Multimodal Prediction of Bipolar Disorder using Neuroimaging and Polygenic Risk Scores
Poster No:
474
Submission Type:
Abstract Submission
Authors:
Nadine Parker1, Kevin O'Connell1, Alexey Shadrin1, Pravesh Parekh1, Espen Hagen1, Jakub Kopal1, Dag Alnæs1, Melody Kang2, Yanghee Im3, Xavier Caseras4, Udo Dannlowski5, Lisa Eyler6, Bartholomeus Haarman7, Tomas Hajek8, Josselin Houenou9, Fleur Howells10, Matthew Kempton11, Tilo Kircher12, Luisa Klahn13, Mikael Landén13, Philip Mitchell14, Edith Pomarol-Clotet15, Joaquim Radua16, Jair Soares17, Scott Sponheim18, Henk Temmingh10, Eduard Vieta15, Sophia Thomopoulos19, Paul Thompson19, Lars Westlye1, Chris Ching2, Anders Dale6, Ole Andreassen1
Institutions:
1University of Oslo, Oslo, Norway, 2University of Southern California, California, United States, 3USC, Los Angeles, United States, 4Cardiff University, Cardiff, United Kingdom, 5Institute for Translational Psychiatry, Münster, Germany, 6University of California San Diego, San Diego, United States, 7University of Groningen, Groningen, Netherlands, 8Dalhousie University, Halifax, Canada, 9Université Paris-Saclay, Paris, France, 10University of Cape Town, Cape Town, South Africa, 11King's College London, London, United Kingdom, 12Department of Psychiatry and Psychotherapy, University of Marburg, Marburg, Germany, 13University of Gothenburg, Gothenburg, Sweden, 14University of New South Wales, Sydney, Australia, 15Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Barcelona, Spain, 16Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain, 17UT Health Science Center at Houston, Houston, United States, 18University of Minnesota, Minneapolis, United States, 19University of Southern California, Los Angeles, United States
First Author:
Co-Author(s):
Edith Pomarol-Clotet
Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM)
Barcelona, Spain
Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM)
Barcelona, Spain
Introduction:
Compared to healthy controls, variations in brain structure and genetics have been associated with bipolar disorder (BD) (Hibar et al., 2016, 2018; Mullins et al., 2021; O'Connell et al., 2024). Previous studies have used either brain structural features or polygenic risk scores (PRS) to predict BD, but little is known about their combined predictive ability (Mullins et al., 2021; Nunes et al., 2020; O'Connell et al., 2024). Here, we develop an imaging risk score (IRS) and determine if there is added predictive value for BD when combined with a PRS for BD.
Methods:
For development of the IRS, a total of 3,653 participants [58.86% female, mean age(sd)=34.59 (13.50)] across 16 cohorts from the Enhancing Neuroimaging Genetics through Meta-Analysis bipolar disorder (ENIGMA-BD) working group were included. This sample consisted of 1,369 BD cases, of whom, 1,012 were diagnosed as BD Type 1 (BD1) and 300 as BD Type 2 (BD2). Each participant's MRI scan was processed using the ENIGMA-standard FreeSurfer and visual quality control procedure and we used 157 imaging features for development of the IRS (bilateral regional cortical thickness, regional cortical surface areas, and subcortical volumes) (Desikan et al., 2006). A 5-fold cross-validation framework was used to select a model (GLM, SVM, LASSO, RIDGE, and E-Net) that best predicts BD, BD1, and BD2 using the 157 features. Each imaging feature was pre-residualized for age, age2, sex, and scanner. Model selection was based on the area under the receiver operator curve (AUC). Coefficients from the best performing model were then used as weights to derive an IRS (i.e., a weighted sum across the 157 features). We generated an IRS in a held-out sample of 1,012 participants [48.02% female, mean age(sd)=32.76 (10.13)] with 266 BD cases (nBD1=151, nBD2=102) and 746 controls. Also, in the held-out sample, a PRS was generated for BD, BD1, and BD2 using the latest Psychiatric Genomics Consortium GWAS of BD and the tool LDpred2 (O'Connell et al., 2024; Privé et al., 2020). For direct comparison of IRS and PRS performance, we pre-residualize the PRS for age, age2, sex, genetic batch, and the first 20 genetic principal components before estimating the prediction of BD. Finally, we combine IRS and PRS in linear models to assess the multimodal prediction of BD.
Results:
The best performing model for BD and BD1 was RIDGE regression (AUC-BD=0.61, AUC-BD1=0.64; Figure 1). For BD2, a non-regularized logistic GLM performed best (AUC-BD2=0.54). In the held-out sample, the IRS for BD1 had the best prediction performance (AUC-BD=0.61, AUC-BD1=0.62, AUC-BD2=0.51). For BD and each subtype, the PRS performed better than the IRS (Figure 1). Prediction of BD2 was lowest for both PRS and IRS when compared to BD and BD1. When combining IRS and PRS for multimodal prediction of BD, we observed an increase in AUC above that of the PRS prediction for BD and BD1 but not BD2 (AUC-BD=0.71, AUC-BD1=0.73, AUC-BD2=0.62).
·Figure 1. Performance of Imaging and Polygenic Risk Scores. A depiction of the methodological approach and prediction performance. BD: bipolar disorder, BD1: bipolar disorder type 1, BD2: bipolar diso
Conclusions:
Both brain structural and genetic variations can be leveraged to predict BD. With continued development, the combination of imaging and genetics for multimodal prediction may form a basis to develop clinically useful tools for brain disorders.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Genetics:
Genetic Association Studies 2
Modeling and Analysis Methods:
Classification and Predictive Modeling
Keywords:
Other - Biopolar Disorder, Imaging-Genetics, Multimodal Prediction
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?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
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.
Hibar, D. P., Westlye, L. T., Doan, N. T., … Andreassen, O. A. (2018). Cortical abnormalities in bipolar disorder: An MRI analysis of 6503 individuals from the ENIGMA Bipolar Disorder Working Group. Molecular Psychiatry, 23(4), 932–942. https://doi.org/10.1038/mp.2017.73
Hibar, D. P., Westlye, L. T., van Erp, … Andreassen, O. A. (2016). Subcortical volumetric abnormalities in bipolar disorder. Molecular Psychiatry, 21(12), Article 12. https://doi.org/10.1038/mp.2015.227
Mullins, N., Forstner, A. J., O’Connell, K. S., … Andreassen, O. A. (2021). Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nature Genetics, 53(6), Article 6. https://doi.org/10.1038/s41588-021-00857-4
Nunes, A., Schnack, H. G., Ching, C. R. K., … Hajek, T. (2020). Using structural MRI to identify bipolar disorders – 13 site machine learning study in 3020 individuals from the ENIGMA Bipolar Disorders Working Group. Molecular Psychiatry, 25(9), Article 9. https://doi.org/10.1038/s41380-018-0228-9
O’Connell, K. S., Koromina, M., Veen, T. van der, … Andreassen, O. A. (2024). Genomics yields biological and phenotypic insights into bipolar disorder (p. 2023.10.07.23296687). medRxiv. https://doi.org/10.1101/2023.10.07.23296687
Privé, F., Arbel, J., & Vilhjálmsson, B. J. (2020). LDpred2: Better, faster, stronger. Bioinformatics, 36(22–23), 5424–5431. https://doi.org/10.1093/bioinformatics/btaa1029