BACKGROUND: Recent advances in data-driven computational approaches have been helpful in devising tools to objectively diagnose psychiatric disorders. However, current machine learning studies limited to small homogeneous samples, different methodologies, and different imaging collection protocols, limit the ability to directly compare and generalize their results. Here we aimed to classify individuals with PTSD versus controls and assess the generalizability using a large heterogeneous brain datasets from the ENIGMA-PGC PTSD Working group. METHODS: We analyzed brain MRI data from 3,477 structural-MRI; 2,495 resting state-fMRI; and 1,952 diffusion-MRI. First, we identified the brain features that best distinguish individuals with PTSD from controls 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 in each modality (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: These results have the potential to provide a baseline classification performance for PTSD when using large scale neuroimaging datasets. Our findings show that the control group used can heavily affect classification performance. The DVAE framework provided better generalizability for the multi-site data. This may be more significant in clinical practice since the neuroimaging-based diagnostic DVAE classification models are much less site-specific, rendering them more generalizable.

Baylor College of Medicine Houston TX USA; Yale University School of Medicine New Haven CT USA

Brain Health Research Centre Department of Anatomy University of Otago Dunedin New Zealand

Brain Research and Innovation Centre Ministry of Defence Utrecht The Netherlands

Center for Trauma Recovery Department of Psychological Sciences University of Missouri St Louis St Louis MO USA

Center of Excellence for Stress and Mental Health VA San Diego Healthcare System San Diego CA USA

Charité Universitätsmedizin Berlin Campus Charite Mitte Charite Universitatsmedizin Berlin Berlin Germany

Civil Aerospace Medical Institute US Federal Aviation Administration Oklahoma City OK USA

Cognitive and Clinical Neuroimaging Core McLean Hospital Belmont MA USA

Department of Child and Adolescent Psychiatry Psychosomatic and Psychotherapy Ludwig Maximilian University of Munich Munich Germany; Psychiatry Neuroimaging Laboratory Brigham and Women's Hospital Boston MA USA

Department of Medical Imaging Jinling Hospital Medical School of Nanjing University Nanjing Jiangsu China

Department of Neuroscience Western University London ON Canada

Department of Pediatrics University of Minnesota Minneapolis MN USA

Department of Psychiatry Amsterdam University Medical Centers Academic Medical Center University of Amsterdam Amsterdam The Netherlands

Department of Psychiatry Amsterdam University Medical Centers VU University Medical Center VU University Amsterdam The Netherlands

Department of Psychiatry and Behavioral Health Ohio State University Columbus OH USA

Department of Psychiatry Columbia University Medical Center New York NY USA

Department of Psychiatry Columbia University Medical Center New York NY USA; New York State Psychiatric Institute New York NY USA

Department of Psychiatry University of Texas at Austin Austin TX USA

Department of Psychology and Neuroscience Baylor University Waco TX USA

Department of Psychology Vanderbilt University Nashville TN USA

Department of Radiology The Affiliated Yixing Hospital of Jiangsu University Yixing Jiangsu China

Department of Radiology Washington University School of Medicine St Louis MO USA

Departments of Psychology and Psychiatry Neuroscience Program Western University London ON Canada; Department of Psychology University of British Columbia Okanagan Kelowna British Columbia Canada

Division of Neuroradiology Brigham and Women's Hospital Boston MA USA

Division of Women's Mental Health McLean Hospital Belmont MA USA

Donders Institute for Brain Cognition and Behavior Centre for Cognitive Neuroimaging Radboud University Nijmegen Nijmegen The Netherlands

Duke University Durham NC USA

Emory University Department of Psychiatry and Behavioral Sciences Atlanta GA USA

Ghent University Ghent Belgium

Harvard University Boston MA USA

Heidelberg University Heidelberg Germany

Imaging Genetics Center Mark and Mary Stevens Neuroimaging and Informatics Institute Keck School of Medicine of the University of Southern California Marina del Rey CA USA

Institute of Psychology Chinese Academy of Sciences Beijing China

Institute of Psychology Chinese Academy of Sciences Beijing China; Department of Psychology University of Chinese Academy of Sciences Beijing China

