Identifying joint brain structure and function change using multimodal age prediction on large data
Poster No:
930
Submission Type:
Abstract Submission
Authors:
Yuhui Du1, Ruotong Li1, Vince Calhoun2
Institutions:
1Shanxi University, Taiyuan, Shanxi, 2GSU/GATech/Emory, Atlanta, GA
First Author:
Co-Author(s):
Introduction:
There has been evidence showing both structural and functional alterations in the aging brain (Cole, 2020; Fotiadis et al., 2024). Understanding the parallel evolution of structural and functional changes is essential for elucidating the mechanisms underlying cognitive aging and facilitating the development of interventions to mitigate age-related cognitive decline. However, the mechanisms underlying the joint changes in brain structure and function during the complex aging process remain largely unknown. In this study, we reveal the mechanisms underlying the joint changes in brain structure and function with aging using a large dataset from 27,796 healthy adults based on unimodal and multimodal age prediction models as well as a new multimodal joint analysis method.
Methods:
Brain functional network connectivity (FNC) and gray matter volume (GMV) of 27,796 healthy adults aged between 49 and 76 years from the UK Biobank are analyzed. Utilizing LASSO that can select important features while eliminating irrelevant or redundant ones (Xiong et al., 2023), we train age prediction models using FNC and GMV separately, employing a strictly unbiased cross-validation framework to identify reliable FNC and GMV associated with aging. Then, we validate these features using a multimodal age prediction model, and finally propose a new method to explore their joint changes along with the aging. In particular, to explore comprehensive associations between the multimodal features, FNC are divided into four patterns according to their strengths and correlations with age: age-positively-related positive (APRP) FNC, age-negatively-related positive (ANRP) FNC, age-positively-related negative (APRN) FNC, and age-negatively-related negative (ANRN) FNC, while GMV include two patterns: age-positively-related (AP) GMV and age-negatively-related (AN) GMV. Hence, we not only investigate significant synergistic decline changes between FNC and GMV along with the aging, but also explore their contradictory changes.
Results:
In the FNC prediction model, important FNC features were identified from different functional domains, with a greater number belonging to the ANRP and APRN patterns, indicating that brain functional interactions primarily show declined connectivity (Fig 1A-F). In the GMV prediction model, a widespread reduction in GMV is observed across the brain, with most regions showing symmetry, except for the thalamus and putamen. The GMV of the left and right thalamus negatively and positively correlate with chronological age, respectively, while the GMV of the left and right putamen show the opposite pattern. This suggests potential functional lateralization (Fig. 1G).
The multimodal prediction significantly outperforms the unimodal prediction (Fig. 2A), with prediction accuracy using GMV being better than that using FNC. In the multimodal joint change analysis (Fig. 2B), we revealed that in most areas, FNC and GMV decreased together, especially in the cerebellum and frontal regions. However, specific regions, such as the occipital pole, lateral occipital cortex, and frontal pole, exhibited increased FNC despite a reduction in GMV, suggesting a compensatory mechanism against brain aging. The left precentral gyrus showed complex changes, indicating both degeneration and compensatory responses in brain aging.
The multimodal prediction significantly outperforms the unimodal prediction (Fig. 2A), with prediction accuracy using GMV being better than that using FNC. In the multimodal joint change analysis (Fig. 2B), we revealed that in most areas, FNC and GMV decreased together, especially in the cerebellum and frontal regions. However, specific regions, such as the occipital pole, lateral occipital cortex, and frontal pole, exhibited increased FNC despite a reduction in GMV, suggesting a compensatory mechanism against brain aging. The left precentral gyrus showed complex changes, indicating both degeneration and compensatory responses in brain aging.
Conclusions:
Given the importance of exploring brain aging, our study uncovers joint changes between FNC and GMV by combining unimodal and multimodal age prediction, followed by a comprehensive joint analysis. Our study identified areas with both synergistic and contradictory changes between FNC and GMV. The cerebellum and frontal regions showed declines in both FNC and GMV, affecting motor control and emotions. Conversely, in the occipital pole and lateral occipital cortex, which are key areas for vision, and the frontal pole associated with cognition, FNC increased despite a decrease in GMV, which may compensate for structural losses with aging.
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Keywords:
Aging
FUNCTIONAL MRI
Machine Learning
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:
Provide references using APA citation style.
Fotiadis, P., Parkes, L., Davis, K. A., Satterthwaite, T. D., Shinohara, R. T., & Bassett, D. S. (2024). Structure-function coupling in macroscale human brain networks. Nat Rev Neurosci. doi:10.1038/s41583-024-00846-6
Xiong, M., Lin, L., Jin, Y., Kang, W., Wu, S., & Sun, S. (2023). Comparison of Machine Learning Models for Brain Age Prediction Using Six Imaging Modalities on Middle-Aged and Older Adults. Sensors (Basel), 23(7). doi:10.3390/s23073622