Empowering Brain-Computer Interface (BCI): Al-Based Interpretation of Motor Imagery in EEG Signals
Bachir Daoud, Fatima Zohra Nardjis Belounis, Sabrine Ammar , Sarra Mousselmal
كلية الرياضيات وعلوم الحاسب-جامعة ابن خلدون – تيارت · الجزائر
Brain-Computer Interface (BCI) technology facilitates direct interaction between individuals and computer systems by capturing brain activities, specifically through electroencephalogram (EEG), without relying on physical movement. Our research delves into the realm of 'Motor Imagery,' a BCI modality where users simulate movements mentally, bypassing the need for physical execution. This methodology harnesses brain signals evoked by imagined actions to command external devices. Particularly beneficial for individuals afflicted by conditions like cerebral palsy, stroke, amyotrophic lateral sclerosis, or spinal cord injuries, BCIs aim to restore impaired neural pathways, thereby reinstating lost motor functions. Moreover, leveraging artificial intelligence, notably via Artificial Intelligence model-enabled data collection, extraction, and signal classification, opens avenues for innovative applications such as 'Thought-Controlled Driving' or 'Mind-Powered Device Operation,' seamlessly translating classification outcomes into actionable commands. Additionally, we believe these technologies can be further enhanced by incorporating the results and performance of the random forest model used in our prototypes, achieving an accuracy of 82.0%, a recall rate of 81.8%, and an F1 score of 81.4%.