Exploring Brain Age Prediction Through MRI Molecular-Enriched Functional Connectivity
Poster No:
884
Submission Type:
Abstract Submission
Authors:
Marco Pinamonti1, Manuela Moretto1,2, Valentina Sammassimo1, Mattia Veronese1,2
Institutions:
1Department of Information Engineering, University of Padua, Padua, Italy, 2Institute of Psychiatry, Psychology & Neuroscience, Department of Neuroimaging, King’s College London, London, United Kingdom
First Author:
Co-Author(s):
Manuela Moretto
Department of Information Engineering, University of Padua|Institute of Psychiatry, Psychology & Neuroscience, Department of Neuroimaging, King’s College London
Padua, Italy|London, United Kingdom
Department of Information Engineering, University of Padua|Institute of Psychiatry, Psychology & Neuroscience, Department of Neuroimaging, King’s College London
Padua, Italy|London, United Kingdom
Mattia Veronese
Department of Information Engineering, University of Padua|Institute of Psychiatry, Psychology & Neuroscience, Department of Neuroimaging, King’s College London
Padua, Italy|London, United Kingdom
Department of Information Engineering, University of Padua|Institute of Psychiatry, Psychology & Neuroscience, Department of Neuroimaging, King’s College London
Padua, Italy|London, United Kingdom
Introduction:
MRI-based functional connectivity (FC) has emerged as a powerful tool for predicting brain age, a biomarker of brain health and neurodevelopment. While previous studies have demonstrated the utility of whole-brain FC in estimating brain age (Liem et al., 2017), these methods lack biological specificity. By integrating molecular information from PET/SPECT data with functional MRI data, a molecular-enriched FC can be obtained, offering novel lens for investigating brain functions. This study aims to determine whether molecular-enriched FC is predictive of brain age in a large cohort of healthy subjects.
Methods:
The study included MRI data of 627 healthy subjects (315F/312M, 53.76±18.47years) of the Cam-CAN dataset (Cam-CAN et al., 2014). The protocol included an anatomical T1w image (TR/TE=2250/2.99ms; voxel size=1x1x1mm) and resting-state functional MRI (rs-fMRI) data acquired with EPI sequence (TR/TE=1970/30ms; voxel size=3x3x4.44mm; 261 volumes).
MRI data was preprocessed using the fMRIPrep 23.2.2 pipeline (Esteban et al., 2019). Subsequently, the XCP-D 0.7.4 pipeline (Mehta et al., 2024), was used for further rs-fMRI preprocessing which included nuisance regression (Muschelli et al., 2014), cubic spline interpolation of volumes with a framewise displacement higher than 0.5mm, and high-pass filtering at 0.001Hz. Finally, data was smoothed using a Gaussian kernel of 6mm.
Preprocessed rs-fMRI data was employed to obtain subject-specific molecular-enriched FC maps, by adopting the Receptor-Enriched Analysis of functional Connectivity by Targets (REACT) (Dipasquale et al., 2019) framework, using available templates of Dopamine, Noradrenaline and Serotonin transporters (DAT, NET and SERT). Thus, after obtaining three spatial maps reflecting the specific transporter-enriched FC, the mean FC values were extracted from 247 regions (ROIs) by merging three atlases including 200 ROIs belonging to 7 networks from the Schaefer atlas, 15 subcortical ROIs from the Harvard-Oxford atlas, and 32 cerebellar ROIs from the Diedrichsen atlas.
For each molecular system, the most important features associated to age were identified using stepwise backward selection, and a linear regression model was fitted to this subset of features to assess the amount of age variance explained by each system. The percentage of selected ROIs per functional network was computed to determine which networks were more influential in predicting brain age.
Finally, to assess the predictive power of FC features enriched by each molecular system separately, a Support Vector Regression (SVR) model with 10-fold cross validation was trained on the selected features.
MRI data was preprocessed using the fMRIPrep 23.2.2 pipeline (Esteban et al., 2019). Subsequently, the XCP-D 0.7.4 pipeline (Mehta et al., 2024), was used for further rs-fMRI preprocessing which included nuisance regression (Muschelli et al., 2014), cubic spline interpolation of volumes with a framewise displacement higher than 0.5mm, and high-pass filtering at 0.001Hz. Finally, data was smoothed using a Gaussian kernel of 6mm.
Preprocessed rs-fMRI data was employed to obtain subject-specific molecular-enriched FC maps, by adopting the Receptor-Enriched Analysis of functional Connectivity by Targets (REACT) (Dipasquale et al., 2019) framework, using available templates of Dopamine, Noradrenaline and Serotonin transporters (DAT, NET and SERT). Thus, after obtaining three spatial maps reflecting the specific transporter-enriched FC, the mean FC values were extracted from 247 regions (ROIs) by merging three atlases including 200 ROIs belonging to 7 networks from the Schaefer atlas, 15 subcortical ROIs from the Harvard-Oxford atlas, and 32 cerebellar ROIs from the Diedrichsen atlas.
