Brian Electrical Activity Acquisition and Decoding Algorithms For Using BMIs

Abstract

Physical disabilities in limbs often lead to severe impacts on personal life and work. However, BMI technology offers hope for amputees or paralyzed individuals to regain control and motor abilities. By reading signals from the brain's motor cortex, BMI technology enables the control of robotic arms, cursors, and more, facilitating functional recovery. The key lies in identifying and extracting information related to specific tasks or intentions from brain signals, where the decoder determines the accuracy, reliability, and sensitivity of the final movements. This paper aims to provide a brief overview of the decoding steps in BMIs. The paper will discuss how the raw signals are acquired from the brain, including invasive BMI and noninvasive BMI, along with the respective advantages and disadvantages of these two approaches. Then, it will explore how these signals are processed into specific commands through both linear and nonlinear algorithms. Additionally, the paper will discuss the strengths, limitations, and feasibility of different decoding methods in human-computer interaction applications. This review offers researchers a preliminary understanding of the acquisition and decoding algorithms for brain electrical activity in BMIs. By evaluating the strengths and limitations of different algorithms, we can provide guidance for designing and improving BMI systems, thereby supporting the achievement of more accurate and efficient human-computer interaction.