Comparative Analysis of Signal Processing Techniques for Mental State Recognition in Brain-Computer Interfaces (BCI)

The BCI (Brain-computer interface) is a new-age tool in which human or artificial body parts like prosthetic arms can be controlled by sensing the EEG signals. To understand the mental state of the brain, the study of EEG is very important but this technique has its own limitation, which has been identified in this work. Various signal-processing techniques of EEG have been studied to analyze the mental state of the brain. A novel cross-technique will be introduced to improve EEG signal processing techniques. The spatial and temporal resolution trade-off problem is the biggest challenge in EEG signal acquisition systems. This is resolved by taking the fMRI signal and then transforming it into an EEG signal by decomposition and comparing modules for better spatial resolution in the EEG signal. The main limitations like frequency overlapping and decomposition, of the feature extraction technique, have been investigated and an improvised general algorithm will be introduced in this proposed work. This work also touches briefly on different classification algorithms which follow different learning criteria and optimized techniques will be used to achieve better performance. The main objective of this proposed work is the identification of the best-suited algorithm for each step of signal processing to understand the mental state of the brain, brain oscillation characteristics, and some old and new techniques trends companion for the analysis of signal processing techniques of the brain.