Motor Imagery-Based Brain-Computer Interfaces: An Exploration of Multiclass Motor Imagery-Based Control for Emotiv EPOC X

Paulina Tarara¹Iwona Przybył²Julius Schöning³Artur Gunia^4*

• ¹Academy of Fine Arts and Design, Katowice, Poland
• ²Business Service Galop, Katowice, Poland
• ³Osnabrück University of Applied Sciences, Osnabrück, Germany
• ⁴Jagiellonian University, Kraków, Poland

Enhancing motor imagery (MI)-based brain-computer interfaces (BCIs) command capacity is a key challenge in neuroinformatics, particularly for real-world assistive applications. To address this, a multiclass BCI system was developed that classifies resting state and five distinct MI tasks-left and right hand, tongue movement, and left and right lateral bending-using electroencephalography (EEG) data acquired via the Emotiv EPOC X headset. Machine learning algorithms were applied to identify discriminative patterns in the recorded signals. Participants underwent a body awareness training protocol combining mindfulness and physical exercises to improve MI performance. Seven individuals participated in experimental sessions, evaluating system feasibility through behavioral assessments and qualitative feedback. Results indicated modest gains in MI proficiency post-training; however, classification performance was hindered by signal variability and hardware limitations. These findings underscore the importance of integrating training strategies and optimizing signal quality in low-cost BCI systems, while highlighting the potential of multiclass MI paradigms for scalable, user-centered neurotechnologies.