Why the study?
PTSD is a serious condition linked to severe anxiety, depression, and cardiovascular complications, making early and accurate detection critical.
Does deep learning analysis of ECG signals improve the detection accuracy of PTSD compared to traditional machine learning?
Comparison
CNN models using ECG scalograms vs traditional ML classifiers using statistical features
Design
Algorithm development and evaluation study
Key result
A deep learning model (ResNet50) using 5-second ECG signal segments achieved 94.92% accuracy and an AUC of 0.99 for detecting PTSD, outperforming traditional machine learning approaches.
Authors
Loading...
Hypothesis-generating for ECG-AI PTSD screening; prospective validation needed before clinical adoption.
Does deep learning analysis of ECG signals improve the detection accuracy of PTSD compared to traditional machine learning?
Effect estimate: AUC 0.99
Deep learning models, particularly ResNet50, can accurately detect PTSD using short segments of ECG signals, offering a potential non-invasive diagnostic tool.
Tasouj et al. (2025) studied Post-traumatic stress disorder (PTSD). Deep learning-based convolutional neural networks (ResNet50) using ECG signals vs. Traditional machine learning classifiers was evaluated on Classification accuracy for PTSD detection (AUC 0.99). A deep learning model (ResNet50) using 5-second ECG signal segments achieved 94.92% accuracy and an AUC of 0.99 for detecting PTSD, outperforming traditional machine learning approaches.