11. Motor imagery based BCI for control of FES

Abstract

Purpose: Brain-Computer Interface (BCI) system that can translate subjects’ motor imagery (MI) into real movements using Functional Electrical Stimulation (FES) was developed. MI identification is based on on-line detection of EEG alpha rhythm Event-Related Desynchronisation (ERD). Method: Five healthy subjects participated in this study. Recordings were made using bipolar EEG channel, C3–C4 (international 10–20 standard). We applied FES with a multi-pad electrode system capable of dynamic change of stimulation pattern in real-time, while communicating wirelessly with a PC. Subjects were instructed to imagine the opening movement of the right hand when a visual cue appeared on the computer screen. ERD curves were determined on-line by calculating EEG band-power: band-pass filtering, squaring and smoothing. Parameters for MI detection were frequency band of interest and power threshold. Both parameters were determined during the 5 min training session. When MI was detected, prearranged stimulation pattern was activated. This pattern was generated particularly for hand opening, using specific subset of electrode pads and predefined current amplitudes. Results: Subjects were able to trigger FES using MI with a mean accuracy of 98% and zero false positives. These results indicate that the developed BCI system may be used in motor imagery based training/motor relearning after stroke. Acknowledgment: Research was partly supported by the Ministry of Science and Technological Development of Republic of Serbia, Belgrade (Contract No: 175016).