Print Close

Effective workflow from multi-modal MRI data to model-based prediction

Poster No:

1086

Submission Type:

Abstract Submission

Authors:

Kyesam Jung¹^,2, Kevin Wischnewski¹^,2^,3, Simon Eickhoff¹^,2, Oleksandr Popovych¹^,2

Institutions:

¹Institute of Neurosciences and Medicine - Brain and Behaviour (INM-7), Research Centre Jülich, Jülich, Germany, ²Institute of Systems Neuroscience, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, Düsseldorf, Germany, ³Institute of Mathematics, Faculty of Mathematics and Natural Sciences, Heinrich Heine University Düsseldorf, Düsseldorf, Germany

First Author:

Kyesam Jung
Institute of Neurosciences and Medicine - Brain and Behaviour (INM-7), Research Centre Jülich|Institute of Systems Neuroscience, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf
Jülich, Germany|Düsseldorf, Germany

Co-Author(s):

Kevin Wischnewski
Institute of Neurosciences and Medicine - Brain and Behaviour (INM-7), Research Centre Jülich|Institute of Systems Neuroscience, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf|Institute of Mathematics, Faculty of Mathematics and Natural Sciences, Heinrich Heine University Düsseldorf
Jülich, Germany|Düsseldorf, Germany|Düsseldorf, Germany

Simon Eickhoff
Institute of Neurosciences and Medicine - Brain and Behaviour (INM-7), Research Centre Jülich|Institute of Systems Neuroscience, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf
Jülich, Germany|Düsseldorf, Germany

Oleksandr Popovych
Institute of Neurosciences and Medicine - Brain and Behaviour (INM-7), Research Centre Jülich|Institute of Systems Neuroscience, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf
Jülich, Germany|Düsseldorf, Germany

Introduction:

Understanding the relationship between brain structure and function remains a key challenge in connectome research. Despite significant advances in neuroimaging, the mechanisms underlying this relationship remain elusive, as the correlation between structural (SC) and functional connectivity (FC) is typically weak (Suarez et al., 2020), which limits the utility of this data for studying cognitive functions and brain disorders. Model-based approaches offer a promising solution by simulating FC from SC and providing a framework (Fig. 1) to explore hidden dynamics and test hypotheses about brain activity (Popovych et al., 2019). These simulations reveal distinct connectome patterns beyond empirical data, including stronger correlations with clinical scores (Jung et al., 2024), enhanced reliability and improved subject specificity (Domhof et al., 2022). Furthermore, employing simulated connectome features to machine learning can lead to improved prediction performance, which may contribute to a better understanding of the brain-behavior relationship.

·Figure 1. A workflow for model-based prediction research.

Methods:

This study included 270 young adults from the Human Connectome Project (Van Essen et al., 2013). The Schaefer atlas with 100 parcels (Schaefer et al., 2018) and the Harvard-Oxford atlas with 96 parcels (Desikan et al., 2006) were used for brain connectomes of SC and FC. Then we performed the whole-brain dynamical modeling, where the model was fit to empirical data in low-dimensional (two parameters) and high-dimensional (about hundred parameters) model parameter spaces. Connectome relationships were calculated by Pearson's correlation between empirical SC (eSC), empirical FC (eFC), and simulated FC (sFC) and used as features for machine-learning via cross-validated regression. To investigate the contribution of connectome relationships to prediction, we prepared two distinct feature sets: empirical (eSC vs. eFC) and simulated (eFC vs. sFC) features. The suggested model-based prediction approach was illustrated for sex classification, prediction of cognitive scores and five personality traits of individual subjects (McCrae & Costa, 2004).

