Machine learning for post‐traumatic stress disorder identification utilizing resting‐state functional magnetic resonance imaging

Abstract

Early detection of post-traumatic stress disorder (PTSD) is essential for proper treatment of the patients to recover from this disorder. The aligned purpose of this study was to investigate the performance deviations in regions of interest (ROI) of PTSD than the healthy brain regions, to assess interregional functional connectivity and applications of machine learning techniques to identify PTSD and healthy control using resting-state functional magnetic resonance imaging (rs-fMRI). The rs-fMRI data of 10 ROI was extracted from 14 approved PTSD subjects and 14 healthy controls. The rs-fMRI data of the selected ROI were used in ANOVA to measure performance level and Pearson's correlation to investigate the interregional functional connectivity in PTSD brains. In machine learning approaches, the logistic regression, K-nearest neighbor (KNN), support vector machine (SVM) with linear, radial basis function, and polynomial kernels were used to classify the PTSD and control subjects. The performance level in brain regions of PTSD deviated as compared to the regions in the healthy brain. In addition, significant positive or negative functional connectivity was observed among ROI in PTSD brains.The rs-fMRI data have been distributed in training, validation, and testing group for maturity, implementation of machine learning techniques. The KNN and SVM with radial basis function kernel were outperformed for classification among other methods with high accuracies (96.6%, 94.8%, 98.5%) and (93.7%, 95.2%, 99.2%) to train, validate, and test datasets, respectively. The study's findings may provide a guideline to observe performance and functional connectivity of the brain regions in PTSD and to discriminate PTSD subject using only the suggested algorithms. HighlightsMachine learning techniques are applied to identify cancerous patients utilizing resting-state fMRI data. Among several MLAs, the SVM-RBF performed a better prediction rate at higher reliability than other competitive models.

Arabic Speech Analysis for Classification and Prediction of Mental Illness due to Depression Using Deep Learning

“…Depression leads to somatic problems, mental disorders, sleep disorders, and gastrointestinal problems. The self-confidence and rumination symptoms show in depression-related patients [ 3 , 4 ]. It affects the functioning or performance of patients at school, family, and work.…”

“…e selfcon dence and rumination symptoms show in depressionrelated patients [3,4]. It a ects the functioning or performance of patients at school, family, and work.…”

Application of machine learning techniques in the diagnostic approach of PTSD using MRI neuroimaging data: A systematic review

“…Medial prefrontal cortex, amygdala, thalamus, hippocampus, and precuneus were chosen as the region of interests （ROIs） and compared the functional connections among them in another study. Saba T et al [ 21 ] tested five machine learning algorithms and discovered that SVM had the greatest classification accuracy of 99.2%. The author concluded that this will identify the best algorithm for providing identification recommendations.…”

Machine Learning Models Based on Hippocampal T2-Weighted-Fluid-Attenuated Inversion Recovery Radiomics for Diagnosis of Posttraumatic Stress Disorder

“…These methods mostly had AUCs of 0.75-0.96, ranging in accuracy from 57.58% to 92.5% [45][46][47][48] . MRI techniques, such as fMRI and DTI, combined with machine learning have been used in the prediction, diagnosis, and assessment of PTSD [49,50] . However, machine learning models based on conventional MRI radiomics for PTSD diagnosis have not been reported.…”

“…Liu et al [57] used the combination of multi-level features extracted from fMRI images and multi-kernel learning to classify PTSD cases and healthy controls with an accuracy of 92.5%. Saba et al [49] used machine learning combined with rs-fMRI to classify PTSD cases and healthy controls with an accuracy of 93.7-99.2% (K-nearest neighbor and SVM with radial basis function kernel) in the training, validation, and test groups. This study had a lower diagnostic performance than previous clinical studies due to the following reasons: differences in the types of trauma, assessment methods, demographic characteristics of participants, MRI techniques, brain region selection, or machine learning algorithms.…”