Towards a Methodology for Using EEG Motor Imagery to Control Video Games

Nowadays, enhancing the quality of life for individuals with reduced mobility or physical disabilities is a crucial challenge. In recent years, new technologies and techniques, such as Brain-Computer Interfaces, have emerged as viable solutions. The research has focused on developing applications and devices to help people complete everyday tasks autonomously. These tools reduce the need for external caregivers. In this work, we present three main contributions: a near-real-time Electroencephalography Motor Imagery classification system, a pipeline that integrates the classification system and seamlessly controls external devices or applications via WebSockets and a Generative Adversarial Network to perform data augmentation and allow rapid and efficient testing. We will also present a virtual environment that we used to test the pipeline usability in real-world case scenarios. Our work, while being a proof of concept and an initial study shows promising results in terms of accuracy, resource efficiency and latency, outperforming the systems and models found during the literature review in terms of EEG recording duration, with comparable accuracy.