Leiden University Medical Center Leiden The Netherlands

Marquette University Milwaukee WI USA

Masaryk University Brno Czechia

McLean Hospital Belmont MA USA; Harvard Medical School Boston MA USA

Medical College of Wisconsin Milwaukee WI USA

Minneapolis VA Health Care System Minneapolis MN USA

Minneapolis VA Health Care System Minneapolis MN USA; University of Minnesota Minneapolis MN USA

Munroe Meyer Institute University of Nebraska Medical Center Omaha NE USA

New York State Psychiatric Institute New York NY USA

Northwestern Neighborhood and Networks Initiative Northwestern University Institute for Policy Research Evanston IL USA

Psychiatry and Behavioral Science Texas A and M University Health Science Center College Station TX USA

Psychiatry Neuroimaging Laboratory Brigham and Women's Hospital Boston MA USA

Sanford School of Medicine University of South Dakota Vermillion SD USA

School of Medicine and Public Health University of Wisconsin Madison Madison WI USA

School of Psychology University of New South Wales Sydney NSW Australia

School of Psychology University of New South Wales Sydney NSW Australia; Neuroscience Research Australia Randwick NSW Australia

Stanford University Stanford CA USA

Tel Aviv University Tel Aviv Israel

UMR 1253 CIC 1415 University of Tours CHRU de Tours INSERM France

University of California Irvine Irvine CA USA

University of California San Diego La Jolla CA USA

University of Cape Town Cape Town South Africa

University of Groningen Groningen The Netherlands

University of Haifa Haifa Israel

University of Illinois at Chicago Chicago IL USA

University of Iowa Iowa City IA USA

University of Michigan Ann Arbor MI USA

University of Münster Münster Germany

University of North Carolina at Chapel Hill Chapel Hill NC USA

University of Rochester Rochester NY USA

University of South Dakota Vermillion SD USA

University of Toledo Toledo OH USA

University of Utah School of Medicine Salt Lake City UT USA

University of Washington Seattle WA USA

University of Wisconsin Madison Madison WI USA

University of Wisconsin Milwaukee Milwaukee WI USA

VISN 17 Center of Excellence for Research on Returning War Veterans Waco TX USA

Wayne State University School of Medicine Detroit MI USA

Westmead Institute for Medical Research Westmead NSW Australia

Yale University School of Medicine New Haven CT USA

Zobrazit více v PubMed

Altmann A, Tolosi L, Sander O, et al., 2010. Permutation importance: a corrected feature importance measure. Bioinformatics 26, 1340–1347. PubMed

Arbabshirani MR, Plis S, Sui J, et al., 2017. Single subject prediction of brain disorders in neuroimaging: promises and pitfalls. Neuroimage 145, 137–165. PubMed PMC

Bai X, Wang X, Liu X, et al., 2021. Explainable deep learning for efficient and robust pattern recognition: a survey of recent developments. Pattern Recognit. 120, 108102.

Behzadi Y, Restom K, Liau J, et al., 2007. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 37, 90–101. PubMed PMC

Belleau EL, Ehret LE, Hanson JL, et al., 2020. Amygdala functional connectivity in the acute aftermath of trauma prospectively predicts severity of posttraumatic stress symptoms. Neurobiol. Stress 12, 100217. PubMed PMC

Ben-Zion Z, Zeevi Y, Keynan NJ, et al., 2020. Multi-domain potential biomarkers for post-traumatic stress disorder (PTSD) severity in recent trauma survivors. Transl. Psychiatry 10, 208. PubMed PMC

Breiman L, 2001. Random forests. Mach. Learn. 45, 5–32.

Bruin WB, Taylor L, Thomas RM, et al., 2020. Structural neuroimaging biomarkers for obsessive-compulsive disorder in the ENIGMA-OCD consortium: medication matters. Transl. Psychiatry 10, 342. PubMed PMC

Bzdok D, Meyer-Lindenberg A, 2018. Machine learning for precision psychiatry: opportunities and challenges. Biol. Psychiatry Cogn. Neurosci. Neuroimaging 3, 223–230. PubMed

Cai XL, Xie DJ, Madsen KH, et al., 2020. Generalizability of machine learning for classification of schizophrenia based on resting-state functional MRI data. Hum. Brain Mapp. 41, 172–184. PubMed PMC

