user ́s cognitive state. Prediction requires real-time processing of brain activities related to motor preparation, that occur and are detectable before movement-onset. Two suitable candidates among brain patterns are Movement Related Cortical Potentials (MRCPs) and Event Related Desynchronization/Synchronizations (ERD/ERSs). These patterns differ from many prospectives, it is thus plausible that they differ also for what concerns their applicability and effectiveness in movement prediction. The aim of the present Master Thesis will be to compare motion prediction based on ERD/ERS and MRCP. Most part of the work will focus on what concerns ERD/ERS patterns. The workflow will begin with literature research about the two brain signals, aiming at acquiring a broad knowledge from a neurobiological point of view and what concerns the methods commonly used for their processing, their applicability and limits. This essential stage will bring to the definition of the methods used for processing and of the terms of the comparison. Features like the time of occurrence or the accuracy of predictions are possible candidates as terms of the comparison. The most promising procedures will be implemented in the DFKI software framework “pySPACE”, if not already available; the methods will be finally evaluated and the prediction based on ERD/ERS will be compared with that based on MRCP. The procedures implementation will be based on existing data recorded in a controlled experimental setup; the applicability to data recorded in a more realistic scenario will be assessed in the final phase. Laura Manca Master Thesis Proposal: Comparison between ERD/ERS and MRCP based movement prediction