Classification of left and right foot movement intention based on steady-state somatosensory evoked potentials

Abstract

Recently, steady-state somatosensory evoked potentials (SSSEPs) which are brain responses to tactile stimulation of specific frequency in somatosensory have been researched in brain-computer interface (BCI) groups. Classification of both feet is important in gait control system. Previous SSSEP studies have mainly researched a feasibility of discrimination by stimulator attached on upper limb (e.g., finger or arm). However, SSSEP-based classification of both feet could be useful in BCI-based gait rehabilitation system. Hence, we investigate a possibility of discrimination of both feet using SSSEP. To this end, we obtain optimal stimuli frequencies in the screening session. In subsequence test session, the optimal stimuli were attached on the left and right foot, respectively. Six healthy subjects conducted the task which was the subjects concentrate on the tactile stimuli following by random visual cue. The classification results show 72.6% and 72.2% in two methods (i.e., common spatial pattern (CSP) and power spectral density (PSD)). Furthermore, we analyzed differences of spatial and spectral features for reliable BCI performance. These results suggest that classification both feet can be available in SSSEP-based BCI for gait rehabilitation.