Neural network architectures for Schizophrenia patients-versus-Controls classification based on Amygdala connectivity

Abstract

A few recent neuroimaging studies reported the role of amygdala connectivity in patients with schizophrenia. However, thus far in the fMRI literature, the predictive capability of amygdala connectivity in classifying schizophrenia patients and controls has not been explored using advanced machine learning techniques. In this brief report, we present results from analysis utilizing classification methods based on deep neural networks and convolutional neural networks for predicting schizophrenia versus healthy control using amygdala’s connectivity to other brain regions. Median accuracy rates of 62.9%, 60% and 60% were obtained for classification based on a deep neural network, convolutional neural network and ResNet34 architectures, respectively.

References

[1] Meisner OC, Nair A, Chang SWC. Amygdala connectivity and implications for social cognition and disorders. Handbook Clin Neurol. 2022;187:381-403. doi: 10.1016/B978-0-12-823493-8.00017-1.

[2] Kim MJ, Loucks RA, Palmer AL, Brown AC, Solomon KM, Marchante AN, Whalen PJ. The structural and functional connectivity of the amygdala: from normal emotion to pathological anxiety. Behav Brain Res. 2011 Oct 1;223(2):403-10. doi: 10.1016/j.bbr.2011.04.025.

[3] Rasetti R, Mattay VS, Wiedholz LM, Kolachana BS, Hariri AR, Callicott JH, Meyer-Lindenberg A, Weinberger DR. Evidence that altered amygdala activity in schizophrenia is related to clinical state and not genetic risk. Am J Psychiatry. 2009 Feb;166(2):216-25. doi: 10.1176/appi.ajp.2008.08020261.

[4] Bogerts, B., Lieberman, J. A., Ashtari, M., Bilder, R. M., Degreef, G., Lerner, G., Johns, C., & Masiar, S. (1993). Hippocampus-amygdala volumes and psychopathology in chronic schizophrenia. Biological Psychiatry, 33(4), 236–246. https://doi.org/10.1016/0006-3223(93)90289-P

[5] Lawrie SM, Abukmeil SS. Brain abnormality in schizophrenia. A systematic and quantitative review of volumetric magnetic resonance imaging studies. Br J Psychiatry. 1998 Feb;172:110-20. doi: 10.1192/bjp.172.2.110. PMID: 9519062.

[6] Nelson MD, Saykin AJ, Flashman LA, Riordan HJ. Hippocampal volume reduction in schizophrenia as assessed by magnetic resonance imaging: a meta-analytic study. Arch Gen Psychiatry. 1998 May;55(5):433-40. doi: 10.1001/archpsyc.55.5.433. PMID: 9596046.

[7] Wright IC, Rabe-Hesketh S, Woodruff PW, David AS, Murray RM, Bullmore ET. Meta-analysis of regional brain volumes in schizophrenia. Am J Psychiatry. 2000 Jan;157(1):16-25. doi: 10.1176/ajp.157.1.16. PMID: 10618008.

[8] Aleman A, Kahn RS. Strange feelings: do amygdala abnormalities dysregulate the emotional brain in schizophrenia? Prog Neurobiol. 2005 Dec;77(5):283-98. doi: 10.1016/j.pneurobio.2005.11.005. Epub 2005 Dec 13. PMID: 16352388.

[9] Mukherjee P, Whalley HC, McKirdy JW, Sprengelmeyer R, Young AW, McIntosh AM, Lawrie SM, Hall J. Altered amygdala connectivity within the social brain in schizophrenia. Schizophr Bull. 2014 Jan;40(1):152-60. doi: 10.1093/schbul/sbt086.

[10] Kim WS, Shen G, Liu C, Kang NI, Lee KH, Sui J, Chung YC. Altered amygdala-based functional connectivity in individuals with attenuated psychosis syndrome and first-episode schizophrenia. Sci Rep. 2020 Oct 19;10(1):17711. doi: 10.1038/s41598-020-74771-w.

