A hybrid BCI-controlled smart home system combining SSVEP and EMG for individuals with paralysis

Abstract In this study, electromyogram (EMG) signals associated with occlusal movement were integrated with steady-state visual evoked potentials (SSVEPs) to develop a hybrid brain–computer interface (hBCI)-based smart home control system for individuals with paralysis. The SSVEP paradigm was used to develop a system containing one main interface and five sub-interfaces corresponding to several devices during the working state, and one interface for the idle state. Participants controlled the devices by gazing at certain stimuli, which flickered at different frequencies for each function. Classical correlation analysis (CCA) of four channel EEG signals was used to recognize SSVEP features as intended selection. Several particular occlusal EMG patterns from the single channel of temporalis muscle were used to confirm the selected function, return from the sub-interface to the main interface, and switch the system on/off, respectively. Five healthy participants and five individuals with paralysis completed the system control experiment. The average target selection accuracy reached 97.5% and 83.6% in healthy participants and patients, while the confirmation accuracy in each group reached 97.6% and 96.9%, respectively. When SSVEPs were combined with EMG signals from occlusal movement to confirm the target selection, the actual control accuracy was maximized to 100%, and the information transmission rate (ITR) reached 45 bit/min among patients. Operation of the hBCI-based smart home control system did not cause higher mental or physical workload in patients compared to healthy participants. Our findings indicate that combining SSVEP and EMG signals effectively enhances the safety and interactivity of hBCI-based smart home systems.