An Imaging Risk Score for Bipolar Disorder: Multimodal Comparison and Integration with Polygenic Risk
- 1Centre for Precision Psychiatry, Division of Mental Health and Addiction, Oslo University Hospital & University of Oslo, Oslo, Norway
Abstract
BackgroundBiological risk scores provide objective markers of bipolar disorder (BD) liability. Large-scale genome-wide association studies (GWASs) have led to polygenic risk scores (PRSs) that capture genetic liability to BD at the individual level. We lack a comparable large-scale effort to leverage a brain-wide association study (BWAS) to create an imaging risk score (IRS) that captures brain structural liability to BD at the individual-level. Here, we conduct the largest multivariate BWAS of BD and develop a BD IRS. We compare the BD IRS to the PRS and assess their combined predictive performance in multimodal risk score (MRS) models.
MethodsOur multivariate BD BWAS used 153 structural neuroimaging features from 3902 cases and 6855 controls (mean age (sd)=35.47 years (13.42), 56.70% female) in a cross-validated machine learning approach. Using the resulting model coefficients, we constructed an IRS. We then evaluated the out-of-sample performance of the IRS, PRS derived from the largest GWAS to date, and combined MRS model using logistic regression in a subset of individuals with imaging and genetic data (nBD=486, ncontrols=1701, age (sd)=33.10 (12.02), 56.47% female). We also performed stratified analyses for demographic and BD clinical groups. Prediction performance was quantified using the area under the receiver operating characteristic curve (AUC). Subsequent analyses examined associations between each risk score and BD relevant symptoms in developing youth.
OutcomesIn predicting case status, the BD BWAS model achieved a mean AUCBD of 0.62 (sdBD=0.08) and in stratified analyses, had higher prediction than subgroup-specific BWAS models. The BD type II BWAS was the only exception (AUCBDII=0.60, sdBDII=0.07 vs AUCBD=0.58, sdBD=0.09). The BD PRS had an AUCBD of 0.65 (CI95%=0.63-0.68) and a higher AUC than subtype-specific PRS in stratified analyses. The PRS typically attained a higher AUC than the IRS. Notably, the IRS and PRS were uncorrelated (r=0.03, p=0.11). MRS models outperformed the prediction of individual PRS and IRS with the strongest prediction for BD type I (AUCBD=0.74, CI95%=0.70-0.78). The IRS had the strongest associations with BD-related symptoms in developing youth including increased energy (OR=1.21, pfdr=0.03) and distress from psychotic-like experiences (OR=1.23, pfdr=0.04).
InterpretationOur BD IRS is a novel biological indicator of BD risk. The lack of correlation between the IRS and PRS along with improved MRS prediction suggests the two risk scores capture some orthogonal liability to BD. We show that integrating imaging and genetics is useful for biological characterization of BD and should be considered for the development of clinically relevant prediction models.
Related
- Mapping the natural course of comorbid bipolar disorder via integration of multimodal data (MapCoBD): a study protocol.
- The molecular and neurochemical basis of altered intrinsic neural timescales in major depressive disorder: a multimodal imaging-transcriptomic investigation of esketamine and sertraline treatment.
