Anxiety detection using EEG data

Abstract

This research paper delves into the exploration of brainwave patterns linked to anxiety using Electroencephalography (EEG) data. The study focuses on understanding the neural underpinnings of emotions, particularly anxiety, by leveraging EEG’s non-invasive and real-time data capture capabilities. The literature review encompasses an analysis of key EEG datasets of 29 subjects(14 Males + 15 Females) and multiple machine learning algorithms mainly random forest applied to emotional classification using EEG data. The experimental setup involves data collection with the NeuroSky Brainwave starter kit, data preprocessing, and the application of machine learning algorithms, leading to the identification of the Random Forest model as the most effective. Additionally, Explainable AI (XAI) techniques, specifically SHAP, are utilized to unveil the critical EEG frequency bands contributing to anxiety. The findings underscore the significance of beta(13Hz - 30Hz) and lower gamma(30Hz - 60Hz) frequency bands in the EEG signals of individuals experiencing anxiety, thereby providing valuable insights into emotional processing and affective neuroscience