Development of a BCI system for Enhancing Human-Robot Interaction in Cognitive Stimulation Therapy

Cognitive stimulation therapy (CST) has emerged as an effective treatment for various mental health disorders such as dementia in elderly. CST aims to actively stimulate the patient's brain through multi-sensory experiences. To address the limitation associated with traditional group therapy settings, therapy robots have been introduced for at-home CST. However, these robots rely on physical light and sound sensors, which may not fully assess the patient's mental activity, leading to potential conflicts and reduced therapeutic efficacy. To enhance human-robot interaction and to improve the effectiveness of therapy robots, a brain-computer interface (BCI) feedback system has been developed. The proposed solution features a lightweight, low-cost, and single channel electroencephalogram (EEG) measuring device and Python-based software for real-time EEG processing and classification. The system classifies the patient's alertness level into high or low categories, providing valuable feedback for the therapy robot. Beta power ratio was extracted from the EEG signal and utilized in a support vector machine (SVM) model. The subsystems were evaluated, and the overall system was integrated with a custom-built therapy robot designed to meet CST requirements. The functionality and performance of the new feedback system were demonstrated and assessed in an online setting.