Identifying Patients With PTSD Utilizing Resting-State fMRI Data and Neural Network Approach

Purpose: The primary aim of the study is to identify existence post-traumatic stress disorder (PTSD) in an individual and detect dominance level each affected brain region PTSD using rs-fMRI data. This will assist psychiatrists neurologists distinguish impartially between individuals healthy controls for brain-based treatment PTSD. Methods: Twenty-eight (14 with PTSD, 14 controls) were assessed obtain data their six regions-of-interest. analyzed by Artificial Neural Network (ANN), adopting training-validation-testing approach classify most due classification accuracy justified a variety different methods metrics. Results: Three ANN models established attain study’s purpose susceptible regions right, left, both hemispheres achieved 79%, 93.5%, 94.5%, respectively. prediction even increased independent holdout sample trained models. developed are reliable, intellectually attractive, generalize. Additionally, dominant was left hippocampus least right hippocampus. Conclusion: present investigation high identified that highly contributed differentiating from controls. results indicated individuals. Therefore, our findings helpful practitioners diagnostic, medication, therapy knowing strength infected regions.