S. Gruber,K. McLaughlin,A. Etkin,Kelene A. Fercho,J. Daniels,E. Dennis,Xi Zhu,M. Sheridan,J. Fitzgerald,R. Bryant,R. Qi,C. Baird,V. Magnotta,J. Théberge,Evan M. Gordon,J. Blackford,T. Jovanović,Pavel Říha,Jennifer S Stevens,A. Gonenç,Philipp Kinzel,Delin Sun,R. Morey,Ye Zhu,Matthew Peverill,D. Hofmann,M. Stein,S. Sponheim,Sheri M. Koopowitz,D. Grupe,P. Thompson,B. Suarez-Jimenez,I. Koerte,M. van Zuiden,Anna R. Hudson,M. Kennis,Jeffrey P. Guenette,S. M. Nelson,Gen Li,Geoffery May,A. Lazarov,N. Fani,M. Debellis,L. Lebois,Orren Ravid,Chiahao Shih,M. Ross,Raluca M. Simons,Li Wang,T. Varkevisser,R. Herringa,A. King,S. Mueller,S. Disner,R. Neufeld,Maurizio Sicorello,Lee A. Baugh,N. Davenport,N. Jahanshad,S. V. van Rooij,S. Bruce,T. V. van Erp,Kyle Choi,J. Krystal,M. Shenton,C. Larson,E. Geuze,A. Sierk,T. deRoon-Cassini,I. Liberzon,W. E. Hage,D. Stein,Meilin Jia-Richards,T. Straube,A. Simmons,Xin Wang,Xiaofu He,M. Densmore,I. Rektor,C. Haswell,G. Lu,Lauren E. Salminen,I. Rosso,Yoo-Ho Kim,S. Winternitz,A. Cotton,J. Simons,Bobak Hosseini,S. Thomopoulos,D. Veltman,Mike Angstadt,R. Vermeiren,M. Kaufman,C. Weis,Kelly A. Sambrook,G. Forster,J. Frijling,Seonjoo Lee,C. Abdallah,Hong Xie,R. Lanius,E. M. O'Leary,K. Ressler,J. Ipser,Y. Quidé,A. Maron-Katz,J. Bomyea,R. Davidson,S. V. D. van der Werff,S. Koch,Z. Cao,S. Zilcha-Mano,A. Huggins,C. Schmahl,K. Phan,B. Olatunji,J. Cisler,Yi-feng Luo,A. Manthey,M. Korgaonkar,Y. Neria,C. Averill,N. J. van der Wee,L. Nawijn,J. Nitschke,M. Olff,J. Herzog,H. Walter,M. Wall

Abstract:Background Current clinical assessments of Posttraumatic stress disorder (PTSD) rely solely on subjective symptoms and experiences reported by the patient, rather than objective biomarkers of the illness. Recent advances in data-driven computational approaches have been helpful in devising tools to objectively diagnose psychiatric disorders. Here we aimed to classify individuals with PTSD versus controls using heterogeneous brain datasets from the ENIGMA-PGC PTSD Working group. Methods We analyzed brain MRI data from 3,527 structural-MRI; 2,502 resting state-fMRI; and 1,953 diffusion-MRI. First, we identified the brain features that best distinguish individuals with PTSD from controls (TEHC and HC) using traditional machine learning methods. Second, we assessed the utility of the denoising variational autoencoder (DVAE) and evaluated its classification performance. Third, we assessed the generalizability and reproducibility of both models using leave-one-site-out cross-validation procedure for each modality. Results We found lower performance in classifying PTSD vs. controls with data from over 20 sites (60% test AUC for s-MRI, 59% for rs-fMRI and 56% for d-MRI), as compared to other studies run on single-site data. The performance increased when classifying PTSD from HC without trauma history across all three modalities (75% AUC). The classification performance remained intact when applying the DVAE framework, which reduced the number of features. Finally, we found that the DVAE framework achieved better generalization to unseen datasets compared with the traditional machine learning frameworks, albeit performance was slightly above chance. Conclusion Our findings highlight the promise offered by machine learning methods for the diagnosis of patients with PTSD. The utility of brain biomarkers across three MRI modalities and the contribution of DVAE models for improving generalizability offers new insights into neural mechanisms involved in PTSD. Significance ⍰ Classifying PTSD from trauma-unexposed healthy controls (HC) using three imaging modalities performed well (∼75% AUC), but performance suffered markedly when classifying PTSD from trauma-exposed healthy controls (TEHC) using three imaging modalities (∼60% AUC). ⍰ Using deep learning for feature reduction (denoising variational auto-encoder; DVAE) dramatically reduced the number of features with no concomitant performance degradation. ⍰ Utilizing denoising variational autoencoder (DVAE) models improves generalizability across heterogeneous multi-site data compared with the traditional machine learning frameworks

A deep transfer learning approach for improved post-traumatic stress disorder diagnosis

Hierarchical Attention Transfer Networks for Depression Assessment from Speech

Automatic Assessment of Depression from Speech Via a Hierarchical Attention Transfer Network and Attention Autoencoders

A Novel Stochastic Transformer-based Approach for Post-Traumatic Stress Disorder Detection using Audio Recording of Clinical Interviews

Speech-based recognition and estimating severity of PTSD using machine learning

Hybrid Network Feature Extraction for Depression Assessment from Speech

Paying attention to uncertainty: A stochastic multimodal transformers for post-traumatic stress disorder detection using video

Interdisciplinary approach to identify language markers for post-traumatic stress disorder using machine learning and deep learning

Deep Learning based Diagnosis of PTSD using 3D-CNN and resting-state fMRI data

Combining Deep Learning and Graph-Theoretic Brain Features to Detect Posttraumatic Stress Disorder at the Individual Level.

Speech‐based markers for posttraumatic stress disorder in US veterans

Multimodal MRI-Based Classification of Trauma Survivors with and Without Post-Traumatic Stress Disorder

Towards Automatic Depression Detection: A BiLSTM/1D CNN-Based Model

PTSD-MDNN : Fusion tardive de réseaux de neurones profonds multimodaux pour la détection du trouble de stress post-traumatique

Volumetric Integrated Classification Index: an Integrated Voxel-Based Morphometry and Machine Learning Interpretable Biomarker for Post-Traumatic Stress Disorder

Using deep learning to classify pediatric posttraumatic stress disorder at the individual level

Combination Of Grey Matter And White Matter Features For Early Prediction Of Posttraumatic Stress Disorder

The application of machine learning techniques in posttraumatic stress disorder: a systematic review and meta-analysis

Multimodal Imaging-Based Classification of PTSD Using Data-Driven Computational Approaches: A Multisite Big Data Study from the ENIGMA-PGC PTSD Consortium

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

FPT-Former: A Flexible Parallel Transformer of Recognizing Depression by Using Audiovisual Expert-Knowledge-Based Multimodal Measures