Multimodal Neuroimaging Predictive Models of Cognition Generalize from Healthy Populations to ADHD
Presented During:
Thursday, June 26, 2025: 11:30 AM - 12:45 PM
Brisbane Convention & Exhibition Centre
Room:
M3 (Mezzanine Level)
Poster No:
764
Submission Type:
Abstract Submission
Authors:
Farzane Lal Khakpoor1, Yue Wang1, William van der Vliet1, Alina Tetereva2, Narun Pat1
Institutions:
1University of Otago, Dunedin, Otago, 2New Zealand Brain Research Institute, Christchurch, Canterbury
First Author:
Co-Author(s):
Introduction:
Integrating neuroimaging data with advanced predictive modeling offers a promising avenue for developing predictive biomarkers for complex traits, such as cognitive functioning (Tetereva et al., 2022). However, the predictive utility of these biomarkers has been predominantly tested on healthy participants (Rasero et al., 2021; Sripada et al., 2020; Tetereva et al., 2022, 2024), leaving it unclear how applicable they are to individuals with neurocognitive conditions, such as Attention-Deficit/Hyperactivity Disorders (ADHD). Addressing this gap brings us closer to biomarkers that can be clinically utilized to trace brain signals related to cognitive abilities during the progression or treatment of ADHD.
Methods:
As a proof of concept, we developed and validated predictive models for cognitive functioning using nested leave-one-site-out cross-validation on data from over 11,000 children in the Adolescent Brain Cognitive Development (ABCD) dataset. We further tested their predictive abilities on children without ADHD (n = 10,708) and with ADHD (n ranging from 61 to 1,034, depending on definitions). Our models leveraged 81 neuroimaging modalities, including cortical thickness, functional connectivity, diffusion metrics, and task-fMRI contrasts. To account for the cumulative effects of brain features recommended for the prediction of higher-level brain functions and related health conditions (Engemann et al., 2020; Mooney et al., 2024), predictions were generated using a two-layer stacking framework. This framework combined Elastic Net regression for single modalities (81 models) and Random Forest for stacked modalities (15 models). Generalizability was then assessed on an independent dataset of 79 children (35 with ADHD). Performance was evaluated across ADHD and non-ADHD groups, with feature importance assessed using Elastic Net coefficients and SHAP values.
Results:
Our results revealed similar predictive abilities of the biomarkers across children with and without ADHD in the internal validation set. No significant differences in performance were observed between the non-ADHD and ADHD groups for the top-performing models and for most single and stacked models in general. The biomarker combining all neuroimaging modalities demonstrated the highest performance, with an out-of-sample correlation of approximately 0.57. Among the single-modality biomarkers, the best performer was the working memory N-back two-minus-zero-back contrast, which achieved an out-of-sample correlation of 0.48. A similar pattern was observed in the external validation, where six out of eight stacked models demonstrated comparable performance between the ADHD and non-ADHD groups. The best-performing biomarker in this validation was the stacked working memory task contrasts, with a correlation of approximately 0.44 and no significant difference in performance between the groups. In contrast, the single-modality models showed mixed results, with more than half exhibiting significantly different performance between the ADHD and non-ADHD groups. These differences may reflect variations in cognitive functioning measures, fMRI tasks, and MRI acquisition properties between the internal and external validation datasets.
Conclusions:
Our findings underscore the potential of multimodal neuroimaging-based predictive biomarkers for tracing cognitive functioning in children with ADHD. Moreover, they highlight the robustness of these biomarkers across datasets, despite differences in data acquisition pipelines.
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism)
Higher Cognitive Functions:
Higher Cognitive Functions Other 1
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Classification and Predictive Modeling 2
Connectivity (eg. functional, effective, structural)
Keywords:
Attention Deficit Disorder
Cognition
Computational Neuroscience
FUNCTIONAL MRI
Machine Learning
Modeling
STRUCTURAL MRI
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.
Jernigan, T. L., Brown, S. A., Dale, A. M., Tapert, S., Sowell, E. S., Herting, M., Laird, A., Gonzalez, R., Squeglia, L., Gray, K., Paulus, M. P., Aupperle, R., Feldstein Ewing, S. W., Nagel, B. J., Fair, D. A., Baker, F., Colrain, I. M., Bookheimer, S. Y., Dapretto, M., … Gee, D. (2023). Adolescent Brain Cognitive Development Study (ABCD) - Annual Release 5.1. ABCD Study.
Lytle, M. N., Hammer, R., & Booth, J. R. (2020). A neuroimaging dataset on working memory and reward processing in children with and without ADHD. Data in Brief, 31, 105801.
Mooney, M. A., Hermosillo, R. J. M., Feczko, E., Miranda-Dominguez, O., Moore, L. A., Perrone, A., Byington, N., Grimsrud, G., Rueter, A., Nousen, E., Antovich, D., Feldstein Ewing, S. W., Nagel, B. J., Nigg, J. T., & Fair, D. A. (2024). Cumulative Effects of Resting-State Connectivity Across All Brain Networks Significantly Correlate with Attention-Deficit Hyperactivity Disorder Symptoms [Article]. The Journal of Neuroscience, 44(10), e1202232023. https://doi.org/10.1523/JNEUROSCI.1202-23.2023
Rasero, J., Sentis, A. I., Yeh, F. C., Verstynen, T., & Robinson, E. C. (2021). Integrating across neuroimaging modalities boosts prediction accuracy of cognitive ability [Article]. PLoS Computational Biology, 17(3), e1008347–e1008347. https://doi.org/10.1371/journal.pcbi.1008347
Sripada, C., Angstadt, M., Rutherford, S., Taxali, A., & Shedden, K. (2020). Toward a “treadmill test” for cognition: Improved prediction of general cognitive ability from the task activated brain [Article]. Human Brain Mapping, 41(12), 3186–3197. https://doi.org/10.1002/hbm.25007
Tetereva, A., Knodt, A., Melzer, T., van der Vliet, W., Gibson, B., Hariri, A., Whitman, E., Li, J., Deng, J., Ireland, D., Ramrakha, S., & Pat, N. (preprint). Improving Predictability, Test-Retest Reliability and Generalisability of Brain-Wide Associations for Cognitive Abilities via Multimodal Stacking [Article]. bioRxiv, 2024-05.
Tetereva, A., Li, L., Deng, J., Stringaris, A., & Pat, N., (2022). Capturing Brain-Cognition Relationship: Integrating Task-Based fMRI Across Tasks Markedly Boosts Prediction and Test-Retest Reliability [Article]. Neuro.