Artificial Intelligence in Brain-Computer Interfaces: Current Studies and Emerging Trends

Abstract

: Neurological disorders such as amyotrophic lateral sclerosis (ALS) or stroke impair the control of nerves over muscles, preventing people from living autonomously. Brain Computer Interface (BCI) offers a pragmatic answer, but suffers from scalability issues. This paper discussed the drawbacks of existing BCIs and how AI can be used in order to overcome them. By systematically reviewing and analyzing recent literature, this paper discusses how generative models and adaptive algorithms overcome the bottlenecks of traditional BCI. The combination of artificial intelligence (AI) and BCI solves the fundamental bottlenecks in traditional BCI, including low generalization capability, lack of sufficient data, and high cost. Generative AI-driven data augmentation improves the classification performance in BCIs up to 10-18%. The introduction of intelligent algorithms to adaptively optimize a BCI can significantly reduce its dial-in period, allowing it to operate without expert supervision while maintaining an adaptive response in real time.