Daniels JK, McFarlane AC, Bluhm RL, et al., 2010. Switching between executive and default mode networks in posttraumatic stress disorder: alterations in functional connectivity. J. Psychiatry Neurosci. 35, 258–266. PubMed PMC

Dennis EL, Disner SG, Fani N, et al., 2021. Altered white matter microstructural organization in posttraumatic stress disorder across 3047 adults: results from the PGC-ENIGMA PTSD consortium. Mol. Psychiatry 26, 4315–4330. PubMed PMC

Dinsdale NK, Jenkinson M, Namburete AIL, 2021. Deep learning-based unlearning of dataset bias for MRI harmonisation and confound removal. Neuroimage 228, 117689. PubMed PMC

Du B, Xiong W, Wu J, et al., 2017. Stacked convolutional denoising auto-encoders for feature representation. IEEE Trans. Cybern. 47, 1017–1027. PubMed

Eslami T, Mirjalili V, Fong A, et al., 2019. ASD-DiagNet: a hybrid learning approach for detection of autism spectrum disorder using fMRI data. Front Neuroinform 13, 70. PubMed PMC

Gao S, Calhoun VD, Sui J, 2018. Machine learning in major depression: from classification to treatment outcome prediction. CNS Neurosci. Ther. 24, 1037–1052. PubMed PMC

Gong Q, Li L, Tognin S, et al., 2014. Using structural neuroanatomy to identify trauma survivors with and without post-traumatic stress disorder at the individual level. Psychol. Med. 44, 195–203. PubMed PMC

Han K, Wen H, Shi J, et al., 2019. Variational autoencoder: an unsupervised model for encoding and decoding fMRI activity in visual cortex. Neuroimage 198, 125–136. PubMed PMC

Im JJ, Kim B, Hwang J, et al., 2017. Diagnostic potential of multimodal neuroimaging in posttraumatic stress disorder. PLoS One 12, e0177847. PubMed PMC

Jahanshad N, Kochunov PV, Sprooten E, et al., 2013. Multi-site genetic analysis of diffusion images and voxelwise heritability analysis: a pilot project of the ENIGMA-DTI working group. Neuroimage 81, 455–469. PubMed PMC

Ju R, Hu C, Zhou P, et al., 2019. Early diagnosis of Alzheimer’s disease based on resting-state brain networks and deep learning. IEEE/ACM Trans. Comput. Biol. Bioinform. 16, 244–257. PubMed

Ju Y, Ou W, Su J, et al., 2020. White matter microstructural alterations in posttraumatic stress disorder: an ROI and whole-brain based meta-analysis. J. Affect. Disord. 266, 655–670. PubMed

Ke J, Zhang L, Qi R, et al., 2018. Typhoon-related post-traumatic stress disorder and trauma might lead to functional integration abnormalities in intra- and inter-resting state networks: a resting-state fmri independent component analysis. Cell. Physiol. Biochem. 48, 99–110. PubMed

Kessler RC, Aguilar-Gaxiola S, Alonso J, et al., 2017. Trauma and PTSD in the WHO World Mental Health Surveys. Eur. J. Psychotraumatol. 8, 1353383. PubMed PMC

Kim J, Calhoun VD, Shim E, et al., 2016. Deep neural network with weight sparsity control and pre-training extracts hierarchical features and enhances classification performance: evidence from whole-brain resting-state functional connectivity patterns of schizophrenia. Neuroimage 124, 127–146. PubMed PMC

Kingma DPW, 2013. M.: auto-Encoding Variational Bayes. arXiv 1312, 6114.

Kingma DPB, J.: adam: a Method for Stochastic Optimization . arXiv 2014; 1412.6980.

Klein A, Tourville J, 2012. 101 labeled brain images and a consistent human cortical labeling protocol. Front. Neurosci. 6, 171. PubMed PMC

Koch SB, van Zuiden M, Nawijn L, et al., 2016. Aberrant resting-state brain activity in posttraumatic stress disorder: a meta-analysis and systematic review. Depress. Anxiety 33, 592–605. PubMed

Koenen KC, Ratanatharathorn A, Ng L, et al., 2017. Posttraumatic stress disorder in the World Mental Health Surveys. Psychol. Med. 47, 2260–2274. PubMed PMC

