Sleep Disorder Classification from EEG Signals using Fast Fourier Transform and Convolutional Neural Networks

Sleep disorders are a significant global health issue affecting millions, necessitating precise diagnostic methods crucial for effective treatment and patient well-being. This paper addresses the challenges inherent in diagnosing sleep disorders, including interpreting complex electroencephalogram (EEG) signals amidst advancements in diagnostic methodologies. Recent innovations underscore the difficulty of extracting meaningful insights from EEG data due to its dynamic nature and variability. Key challenges include accurately identifying disorder-related patterns amidst noise in EEG recordings and the need for reliable classification frameworks that can accommodate diverse manifestations of sleep disorders. To tackle these challenges, this research study proposes a novel method for classifying sleep disorders from EEG data. This approach integrates the Fast Fourier Transform (FFT) with Convolutional Neural Networks (CNNs) to analyse EEG signals in the frequency domain. FFT facilitates the extraction of critical frequency components that distinguish different sleep disorders, while CNN s leverage their capability to identify spatial and temporal patterns essential for precise categorization. The experimental results demonstrate significant improvements in accuracy, precision, recall, and F1 scores, highlighting the effectiveness of the proposed method. Achieving a high accuracy rate of 98.62% underscores its potential to enhance sleep disorder diagnostics and personalize treatment plans. This research represents a significant advancement in the field, promising improved patient care through more accurate and timely diagnoses, with implications for integrating these findings into clinical practice.