Architectural Review of Co-Adaptive Brain Computer Interface

Abstract

Electroencephalogram (EEG) based brain computer interface (BCI) detect specific EEG pattern from brain scalp and translate them into control commands for external devices. EEG based BCIs are indeed very promising for people suffering from neuromuscular disorder, but still lack adoption as access modalities outside laboratories. Two major factors responsible for this are (a) User variability: huge performance variations among and within user; (b) Signal Variability: High signal variations within or in between BCI sessions. To some extent advanced signal processing and classification methods can adapt with these variability. This makes machines to adapt with users; however these techniques do not focus on cause of these user and signal variability. Co-adaptive BCI systems are newly introduced concept which identify cause of variability and incorporate appropriate action for it. This enables both user and machine adapt with each other. This review paper describes a framework for co-adaptive BCI system as an initial point for any adaptive BCI solution.