Lama RK, Gwak J, Park JS, et al., 2017. Diagnosis of Alzheimer’s disease based on structural mri images using a regularized extreme learning machine and PCA features. J. Healthc. Eng. 2017, 5485080. PubMed PMC

Lanka P, Rangaprakash D, Dretsch MN, et al., 2020. Supervised machine learning for diagnostic classification from large-scale neuroimaging datasets. Brain Imaging Behav. 14, 2378–2416. PubMed PMC

Li X, Zhu D, Jiang X, et al., 2014. Dynamic functional connectomics signatures for characterization and differentiation of PTSD patients. Hum. Brain Mapp. 35, 1761–1778. PubMed PMC

Li Y, Zhu H, Ren Z, et al., 2020a. Exploring memory function in earthquake trauma survivors with resting-state fMRI and machine learning. BMC Psychiatry 20, 43. PubMed PMC

Li G, Han D, Wang C, et al., 2020b. Application of deep canonically correlated sparse autoencoder for the classification of schizophrenia. Comput. Methods Programs Biomed. 183, 105073. PubMed

Liu F, Xie B, Wang Y, et al., 2015. Characterization of post-traumatic stress disorder using resting-state fMRI with a multi-level parametric classification approach. Brain Topogr. 28, 221–237. PubMed

Liu S, Zhao L, Wang X, et al., 2021. Deep spatio-temporal representation and ensemble classification for attention deficit/hyperactivity disorder. IEEE Trans. Neural Syst. Rehabil. Eng. 29, 1–10. PubMed

Madley-Dowd P, Hughes R, Tilling K, et al., 2019. The proportion of missing data should not be used to guide decisions on multiple imputation. J. Clin. Epidemiol. 110, 63–73. PubMed PMC

McCunn P, Richardson JD, Jetly R, et al., 2021. Diffusion tensor imaging reveals white matter differences in military personnel exposed to trauma with and without post-traumatic stress disorder. Psychiatry Res. 298, 113797. PubMed

Morey RA, Clarke EK, Haswell CC, et al., 2020. Amygdala nuclei volume and shape in military veterans with posttraumatic stress disorder. Biol. Psychiatry Cogn. Neurosci. Neuroimaging 5, 281–290. PubMed PMC

Moyer D, Ver Steeg G, Tax CMW, et al., 2020. Scanner invariant representations for diffusion MRI harmonization. Magn. Reson. Med. 84, 2174–2189. PubMed PMC

Mwangi B, Tian TS, Soares JC, 2014. A review of feature reduction techniques in neuroimaging. Neuroinformatics 12, 229–244. PubMed PMC

Nunes A, Schnack HG, Ching CRK, et al., 2020. Using structural MRI to identify bipolar disorders - 13 site machine learning study in 3020 individuals from the ENIGMA Bipolar Disorders Working Group. Mol. Psychiatry 25, 2130–2143. PubMed PMC

O’Doherty DCM, Tickell A, Ryder W, et al., 2017. Frontal and subcortical grey matter reductions in PTSD. Psychiatry Res. Neuroimaging 266, 1–9. PubMed

O’Doherty DCM, Ryder W, Paquola C, et al., 2018. White matter integrity alterations in post-traumatic stress disorder. Hum. Brain Mapp. 39, 1327–1338. PubMed PMC

Pedregosa F, Varoquaux GG, A., Michel V, Thirion B, Grisel O, 2011. Scikit-learn: machine learning in python. J. Mach. Learn. Res. 12, 2825–2830.

Pereira F, Mitchell T, Botvinick M, 2009. Machine learning classifiers and fMRI: a tutorial overview. Neuroimage 45, S199–S209. PubMed PMC

Pinaya WHL, Mechelli A, Sato JR, 2019. Using deep autoencoders to identify abnormal brain structural patterns in neuropsychiatric disorders: a large-scale multi-sample study. Hum. Brain Mapp. 40, 944–954. PubMed PMC

Pomponio R, Erus G, Habes M, et al., 2020. Harmonization of large MRI datasets for the analysis of brain imaging patterns throughout the lifespan. Neuroimage 208, 116450. PubMed PMC

Power JD, Cohen AL, Nelson SM, et al., 2011. Functional network organization of the human brain. Neuron 72, 665–678. PubMed PMC

