An Intuitive Real-Time Brain Control Interface based on Motor Imagery and Execution.

This study introduces a real-time EEG-based Brain-Computer Interface (BCI) for precise motor imagery and execution classification, focusing on the opening and closing of the hand. The proposed BCI aims to offer intuitive control for external devices, such as neuroprosthetic arms or hand orthoses, by using time and frequency domain features of EEG data when performing or imagining these specific hand movements. The temporal and spectral features are processed through a Machine Learning classifier, employing Support Vector Machine (SVM), Logistic Regression and Random Forest algorithms. The algorithm’s accuracy and response time are evaluated in real-time simulations. The results show promising classification accuracy, with training accuracy about 80% for real movement and 75% for imagined movement, and evaluation accuracy exceeding 72% for real movement and 68% for imagined movement, with less than 45 ms latency in all cases. The study addresses a crucial gap in research by focusing on single-hand motor imagery classification, which is essential for intuitive and practical applications in neuroprosthetics and neurorehabilitation.