Towards electroencephalography-based consciousness assessment and cognitive function profiling in prolonged disorders of consciousness

Abstract

Abstract Background: Electroencephalography (EEG) -based motor-imagery brain-computer interface (MI-BCI) systems can facilitate training of wilful modulation of sensorimotor oscillations in patients with prolonged disorders of consciousness, potentially enabling empirical assessment of awareness and movement independent responses and question answering. Methods: A multistage, ten session, protocol was evaluated with twenty-four (N = 24) patients who have prolonged disorders of consciousness (PDoC) (unresponsive wakefulness syndrome (UWS) or minimally conscious state (MCS)) or locked-in syndrome (LIS). Phase one assessed capacity to modulate brain activity and thus detect awareness, phase two involved motor-imagery training with neurofeedback, and phase three assessed patients’ motor-imagery response to closed questions, categorised to assess biographical, numerical, logical, and situational awareness. Results: Patients from all groups show significant capacity to modulate brain activity and respond to closed questions with motor imagery. A conservative approach to statistical testing was necessarily taken, due to a limited sample size and uneven groups. However, significant group differences were observed indicating patients with LIS outperform patients with PDoC. As patients diagnosed with UWS present evidence of command following in this protocol and thus could have MCS, significant differences between UWS and MCS across a range of tests were inconclusive although, in general, group performances are correlated with the level of awareness associated with clinical diagnosis. Conclusions: We conclude that the proposed multistage protocol involving progression from motor imagery training with neurofeedback to closed question responses under the correct conditions may be effective for movement-independent consciousness profiling in prolonged disorders of consciousness.