EEG Classification Tool: Distinguishing Between Normal and Abnormal Brainwaves

Masoom J. Desai, Department of Neurology, University of New Mexico
Joseph Picone, Electrical and Computer Engineering, Temple University

Abstract

Problem Statement: Current EEG interpretation heavily relies on clinical expertise, facing challenges in timeliness and variability, crucial in acute settings like stroke surveillance.

Advancing EEG interpretation with AI/ML means using technology to spot abnormalities quickly. Challenges include different types of data and delays in human analysis. Our model uses TUH Abnormal EEG to classify in real-time, in a rapid and efficient manner. Gemein et al.'s research suggests that focusing on specific features could help doctors make better decisions faster in urgent situations.