Towards a Roadmap for Machine Learning and EEG-Based Brain Computer Interface

Abstract

The technological revolution of the last decades allowed the development of algorithms to perform tasks computationally demanding. Electroencephalography (EEG) signal processing typically requires complex protocols for feature extraction. Machine learning emerged as a potential tool to demystify these complications. Brain-computer interface (BCI) researchers have already implemented machine learning techniques to solve complicated paradigms. These studies achieved significant results and suggested that machine learning could accelerate complex data analysis. This study aims to develop quantitative bibliometric research, i.e., a roadmap, to evaluate the development of studies involving machine learning and BCI applications, especially motor imagery protocols. The results showed that machine learning provides innovative solutions for motor imagery studies. Although, there were few publications related to machine learning and BCI, it is clear that the scientific applications are growing fast, and are developing higher-performance EEG-based approaches.