Cross-Disorder Brain Vulnerability as Predictor of Treatment Response in First-Episode Schizophrenia
Poster No:
438
Submission Type:
Abstract Submission
Authors:
Yanmeng Huang1, Yongbin Wei1, Yingchan Wang2, Yong Liu1, Jijun Wang2
Institutions:
1School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China, 2Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine, Shanghai, China
First Author:
Yanmeng Huang
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
Co-Author(s):
Yongbin Wei
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
Yingchan Wang
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China
Yong Liu
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
School of Artificial Intelligence, Beijing University of Posts and Telecommunications
Beijing, China
Jijun Wang
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China
Shanghai Mental Health Center, Shanghai Jiaotong University School of Medicine
Shanghai, China
Introduction:
Schizophrenia (SCZ) is a complex psychiatric disorder with widespread brain morphological abnormalities. These morphological changes have been reliably delineated through large-scale, multi-site neuroimaging cohorts (Thompson, 2020; van Erp, 2018) and may correlate to symptoms and treatment responses (Kochunov, 2022). Recent transdiagnostic studies suggest the brain alterations in SCZ to overlap with other mental conditions (Chang, 2018). Here, we examined transdiagnostic similarity of brain structure in first-episode SCZ and demonstrated the morphological similarities as important predictors of treatment responses.
Methods:
T1-weighted magnetic resonance imaging (MRI) data of 197 subjects (incl. 97 healthy controls (HCs) and 100 first-episode SCZ) from the Shanghai Mental Health Center (SMHC) were used. Data were processed using FreeSurfer (version 6.0) (Fischl, 2012), with the longitudinal pipeline specifically used for SCZ (Reuter, 2012), yielding brain volumes of 14 subcortical regions and cortical thickness of 68 cortical regions from the Desikan-Killiany atlas (Desikan, 2006). Sex-specific CentileBrain normative models were used to compute z-scores delineating deviations from normative values (Ge, 2024). Correlating z-scores with ENIGMA-derived summary statistics for 8 disorders resulted in 16 transdiagnostic brain similarity scores per subject (Larivière, 2021).
Two-sample t-tests compared cross-disorder similarity metrics between SCZ and HC, and paired-sample t-tests evaluated longitudinal changes in SCZ. A nonlinear support vector machine (SVM) with a radial basis function (RBF) kernel classified SCZ and HC based on similarity metrics (5-fold cross-validation, 100 times). Partial least squares (PLS) analysis explored the correlation between brain morphological features and clinical indicators (positive and negative syndrome scale (PANSS) at baseline and follow-up, ΔPANSS, and PANSS change rate) (Krishnan, 2011). LASSO regression with nested cross-validation predicted treatment responses using the same metrics.
Two-sample t-tests compared cross-disorder similarity metrics between SCZ and HC, and paired-sample t-tests evaluated longitudinal changes in SCZ. A nonlinear support vector machine (SVM) with a radial basis function (RBF) kernel classified SCZ and HC based on similarity metrics (5-fold cross-validation, 100 times). Partial least squares (PLS) analysis explored the correlation between brain morphological features and clinical indicators (positive and negative syndrome scale (PANSS) at baseline and follow-up, ΔPANSS, and PANSS change rate) (Krishnan, 2011). LASSO regression with nested cross-validation predicted treatment responses using the same metrics.
Results:
SCZs had higher cross-disorder similarity in subcortical vulnerabilities to bipolar disorder (BD), major depressive disorder (MDD), and 22q11.2 deletion syndrome (22q) compared to HC at baseline and follow-up (p < 0.05, false discovery rate (FDR) corrected). Longitudinal comparisons showed increasing similarity over time, indicating progressive brain vulnerability changes. Similarly, cortical thickness vulnerabilities were greater in SCZs at both time points, with longitudinal analyses showing increased similarity to BD, MDD, and obsessive-compulsive disorder (OCD).
