Feasibility of an EEG-based brain-computer interface in the intensive care unit

Objective: We tested the feasibility of deploying a commercially available EEG-based brain-computer interface (BCI) in the intensive care unit (ICU) to detect consciousness in patients with acute disorders of consciousness (DoC) or locked-in syndrome (LIS). Methods: Ten patients (9 DoC, 1 LIS) and 10 healthy subjects (HS) were enrolled. The BCI utilized oddball auditory evoked potentials, vibrotactile evoked potentials (VTP) and motor imagery (MoI) to assess consciousness. We recorded the assessment completion rate and the time required for assessment, and we calculated the sensitivity and specificity of each paradigm for detecting behavioral signs of consciousness. Results: All 10 patients completed the assessment, 9 of whom required less than 1 h. The LIS patient reported fatigue before the end of the session. The HS and LIS patient showed more consistent BCI responses than DoC patients, but overall there was no association between BCI responses and behavioral signs of consciousness. Conclusions: The system is feasible to deploy in the ICU and may confirm consciousness in acute LIS, but it was unreliable in acute DoC. Significance: The accuracy of the paradigms for detecting consciousness must be improved and the duration of the protocol should be shortened before this commercially available BCI is ready for clinical implementation in the ICU in patients with acute DoC. (C) 2018 International Federation of Clinical Neurophysiology. Published by Elsevier B.V. All rights reserved.