Georgia Southern University
Self-Supervised Contrastive Learning Using EEG Signals For Mental Stress Assessment
Abstract
dc:description.abstractA study is presented to investigate self-supervised contrastive learning (SSCL) models using physiological data obtained from non-invasive wearable sensors for mental stress assessment. The present work involved acquisition of electroencephalography (EEG) signals using wearable sensors, signal preprocessing, data augmentation, and investigation of self-supervised contrastive learning (SSCL) algorithms for multi-class mental stress assessment. Seven volunteers participated in this study executing various mental tasks while wearing an OpenBCI head cap to acquire EEG signals. The acquired EEG signals were preprocessed and utilized for data augmentation in time and frequency domains with different SSCL models. Optimal data augmentation combinations and SSCL models were identified leading to multi-class mental stress classification accuracy in the range of 98 to 99.9%. The study demonstrates the effectiveness of SSCL models using relatively small-size unlabeled EEG signals for multi-class mental stress assessment with a reasonably high accuracy. The potential of using SSCL models with EEG signals for mental stress assessment and intervention through human-robot interactions (HRI) is outlined.
Degree
thesis:*- Name thesis:degree_name
- Master of Science, Mechanical Engineering
- Level thesis:degree_level
- Dissertation (open access)
- Discipline thesis:degree_discipline
- Department of Mechanical Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Uraechu, Ugochukwu
- Contributors dc:contributor
-
- Jinki Kim
- JungHun Choi
Subjects
dc:subject × 8Rights
- Licence dc:rights.license
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.georgiasouthern.edu/etd/2797