High gamma oscillations enhance the subdural visual speller

The N200 speller is a non-flashing visual brain-computer interface (BCI) using motion-onset visual evoked potentials (mVEPs). Previous N200 speller was implemented at the scalp EEG level. Compared to scalp EEG, electrocorticography (ECoG) provides a broader frequency band that could be utilized in BCI. In this study, we investigated whether the high gamma brain activities recorded from human intracranial electrodes can enhance the performance of the subdural speller. The ERP and high gamma responses of one most task-related subdural electrode were used together for BCI classification and showed that high gamma responses did enhance the performance for the subdural visual motion speller resulted in an average increase of over 8% (p<;0.05, paired t-test).