Samir Brahim Belhaouari

Abstract. This thesis explores significant advancements in EEG signal processing to enhance the understanding of brain function and dysfunction. It begins with an examination of EEG signal acquisition, preprocessing, and feature extraction, comparing various state-of-the-art methods for noise reduction, artifact removal, and feature selection. The integration of advanced machine learning algorithms, such as deep learning models, is investigated to improve the accuracy and efficiency of EEG analysis. The study aims to develop innovative strategies for deciphering EEG patterns associated with neurological conditions like epilepsy, Parkinson’s disease, Alzheimer’s, murmur, and stress disorders, presenting several case studies. Emphasizing real-time EEG analysis, the research also paves the way for wearable EEG devices for continuous brain activity monitoring, potentially revolutionizing healthcare through early detection and personalized treatment plans. By leveraging cutting-edge techniques, the thesis advances the understanding of neurological disorders, leading to more accurate diagnoses and improved patient care.
    

This work is part of a multi-paper series. In the thesis thesis we introduce the methodology and main results of Progressive Fourier Transform, we give another implementation in this paper. We then take the idea further in the form of Forward–Backward Fourier transform in this paper.

Illustration of the proposed experimental design.