MI brain-computer interfaces: A concise overview

Abstract

BCI stands for brain–computer interface, and this is the hottest topic among the research community to convert brain impulses into preset instructions that can be used to interact with each other or operate external equipment. A variety of BCI systems have been created during the last decades to give an alternate communications platform for people with extreme neurological illnesses including “amyotrophic lateral sclerosis, spinal cord damage, and brainstem stroke”. BCI literature reviews have also been produced to give BCI researchers well-organized information on important BCI research topics as well as valuable suggestions for the effective design of BCI systems. However, no one has lately looked into general trends in a range of EEG-based BCI study features. The inability of an EEG to measure neural activity down beyond the top layers of the brain is one of its limitations. Additionally, they have low signal-to-noise ratios and are useless for pinpointing the precise location of medications and neurotransmitters in the brain. In this paper, a literature review is conducted with 60 research appears collected from different international journals, published during different years (2013–2022). To be more valuable, this survey provides information about different aspects like signal acquisition techniques, EEG-based BCI paradigms, and channel selection techniques applied in the collected papers. A brief review of various techniques for feature extraction and classification is also included. In addition, the performance of each of the works in terms of accuracy, sensitivity, and kappa coefficient as well is manifest.