SVM models using combined subcortical and cortical similarities achieved high classification performance between SCZ and HC (AUC = 0.83 for baseline, 0.87 for follow-up) (Fig. 1A, 1B). Ablation experiments revealed subcortical similarity to 22q and cortical similarity to OCD and SCZ as key classification features (Fig. 1C), while brain regions, such as the bilateral caudate, right accumbens, and right rostral anterior cingulate gyrus contributed the most (Fig. 1D).
PLS analysis demonstrated a correlation between cross-disorder brain similarities and clinical variables (p = 0.014, 10000 permutations), with individualized similarity composite scores correlating with clinical composite scores (r = 0.27, p = 0.007) (Fig. 2B). Significant contributions were observed from similarity metrics for autism spectrum disorder (ASD), OCD, and SCZ (Fig. 2C), as well as baseline PANSS, ΔPANSS, and reduction ratio (Fig. 2D). Predictive modeling found subcortical similarity metrics significantly predicted ΔPANSS (r = 0.23, p = 0.021), while cortical metrics showed no significance (Fig. 2A).
SVM models using combined subcortical and cortical similarities achieved high classification performance between SCZ and HC (AUC = 0.83 for baseline, 0.87 for follow-up) (Fig. 1A, 1B). Ablation experiments revealed subcortical similarity to 22q and cortical similarity to OCD and SCZ as key classification features (Fig. 1C), while brain regions, such as the bilateral caudate, right accumbens, and right rostral anterior cingulate gyrus contributed the most (Fig. 1D).
PLS analysis demonstrated a correlation between cross-disorder brain similarities and clinical variables (p = 0.014, 10000 permutations), with individualized similarity composite scores correlating with clinical composite scores (r = 0.27, p = 0.007) (Fig. 2B). Significant contributions were observed from similarity metrics for autism spectrum disorder (ASD), OCD, and SCZ (Fig. 2C), as well as baseline PANSS, ΔPANSS, and reduction ratio (Fig. 2D). Predictive modeling found subcortical similarity metrics significantly predicted ΔPANSS (r = 0.23, p = 0.021), while cortical metrics showed no significance (Fig. 2A).
·ROC plot of SVM classification results for SCZ and HC (repeated 100 times) and contributions of features and brain regions.
·Results of prediction and PLS analysis.
Conclusions:
This study highlights the transdiagnostic similarity of brain morphological abnormalities in SCZ, which might play a role in the prediction of clinical outcomes of SCZ.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Subcortical Structures
Keywords:
MRI
Schizophrenia
Treatment
Other - Classification;Longitudinal
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Desikan, R. S., et al. (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage, 31(3), 968-980.
Fischl, B., et al. (2012). FreeSurfer. Neuroimage, 62(2), 774-781.
Ge, R., et al. (2024). Normative modelling of brain morphometry across the lifespan with CentileBrain: algorithm benchmarking and model optimisation. The Lancet Digital Health, 6(3), e211-e221.
Kochunov, P., et al. (2022). Translating ENIGMA schizophrenia findings using the regional vulnerability index: Association with cognition, symptoms, and disease trajectory. Hum Brain Mapp, 43(1), 566-575.
Krishnan, A., et al. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. Neuroimage, 56(2), 455-475.
Larivière, S., et al. (2021). The ENIGMA Toolbox: multiscale neural contextualization of multisite neuroimaging datasets. Nature Methods, 18(7), 698-700.
Reuter, M., et al. (2012). Within-subject template estimation for unbiased longitudinal image analysis. Neuroimage, 61(4), 1402-1418.
Thompson, P. M., et al. (2020). ENIGMA and global neuroscience: A decade of large-scale studies of the brain in health and disease across more than 40 countries. Translational Psychiatry, 10(1), 100.
van Erp, T. G. M., et al. (2018). Cortical Brain Abnormalities in 4474 Individuals With Schizophrenia and 5098 Control Subjects via the Enhancing Neuro Imaging Genetics Through Meta Analysis (ENIGMA) Consortium. Biol Psychiatry, 84(9), 644-654.