EEG-based assessment of sustained attention in geriatric inpatients: a feasibility study

Sustained attention is a core cognitive function, frequently declines with age and neurological disease, and is highly relevant for functional outcomes in geriatric populations. Electroencephalography (EEG) offers high temporal resolution for assessing attentional processes, but its feasibility in multimorbid geriatric inpatients remains insufficiently characterized. This study evaluated the procedural and methodological feasibility of EEG-based sustained-attention assessments in geriatric inpatients undergoing early rehabilitative geriatric complex treatment (ERGCT). Geriatric participants completed standardized neuropsychological testing and two computerized attention tasks, the Psychomotor Vigilance Task (PVT) and a cued Go-NoGo task, during 32-channel EEG recording, including a resting-state recording. Feasibility was assessed using procedural indicators (recruitment, task completion, follow-up participation) and methodological indicators of EEG data quality, specifically the signal-to-noise ratios (SNR) and identification of expected event-related potential (ERP) components (P100, P300). Of 173 patients referred to the ERGCT programme, 43 could be included. Of those, 38 could undergo structured cognitive testing, and resting-state EEG and task-evoked ERPs could be recorded with adequate signal quality However, procedural feasibility was limited by substantial selection and attrition, resulting in incomplete protocol completion. Exploratory analyses suggest an association between higher resting-state alpha-band signal quality and faster PVT reaction times, whereas no associations were observed with global cognition, executive function, or age. Overall, resting-state and task EEG measures of sustained attention could be acquired with adequate data quality only in a selected subgroup of geriatric inpatients, who were able to participate in structured cognitive assessments, while extended and repeated assessment protocols were constrained by procedural burden, suggesting that EEG-based follow-up assessments are difficult to implement under current clinical conditions.