Convolutional Neural Network–Based Deep Learning Model for Predicting Differential Suicidality in Depressive Patients Using Brain Generalized q-Sampling Imaging

Roger S. McIntyre, MD^f,g,h; Vincent Chin-Hung Chen, MD, PhD^a,b; Jun-Cheng Weng, PhD^b,c,i; Fu-Te Wong, MSc; Yuan-Hsiung Tsai, MD, PhD^a,d; Man Teng Cheok, BS^c; Yi-Peng Eve Chang, MS^e

J Clin Psychiatry 2021;82(2):19m13225

ABSTRACT

Objective: Suicide is a priority health problem. Suicide assessment depends on imperfect clinician assessment with minimal ability to predict the risk of suicide. Machine learning/deep learning provides an opportunity to detect an individual at risk of suicide to a greater extent than clinician assessment. The present study aimed to use deep learning of structural magnetic resonance imaging (MRI) to create an algorithm for detecting suicidal ideation and suicidal attempts.

Methods: We recruited 4 groups comprising a total of 186 participants: 33 depressive patients with suicide attempt (SA), 41 depressive patients with suicidal ideation (SI), 54 depressive patients without suicidal thoughts (DP), and 58 healthy controls (HCs). The confirmation of depressive disorder, SA and SI was based on psychiatrists’ diagnosis and Mini-International Neuropsychiatric Interview (MINI) interviews. In the generalized q-sampling imaging (GQI) dataset, indices of generalized fractional anisotropy (GFA), the isotropic value of the orientation distribution function (ISO), and normalized quantitative anisotropy (NQA) were separately trained in convolutional neural network (CNN)–based deep learning and DenseNet models.

Results: From the results of 5-fold cross-validation, the best accuracies of the CNN classifier for predicting SA, SI, and DP against HCs were 0.916, 0.792, and 0.589, respectively. In SA-ISO, DenseNet outperformed the simple CNNs with a best accuracy from 5-fold cross-validation of 0.937. In SA-NQA, the best accuracy was 0.915.

Conclusions: The results showed that a deep learning method based on structural MRI can effectively detect individuals at different levels of suicide risk, from depression to suicidal ideation and attempted suicide. Further studies from different populations, larger sample sizes, and prospective follow-up studies are warranted to confirm the utility of deep learning methods for suicide prevention and intervention.

Free Access: Please Log In

This content is completely free—but you need to be logged in to read the full article. If you already have an account, please log in below. Otherwise, register for free to unlock instant access.

References (18)

Preventing suicide: A global imperative. World Health Organization. 2014. https://www.who.int/mental_health/suicide-prevention/world_report_2014/en/
Sveticic J, De Leo D. The hypothesis of a continuum in suicidality: a discussion on its validity and practical implications. Ment Illn. 2012;4(2):e15. PubMed NLM
Belsher BE, Smolenski DJ, Pruitt LD, et al. Prediction models for suicide attempts and deaths: a systematic review and simulation. JAMA Psychiatry. 2019;76(6):642–651. PubMed CrossRef NLM
Chung DT, Ryan CJ, Hadzi-Pavlovic D, et al. Suicide rates after discharge from psychiatric facilities: a systematic review and meta-analysis. JAMA Psychiatry. 2017;74(7):694–702. PubMed CrossRef NLM
Bani-Fatemi A, Tasmim S, Graff-Guerrero A, et al. Structural and functional alterations of the suicidal brain: an updated review of neuroimaging studies. Psychiatry Res Neuroimaging. 2018;278:77–91. PubMed CrossRef NLM
Lee Y, Ragguett RM, Mansur RB, et al. Applications of machine learning algorithms to predict therapeutic outcomes in depression: a meta-analysis and systematic review. J Affect Disord. 2018;241:519–532. PubMed CrossRef NLM
Just MA, Pan L, Cherkassky VL, et al. Machine learning of neural representations of suicide and emotion concepts identifies suicidal youth. Nat Hum Behav. 2017;1(12):911–919. PubMed CrossRef NLM
Sheehan DV, Lecrubier Y, Sheehan KH, et al. The Mini-International Neuropsychiatric Interview (MINI): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59(suppl 20):22–33, quiz 34–57. PubMed NLM
Yeh FC, Wedeen VJ, Tseng WY. Generalized q-sampling imaging. IEEE Trans Med Imaging. 2010;29(9):1626–1635. PubMed CrossRef NLM
Abadi M, Agarwal A, Barham P, et al. TensorFlow: large-scale machine learning on heterogeneous distributed systems. Preliminary White Paper: November 9, 2015. https://arxiv.org/pdf/1603.04467.pdf.
Huang G, Liu Z, van der Maaten L, et al. Densely Connected Convolutional Networks. Cornell University. 2017. http://arxivorg/abs/1608.06993
Gosnell SN, Fowler JC, Salas R. Classifying suicidal behavior with resting-state functional connectivity and structural neuroimaging. Acta Psychiatr Scand. 2019;140(1):20–29. PubMed CrossRef NLM
Yeh FC, Verstynen TD, Wang Y, et al. Deterministic diffusion fiber tracking improved by quantitative anisotropy. PLoS One. 2013;8(11):e80713. PubMed CrossRef NLM
Serre T, Wolf L, Bileschi S, et al. Robust object recognition with cortex-like mechanisms. IEEE Trans Pattern Anal Mach Intell. 2007;29(3):411–426. PubMed CrossRef NLM
Bergstra J, Bengio Y. Random search for hyper-parameter optimization. J Mach Learn Res. 2012;13:281–305.
Aydore S, Thirion B, Varoquaux G. Feature grouping as a stochastic regularizer for high-dimensional structured data. Cornell University. 2019. https://arxiv.org/abs/1807.11718
Hoyos-Idrobo A, Varoquaux G, Kahn J, et al. Recursive Nearest Agglomeration (ReNA): fast clustering for approximation of structured signals. IEEE Trans Pattern Anal Mach Intell. 2019;41(3):669–681. PubMed CrossRef NLM
Rubinstein DH. A stress-diathesis theory of suicide. Suicide Life Threat Behav. 1986;16(2):182–197. PubMed CrossRef NLM