Single-Channel Wearable EEG for Monitoring Mental Workload for Telerehabilitation Applications

Telerehabilitation has become a promising solution for delivering physical and cognitive therapy to individuals with limited access to in-person care. Its effectiveness can be enhanced by tailoring therapeutic interventions to the patients' cognitive state and workload. This represents a challenging task, as cognitive states are inherently difficult to quantify. This study proposes a low-cost, non-invasive framework for monitoring mental workload (MWL) during a rehabilitation-oriented exergame from single-channel EEG, using a consumer-grade, wearable EEG device. 50 healthy volunteers (13 females) were recruited and a supervised Machine Learning approach was employed for distinguishing between (1) Rest and Task conditions and (2) Task complexity levels, through objective MWL measurements. The proposed framework achieved an Accuracy of 91% for the binary Rest vs Task classification, and a 76% accuracy in a 3-class configuration (Rest, Easy Level, Complex Level), in a Leave-One-Subject-Out Cross Validation approach (LOSO-CV). These findings suggest the feasibility of characterising cognitive states through EEG-based objective metrics and consumer-grade devices, to improve the quality of telerehabilitation protocols and promote patient's adherence and inclusivity.