[11] Wang X, Yin Z, Sun Q, Jiang X, Chao L, Dai X, Tang Y. Comparative Study on the Functional Connectivity of Amygdala and Hippocampal Neural Circuits in Patients With First-Episode Schizophrenia and Other High-Risk Populations. Front Psychiatry. 2021 Sep 1;12:627198. doi: 10.3389/fpsyt.2021.627198.

[12] Guo H, Ye H, Li Z, Li X, Huang W, Yang Y, Xie G, Xu C, Li X, Liang W, Jing H, Zhang C, Tang C, Liang J. Amygdala signal abnormality and cognitive impairment in drug-naïve schizophrenia. BMC Psychiatry. 2023 Apr 5;23(1):231. doi: 10.1186/s12888-023-04728-6.

[13] Zheng G, Zhou Y, Zhou J, Liang S, Li X, Xu C, Xie G, Liang J. Abnormalities of the Amygdala in schizophrenia: a real world study. BMC Psychiatry. 2023 Aug 22;23(1):615. doi: 10.1186/s12888-023-05031-0.

[14] Feola B, McHugo M, Armstrong K, Noall MP, Flook EA, Woodward ND, Heckers S, Blackford JU. BNST and amygdala connectivity are altered during threat anticipation in schizophrenia. Behav Brain Res. 2021 Aug 27;412:113428. doi: 10.1016/j.bbr.2021.113428.

[15] Ford JM, Palzes VA, Roach BJ, Potkin SG, van Erp TG, Turner JA, Mueller BA, Calhoun VD, Voyvodic J, Belger A, Bustillo J, Vaidya JG, Preda A, McEwen SC; Functional Imaging Biomedical Informatics Research Network; Mathalon DH. Visual hallucinations are associated with hyperconnectivity between the amygdala and visual cortex in people with a diagnosis of schizophrenia. Schizophr Bull. 2015 Jan;41(1):223-32. doi: 10.1093/schbul/sbu031.

[16] Çetin MS, Christensen F, Abbott CC, Stephen JM, Mayer AR, Cañive JM, Bustillo JR, Pearlson GD, Calhoun VD. Thalamus and posterior temporal lobe show greater inter-network connectivity at rest and across sensory paradigms in schizophrenia. Neuroimage. 2014 Aug 15;97:117-26. doi: 10.1016/j.neuroimage.2014.04.009.

[17] Kiparizoska S, Ikuta T. Disrupted Olfactory Integration in Schizophrenia: Functional Connectivity Study. Int J Neuropsychopharmacol. 2017 Sep 1;20(9):740-746. doi: 10.1093/ijnp/pyx045.

[18] Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage. 2012 Feb 1;59(3):2142-54. doi: 10.1016/j.neuroimage.2011.10.018. Epub 2011 Oct 14. Erratum in: Neuroimage. 2012 Nov 1;63(2):999. PMID: 22019881; PMCID: PMC3254728.

[19] Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage. 2014 Jan 1;84:320-41. doi: 10.1016/j.neuroimage.2013.08.048. Epub 2013 Aug 29. PMID: 23994314; PMCID: PMC3849338.

[20] Siegel JS, Power JD, Dubis JW, Vogel AC, Church JA, Schlaggar BL, Petersen SE. Statistical improvements in functional magnetic resonance imaging analyses produced by censoring high-motion data points. Hum Brain Mapp. 2014 May;35(5):1981-96. doi: 10.1002/hbm.22307. Epub 2013 Jul 17. PMID: 23861343; PMCID: PMC3895106.

[21] Makris N, Kaiser J, Haselgrove C, Seidman LJ, Biederman J, Boriel D, Valera EM, Papadimitriou GM, Fischl B, Caviness VS, Kennedy DN, 2006. Human cerebral cortex: a system for the integration of volume and surface-based representations. NeuroImage 33, 139–153.

[22] Geron A. Hands-On Machine Learning with Scikit-Learn & TensorFlow. 2nd ed: O-Reilly Media Inc; 2019.