Research Article
EEG-Based Machine Learning Framework for Detection of Post-Traumatic Stress Disorder
@INPROCEEDINGS{10.4108/eai.28-4-2025.2357805, author={Badhmabanu Rajagopalan and Janani Selvam and Parameswaran Sarvalingam and Asick Ali M}, title={EEG-Based Machine Learning Framework for Detection of Post-Traumatic Stress Disorder}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part I}, publisher={EAI}, proceedings_a={ICITSM PART I}, year={2025}, month={10}, keywords={ptsd eeg machine learning random forest signal processing biomarkers mental health diagnosis feature extraction neural signals electroencephalography classification models auc-roc}, doi={10.4108/eai.28-4-2025.2357805} }- Badhmabanu Rajagopalan
Janani Selvam
Parameswaran Sarvalingam
Asick Ali M
Year: 2025
EEG-Based Machine Learning Framework for Detection of Post-Traumatic Stress Disorder
ICITSM PART I
EAI
DOI: 10.4108/eai.28-4-2025.2357805
Abstract
PTSD is an under- and misdiagnosed psychiatric disorder characterised by complicated neurophysiology, which is sometimes difficult to be diagnosed accurately due to subjective based methods. This work introduces a machine learning driven strategy to identify PTSD in a more objective manner using (EEG) signals. The overall system involves repetitive stages; signal acquisition, artifact removal, feature extraction (time-domain, frequency-domain and nonlinear features), dimensionality reduction and classification with Support Vector Machines (SVM), Random Forest (RF), Logistic Regression (LR) and Multilayer Perceptron (MLP). The Random Forest classifier showed the best performance with accuracy, precision, and AUC-ROC of 90.3%, 91.1% and 93.1% respectively, indicating the possibility of applying EEG-based diagnostic methods. This model offers scale, non-intrusiveness and cost-effective support for clinical decision making in mental health, and may pave the way for AI and neurophysiological signal processing-based PTSD diagnosis.
