medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: March 15, 2025

Abstract Objective This study presents a novel computational approach for analyzing electroencephalogram (EEG) signals, focusing on the distribution and variability of energy in different frequency bands. The proposed method, FFT Weed Plot, systematically encodes EEG spectral information into structured metrics that facilitate quantitative analysis. Methods methodology employs Fast Fourier Transform (FFT) to compute Power Spectral Density (PSD) signals. A encoding technique transforms band distributions six-entry vectors, referred as “words,” which serve basis three key metrics: scalar value vector , matrix H . These are evaluated using dataset comprising recordings from 30 healthy individuals 15 patients with epilepsy. Machine learning classifiers then applied assess discriminatory power features. Results classification models achieved 95.55% accuracy, 93.33% sensitivity, 96.67% specificity, demonstrating robustness distinguishing between control epileptic EEGs. Conclusions Plot method provides signal quantification, improving systematization analysis neurophysiological studies. developed could descriptors automated interpretation, offering potential applications clinical research settings. Highlights From domain probability theory, new ways information. step towards automation medical reading. New global description an recording their machine learning. We present new, reproducible, robust clinically designed improve objectivity practice neurophysiology.

Language: Английский