NEURAL-NETWORK-BASED PREDICTIONS OF HAND MOVEMENTS USING SIMULATED AND REAL EEG DATA

Recently, neural networks have been used as a tool for the classificat ion of spatio-temporal EEG patterns arising from a hand movement exper iment. Results indicated that, based on single-trial EEG data recorded before movement, the side of hand movement can be predicted with fair ly high precision, but variability of results raised the question of t heir validity. In order to validate results obtained with real EEG dat a, an equivalent simulated movement experiment was performed. Alpha ba nd rhythms composed of two components were simulated as a superpositio n of two second-order autoregressive (AR) processes. These simulated E EG data were then filtered, and their power values calculated and used as features in a classification task. Systematic analysis of the sens itivity of the classification results on various simulation parameters was performed. The analysis showed that the Cascade-correlation (CC) network is able to perform satisfactorily in an extremely noisy enviro nment.