- Polygenic Risk Profiles and the Manifestations of Alcohol Use Disorder and Related Outcomes
- Attention‐deficit/hyperactivity disorder polygenic risk score and incidence of cognitive impairment in older adults: A prospective multicohort study
- Genetic Correlates of Treatment Utilization for Alcohol Use Disorder: A Family-Based and Polygenic Investigation
BibTeX
@misc{parker2026imaging,
title = {{An Imaging Risk Score for Bipolar Disorder: Multimodal Comparison and Integration with Polygenic Risk}},
author = {N. Parker and K. S. O'Connell and A. A. Shadrin and P. Parekh and M. J. Y. Kang and J. Kopal and E. Hagen and K. G. Kennedy and Y. Im and D. Alnaes and D. Hughes and M. Alda and N. Alexander and L. Altegoer and A. Aquino-Servin and J. Bauer and D. Bedir and F. Benedetti and K. Berger and M. Berk and H. P. Blumberg and T. Borgers and K. Bozaoglu and B. Bravi and G. F. Busatto and E. J. Canales and D. M. Cannon and X. Caseras and I.-J. Chi and P. F. Claramonte and B. Clementz and L. Colic and E. Corley and O. Dandash and U. Dannlowski and C. Demro and A. M. Diaz-Zuluaga and K. E. Brand and T. Elvsashagen and L. T. Eyler and P. Favre and K. Flinkenflugel and M. F. Forte and Fuller},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.13.26365416},
url = {https://www.medrxiv.org/content/10.64898/2026.10.13.26365416v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - An Imaging Risk Score for Bipolar Disorder: Multimodal Comparison and Integration with Polygenic Risk
AU - N. Parker
AU - K. S. O'Connell
AU - A. A. Shadrin
AU - P. Parekh
AU - M. J. Y. Kang
AU - J. Kopal
AU - E. Hagen
AU - K. G. Kennedy
AU - Y. Im
AU - D. Alnaes
AU - D. Hughes
AU - M. Alda
AU - N. Alexander
AU - L. Altegoer
AU - A. Aquino-Servin
AU - J. Bauer
AU - D. Bedir
AU - F. Benedetti
AU - K. Berger
AU - M. Berk
AU - H. P. Blumberg
AU - T. Borgers
AU - K. Bozaoglu
AU - B. Bravi
AU - G. F. Busatto
AU - E. J. Canales
AU - D. M. Cannon
AU - X. Caseras
AU - I.-J. Chi
AU - P. F. Claramonte
AU - B. Clementz
AU - L. Colic
AU - E. Corley
AU - O. Dandash
AU - U. Dannlowski
AU - C. Demro
AU - A. M. Diaz-Zuluaga
AU - K. E. Brand
AU - T. Elvsashagen
AU - L. T. Eyler
AU - P. Favre
AU - K. Flinkenflugel
AU - M. F. Forte
AU - Fuller
PY - 2026
DA - 2026/10/13
PB - medRxiv
DO - 10.64898/2026.10.13.26365416
UR - https://www.medrxiv.org/content/10.64898/2026.10.13.26365416v1
AB - BackgroundBiological risk scores provide objective markers of bipolar disorder (BD) liability. Large-scale genome-wide association studies (GWASs) have led to polygenic risk scores (PRSs) that capture genetic liability to BD at the individual level. We lack a comparable large-scale effort to leverage a brain-wide association study (BWAS) to create an imaging risk score (IRS) that captures brain structural liability to BD at the individual-level. Here, we conduct the largest multivariate BWAS of BD and develop a BD IRS. We compare the BD IRS to the PRS and assess their combined predictive performance in multimodal risk score (MRS) models. MethodsOur multivariate BD BWAS used 153 structural neuroimaging features from 3902 cases and 6855 controls (mean age (sd)=35.47 years (13.42), 56.70% female) in a cross-validated machine learning approach. Using the resulting model coefficients, we constructed an IRS. We then evaluated the out-of-sample performance of the IRS, PRS derived from the largest GWAS to date, and combined MRS model using logistic regression in a subset of individuals with imaging and genetic data (nBD=486, ncontrols=1701, age (sd)=33.10 (12.02), 56.47% female). We also performed stratified analyses for demographic and BD clinical groups. Prediction performance was quantified using the area under the receiver operating characteristic curve (AUC). Subsequent analyses examined associations between each risk score and BD relevant symptoms in developing youth. OutcomesIn predicting case status, the BD BWAS model achieved a mean AUCBD of 0.62 (sdBD=0.08) and in stratified analyses, had higher prediction than subgroup-specific BWAS models. The BD type II BWAS was the only exception (AUCBDII=0.60, sdBDII=0.07 vs AUCBD=0.58, sdBD=0.09). The BD PRS had an AUCBD of 0.65 (CI95%=0.63-0.68) and a higher AUC than subtype-specific PRS in stratified analyses. The PRS typically attained a higher AUC than the IRS. Notably, the IRS and PRS were uncorrelated (r=0.03, p=0.11). MRS models outperformed the prediction of individual PRS and IRS with the strongest prediction for BD type I (AUCBD=0.74, CI95%=0.70-0.78). The IRS had the strongest associations with BD-related symptoms in developing youth including increased energy (OR=1.21, pfdr=0.03) and distress from psychotic-like experiences (OR=1.23, pfdr=0.04). InterpretationOur BD IRS is a novel biological indicator of BD risk. The lack of correlation between the IRS and PRS along with improved MRS prediction suggests the two risk scores capture some orthogonal liability to BD. We show that integrating imaging and genetics is useful for biological characterization of BD and should be considered for the development of clinically relevant prediction models.
ER -