Impact of responsiveness fluctuations on P300-based classification: an offline study for BCI applications in DoC

Brain-Computer Interfaces (BCIs) represent an important frontier in the clinical care of individuals with Disorders of Consciousness (DoC), with significant potential to improve communication and quality of life. It has not been demonstrated that patients with DoC can reliably control a BCI and this may ultimately be due to fluctuations in responsiveness. Such variability may affect the consistency and reliability of BCI performance in this population. This study investigated the robustness and reliability of an offline classification model in discriminating target from non-target stimuli during auditory oddball experiments over an extended monitoring period, both within the same day and across consecutive days, using EEG data collected from healthy volunteers and patients in a Minimally Conscious State (MCS).