Robust understanding of EEG patterns in silent speech

Abstract

This paper describes the secondary research on feature extraction and selection for decoding the brain electroencephalograph (EEG) signals in designing a prosthetic arm, a Brain Computer Interface (BCI) system. It considers EEG pattern recognition using Principal Component Analysis (PCA) for Feature Extraction. The data used for this research is the EEG signal that is recorded during the imagination of vowels /a/, /e/, /i/, /o/, /u/ by 20 subjects. Since brain signals are very noisy in nature, a robust PCA is also used to extract the best solution to find principal patterns of the data. The final goal of our research is to train the system based on the information in the sample EEG data and make it ready to classify the pattern correctly.