Brain computer interface prototypes for upper limb rehabilitation: A review of principles and experimental results

In this paper we review and discuss three novel Brain Computer Interface (BCI) prototypes for upper limb motor function rehabilitation after stroke or brain injury, recently developed at the Laboratory for Biomedical Engineering and Technologies, University of Belgrade - School of Electrical Engineering. All three systems combine movement imagery as a control strategy and functional electrical stimulation (FES) for executing the action and delivering sensory feedback. The BCI system I is a hybrid version comprising both steady state visual evoked potentials (SSVEP) and event related desynchronization (ERD) based modalities for automated selection and triggering of three FES patterns for assistance of grasping. This system was tested in 6 healthy subjects, and all of them were able to control the device with positive predictive value in a range 0.7-0.96 and with true positive rate in a range 0.89-1. BCI system II is a cue-based version of ERD-driven BCI for FES control and was tested in 4 subjects that achieved positive predictive value of 1 and true positive rate in a range 0.8-1. BCI system III is asynchronous version of an ERD-operated brain switch for triggering of FES and it was tested in 4 subject with results for positive predictive value and true positive rate in ranges 0.66-0.92 and 0.81-0.94 respectively.