For each molecular system, the most important features associated to age were identified using stepwise backward selection, and a linear regression model was fitted to this subset of features to assess the amount of age variance explained by each system. The percentage of selected ROIs per functional network was computed to determine which networks were more influential in predicting brain age.
Finally, to assess the predictive power of FC features enriched by each molecular system separately, a Support Vector Regression (SVR) model with 10-fold cross validation was trained on the selected features.
Results:
Molecular enriched FC explains a good amount of variance in age for all the three considered systems: DAT R²=0.68, NET R²=0.68 and SERT R²=0.69 (Fig. 1).
Across all three models, FC features of the Salience/Ventral Attention network consistently showed the highest percentage of selection (59% in DAT, 59% in NET and 55% in SERT, Fig. 2). However, the contribution of these networks was consistent across receptor systems, suggesting that there is no single network that uniquely drives the selection process in any system.
The SVR model performed best with DAT features (R²=0.53, mean absolute error (MAE)=9.98 years). Slightly worse performances were achieved for NET (R²=0. 53, MAE=10.03 years) and SERT (R²=0.52, MAE=10.37 years) systems.
Across all three models, FC features of the Salience/Ventral Attention network consistently showed the highest percentage of selection (59% in DAT, 59% in NET and 55% in SERT, Fig. 2). However, the contribution of these networks was consistent across receptor systems, suggesting that there is no single network that uniquely drives the selection process in any system.
The SVR model performed best with DAT features (R²=0.53, mean absolute error (MAE)=9.98 years). Slightly worse performances were achieved for NET (R²=0. 53, MAE=10.03 years) and SERT (R²=0.52, MAE=10.37 years) systems.
Conclusions:
A linear regression model shows a good amount of explained variance of the age relationship with molecular-enriched FC maps, in line with previous FC studies (Liem et al., 2017) and highlighting its utility in studying brain aging. Future research will integrate structural data to enhance brain aging prediction accuracy.
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
fMRI Connectivity and Network Modeling
Multivariate Approaches
Keywords:
Aging
Dopamine
FUNCTIONAL MRI
Machine Learning
Multivariate
Norpinephrine
Seretonin
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.
[2] Dipasquale, O., Selvaggi, P., Veronese, M., Gabay, A. S., Turkheimer, F., & Mehta, M. A. (2019). Receptor-Enriched Analysis of functional connectivity by targets (REACT): A novel, multimodal analytical approach informed by PET to study the pharmacodynamic response of the brain under MDMA. NeuroImage, 195, 252–260. https://doi.org/10.1016/j.neuroimage.2019.04.007
[3] Esteban, O., Markiewicz, C. J., Blair, R. W., Moodie, C. A., Isik, A. I., Erramuzpe, A., Kent, J. D., Goncalves, M., DuPre, E., Snyder, M., Oya, H., Ghosh, S. S., Wright, J., Durnez, J., Poldrack, R. A., & Gorgolewski, K. J. (2019). FMRIPrep: A robust preprocessing pipeline for functional MRI. Nature Methods, 16(1), 111–116. https://doi.org/10.1038/s41592-018-0235-4
[4] Liem, F., Varoquaux, G., Kynast, J., Beyer, F., Kharabian Masouleh, S., Huntenburg, J. M., Lampe, L., Rahim, M., Abraham, A., Craddock, R. C., Riedel-Heller, S., Luck, T., Loeffler, M., Schroeter, M. L., Witte, A. V., Villringer, A., & Margulies, D. S. (2017). Predicting brain-age from multimodal imaging data captures cognitive impairment. NeuroImage, 148, 179–188. https://doi.org/10.1016/j.neuroimage.2016.11.005
[5] Mehta, K., Salo, T., Madison, T. J., Adebimpe, A., Bassett, D. S., Bertolero, M., Cieslak, M., Covitz, S., Houghton, A., Keller, A. S., Lundquist, J. T., Luo, A., Miranda-Dominguez, O., Nelson, S. M., Shafiei, G., Shanmugan, S., Shinohara, R. T., Smyser, C. D., Sydnor, V. J., … Satterthwaite, T. D. (2024). XCP-D: A robust pipeline for the post-processing of fMRI data. Imaging Neuroscience, 2, 1–26. https://doi.org/10.1162/imag_a_00257
[6] Muschelli, J., Nebel, M. B., Caffo, B. S., Barber, A. D., Pekar, J. J., & Mostofsky, S. H. (2014). Reduction of motion-related artifacts in resting state fMRI using aCompCor. NeuroImage, 96, 22–35. https://doi.org/10.1016/j.neuroimage.2014.03.028