Radua J, Vieta E, Shinohara R, et al., 2020. Increased power by harmonizing structural MRI site differences with the ComBat batch adjustment method in ENIGMA. Neuroimage 218, 116956. PubMed PMC

Renteria ME, Schmaal L, Hibar DP, et al., 2017. Subcortical brain structure and suicidal behaviour in major depressive disorder: a meta-analysis from the ENIGMA-MDD working group. Transl. Psychiatry 7, e1116. PubMed PMC

Rozycki M, Satterthwaite TD, Koutsouleris N, et al., 2018. Multisite machine learning analysis provides a robust structural imaging signature of schizophrenia detectable across diverse patient populations and within individuals. Schizophr. Bull. 44, 1035–1044. PubMed PMC

Sheela KGDSN, 2013. Review on methods to fix number of hidden neurons in neural networks. Mathematical Problems in Engineering.

Sheynin J, Duval ER, Lokshina Y, et al., 2020. Altered resting-state functional connectivity in adolescents is associated with PTSD symptoms and trauma exposure. Neuroimage Clin. 26, 102215. PubMed PMC

Sheynin S, Wolf L, Ben-Zion Z, et al., 2021. Deep learning model of fMRI connectivity predicts PTSD symptom trajectories in recent trauma survivors. Neuroimage 238, 118242. PubMed PMC

Siehl S, Wicking M, Pohlack S, et al., 2020. Structural white and gray matter differences in a large sample of patients with Posttraumatic Stress Disorder and a healthy and trauma-exposed control group: diffusion tensor imaging and region-based morphometry. Neuroimage Clin 28, 102424. PubMed PMC

Singh A, Sengupta S, Lakshminarayanan V, 2020. Explainable deep learning models in medical image analysis. J. Imaging 6, 52. PubMed PMC

Skatun KC, Kaufmann T, Doan NT, et al., 2017. Consistent functional connectivity alterations in schizophrenia spectrum disorder: a multisite study. Schizophr. Bull. 43, 914–924. PubMed PMC

Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R, 2014. Dropout: a simple way to prevent neural networks from overfitting. J. Machine Learn. Res. 15, 1929–1958.

Steardo L Jr., Carbone EA, de Filippis R, et al., 2020. Application of support vector machine on fmri data as biomarkers in schizophrenia diagnosis: a systematic review. Front. Psychiatry 11, 588. PubMed PMC

Suk HI, Lee SW, Shen D, et al., 2015. Latent feature representation with stacked auto-encoder for AD/MCI diagnosis. Brain Struct. Funct. 220, 841–859. PubMed PMC

Sumpter RE, McMillan TM, 2005. Misdiagnosis of post-traumatic stress disorder following severe traumatic brain injury. Br. J. Psychiatry 186, 423–426. PubMed

van Loo HM, de Jonge P, Romeijn JW, et al., 2012. Data-driven subtypes of major depressive disorder: a systematic review. BMC Med. 10, 156. PubMed PMC

Varoquaux G, 2018. Cross-validation failure: small sample sizes lead to large error bars. Neuroimage 180, 68–77. PubMed

Wang X, Xie H, Chen T, et al., 2021. Cortical volume abnormalities in posttraumatic stress disorder: an ENIGMA-psychiatric genomics consortium PTSD workgroup mega-analysis. Mol. Psychiatry 26, 4331–4343. PubMed PMC

Weng Y, Qi R, Zhang L, et al., 2019. Disturbed effective connectivity patterns in an intrinsic triple network model are associated with posttraumatic stress disorder. Neurol Sci 40, 339–349. PubMed

Woo CW, Chang LJ, Lindquist MA, et al., 2017. Building better biomarkers: brain models in translational neuroimaging. Nat. Neurosci. 20, 365–377. PubMed PMC

Ying X, 2019. An Overview of Overfitting and Its Solutions. IOP Publishing.

Zhao Y, Dong Q, Chen H, et al., 2017. Constructing fine-granularity functional brain network atlases via deep convolutional autoencoder. Med. Image Anal. 42, 200–211. PubMed PMC

Zilcha-Mano S, Zhu X, Suarez-Jimenez B, et al., 2020. Diagnostic and predictive neuroimaging biomarkers for posttraumatic stress disorder. Biol. Psychiatry Cogn. Neurosci. Neuroimag. 5, 688–696. PubMed PMC