Comparative Analysis of Denoising Techniques for Optimizing EEG Signal Processing

Abstract

Electroencephalogram (EEG) is a non-invasive technology widely used to record the brain's electrical activity. However, noise often contaminates the EEG signal, including ocular artifacts and muscle activity, which can interfere with accurate analysis and interpretation. This research aims to improve the quality of EEG signals related to concentration by comparing the effectiveness of two denoising methods: Independent Component Analysis (ICA) and Principal Component Analysis (PCA). Using commercial EEG headsets, this study recorded Alpha, Beta, Delta, and Theta signals from 20 participants while they performed tasks that required concentration. The effectiveness of the denoising technique is evaluated by focusing on changes in standard deviation and calculating the Percentage Residual Difference (PRD) value of the EEG signal before and after denoising. The results show that ICA provides better denoising performance than PCA, as reflected by a significant reduction in standard deviation and a lower PRD value. These results indicate that the ICA method can effectively reduce noise and preserve important information from the original signal.