Non- Invasive EEG-based BCI system for Left or Right Hand Movement

A brain computer interface (BCI) records the activation of the brain and classifies it into different classes. BCIs can be used by both severely motor disabled as well as healthy people to control devices. The study addresses the development and application of a novel medical technology to measure a patient’s brain activity, translated it with intelligent software, and uses the translated signals to drive patient-specific effectors. In this work, the EEG pattern recognition approach is used based on brain computer interfaces for moving hands right and left. Electroencephalographic (EEG) signals produced by the brain were used as input to the proposed BCI system. There are two BCI approaches used in this paper; the offline BCI approach and the online BCI approach. In the offline approach, the Dataset of motor imagery EEG recordings is used, while in the online approach we used our own BCI system to capture EEG recordings. The practical online testing demonstrates the feasibility of using the proposed system with the ability of real-time processing, automatic analysis. The Principle Component Analysis (PCA) is used for both artifact removal and feature extraction. Wavelet Transformation is also developed to extract the important information from EEG recordings. The K-Nearest-Neighbor (KNN) and Neural Networks (NNs) classifiers were used to find out what the user wants. The results show that we can effectively classify two kinds of tasks based on both BCI approaches with best predictive accuracy of 99.2% for offline approach and 98 % for online approach when wavelet transform and Neural Networks used together. This gives an ideal solution for people with severe neuromuscular disorders, such as Amyotrophic Lateral Sclerosis (ALS) or spinal cord injury, people who are totally paralyzed, or “locked-in”, help them to have a communication channel with others.