An evaluation of combined objective neurophysiologic markers to aid assessment of prolonged disorders of consciousness (PDoC)

Objective. Clinical assessments of individuals with Cognitive-Motor Dissociation (CMD) following brain injury are challenging and prone to errors. This prompts investigation of objective, movement-independent neurophysiological markers using electroencephalography (EEG)-based Brain-Computer Interface (BCI) technology. The current pilot study involving adults with prolonged disorders of consciousness (PDoC) investigated the combination of Motor-Imagery BCI (MI-BCI) training and auditory evoked Event Related Potentials (ERPs) using an oddball paradigm to produce complementary biomarkers to improve evaluation of awareness in PDoC .
Approach. EEG data (16 channels) were collected from participants with Unresponsive Wakefulness Syndrome (UWS, n = 2), Minimally Conscious State (MCS, n = 3), and Locked-In Syndrome (LIS, n = 4). The MI-BCI involved assessing sensorimotor rhythm modulation, motor-imagery training with and without auditory feedback, and motor-imagery responses to closed questions over 12 sessions each lasting ~1hour. The oddball protocol was also deployed in 2-3 of those sessions, with ~10 days between first and last, featuring auditory stimuli, comprising two 5-minute sets of standard, deviant beeps plus novel sounds, in a structured ratio. We expected those with the lowest levels of awareness would have reduced ERP components, with highest latencies to peak, as well as lowest accuracy in the motor imagery BCI protocol – and that trends across these metrics would be observed across the three patient groups based on their clinical diagnoses.
Main results. Significant differences in mean N1 component latencies and mean MI Decoding Accuracies (DA, for significant runs) occurred between groups – with shorter N1 latencies for the LIS and MCS groups than for the UWS group (LIS vs. UWS and MCS vs. UWS, p < 0.001), and higher DA for the LIS group compared to MCS and UWS (p < 0.001). Mean DA were found to have a significant negative correlation with mean N1 latencies (two-tailed, p = 0.017).
Significance. The results indicate that neurophysiological markers from the concomitant application of an MI-BCI and auditory-oddball paradigm can augment standard clinical assessments by providing objective measures that produce robust evidence of awareness in people with PDoC.