Systematic Review Of Machine Learning In Ptsd Studies For Automated Diagnosis Evaluation

This article reviews 41 studies that utilized Machine Learning (ML) models to diagnose Post-Traumatic Stress Disorder (PTSD). 21 of these studies used neuroimaging techniques such as fMRI, MEG, and EEG to perform automatic PTSD diagnosis across diverse sample groups. Five studies conducted their experiments on heterogeneous sample groups, combining resting-state (rs) amplitude of low-frequency fluctuation (ALFF) and amygdala complex connectivity maps to detect PTSD via a multiclass Gaussian process classifier. The PRISMA guidelines were followed in the review process, which was managed by utilizing Covidence, an online systematic review management system.