The AppleCatcher Game: A Motor Imagery BCI Platform for Investigating Cortical Activation During Imagined Hand Movements

The AppleCatcher game is an innovative brain-computer interface (BCI) designed to support hand rehabilitation in individuals with motor impairments. This EEG-based system enables users to control a virtual apple-catching game using motor imagery (MI) of hand movements, leveraging brain activity without requiring physical motion. By activating overlapping neural networks involved in motor execution and imagery, AppleCatcher aims to promote neuroplasticity and strengthen motor pathways. Unlike conventional MI-based BCIs that rely on EEG sensor-level data, AppleCatcher uses EEG source imaging (ESI) to localize brain activity at the cortical level. This approach enhances precision in monitoring motor-related regions, improving classification accuracy and reliability. The system thus serves both as a platform for studying motor function and recovery and as rehabilitation tool. Initial evaluations using a public EEG dataset for left/right hand motor imagery (MI) produced encouraging results. These were followed by live sessions involving 40 healthy participants during gameplay with the AppleCatcher interface. Among these, the top three subjects achieved classification accuracies of 88%, 86% and 83% using features extracted from sLORETA-based source power estimates and classified with Linear Discriminant Analysis (LDA). The overall average classification accuracy across all participants was 60%. Cluster analysis revealed three distinct groups based on BCI performance, with average accuracies of 79%, 59%, and 49%, respectively. Notably, for one subject, accuracy improved by 25% points between the first and fourth sessions, indicating potential training effects and progressive development of MI-related skills. These findings support the feasibility of the proposed MI-BCI system. Future work will include testing with stroke patients to assess the clinical applicability and rehabilitation potential of the platform.