Results:

The mean goodness-of-fit (connectome correlation) of the whole-brain model for the high-dimensional optimization was 0.61 (Schaefer) and 0.72 (Harvard-Oxford), compared to the respective 0.30 and 0.50 for the low-dimensional optimization. Machine-learning performance showed that involving the simulated features obtained through high-dimensional model optimization yielded the best outcome in the sex classification (Fig. 2a). Likewise, the low-dimensional optimization condition showed the best prediction of cognition (Fig. 2b). The simulated features also showed the best results in the predictions of four out of five personality traits (Fig. 2c-g), except for the openness, where the empirical features demonstrated the best performance (Fig. 2g). These findings indicate that whole-brain dynamical modeling can enhance machine-learning performance. This is especially evident in prediction of cognition and most of personality traits (Fig. 2b-f), where the empirical features mainly showed correlations near zero or below, whereas the simulated features demonstrated clearly improved results. We also found that employing simulated features of the high-dimensional model optimization for simultaneous prediction of all five personality traits clearly outperformed the prediction results obtained for empirical features (effect size: 0.836) (Fig. 2h).

·Figure 2. Prediction results with different feature conditions.

Conclusions:

By incorporating the simulated features alongside empirical data, we can extensively explore brain connectomes and achieve enhanced performance at behavior prediction. In light of these findings, it is important to elucidate the specific advantages offered by the model-based approach. Consequently, the systematic approach proposed in this report represents a promising method for advancing the application of brain modeling for investigation of inter-individual variability of the brain-behavior relationship.

Modeling and Analysis Methods:

Classification and Predictive Modeling ¹

Connectivity (eg. functional, effective, structural) ²

fMRI Connectivity and Network Modeling

Keywords:

Computational Neuroscience

Machine Learning

Modeling

Other - Brain-Behavior Relationship

^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.

I accept

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:

I am submitting this abstract as an original work to be reproduced. I am available to be the “source party” in an upcoming team and consent to have this work listed on the OSSIG website. I agree to be contacted by OSSIG regarding the challenge and may share data used in this abstract with another team.

Please indicate below if your study was a "resting state" or "task-activation” study.

Resting state

Healthy subjects only or patients (note that patient studies may also involve healthy subjects):

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.

Yes

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.

Not applicable

Please indicate which methods were used in your research:

Functional MRI

Structural MRI

Diffusion MRI

Behavior

Computational modeling

For human MRI, what field strength scanner do you use?

3.0T

Which processing packages did you use for your study?

AFNI

FSL

Free Surfer

Other, Please list - MRtrix, ANTs

Provide references using APA citation style.

Desikan, R. S., Segonne, F., Fischl, B., Quinn, B. T., Dickerson, B. C., Blacker, D., Buckner, R. L., Dale, A. M., Maguire, R. P., Hyman, B. T., Albert, M. S., & Killiany, R. J. (2006, Jul 1). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage, 31(3), 968-980. https://doi.org/10.1016/j.neuroimage.2006.01.021
Domhof, J. W. M., Eickhoff, S. B., & Popovych, O. V. (2022, Aug 15). Reliability and subject specificity of personalized whole- brain dynamical models. Neuroimage, 257, 119321. https://doi.org/10.1016/j.neuroimage.2022.119321
Jung, K., Eickhoff, S. B., Caspers, J., UKD-PD team, & Popovych, O. V. (2024). Simulated brain networks reflecting progression of Parkinson’s disease. Network Neuroscience, 1-31. https://doi.org/10.1162/netn_a_00406
McCrae, R. R., & Costa, P. T. (2004). A contemplated revision of the NEO Five-Factor Inventory. Personality and Individual Differences, 36(3), 587-596. https://doi.org/10.1016/s0191-8869(03)00118-1
Popovych, O. V., Manos, T., Hoffstaedter, F., & Eickhoff, S. B. (2019). What Can Computational Models Contribute to Neuroimaging Data Analytics? Front Syst Neurosci, 12, 68. https://doi.org/10.3389/fnsys.2018.00068
Schaefer, A., Kong, R., Gordon, E. M., Laumann, T. O., Zuo, X. N., Holmes, A. J., Eickhoff, S. B., & Yeo, B. T. T. (2018, Sep 1). Local-Global Parcellation of the Human Cerebral Cortex from Intrinsic Functional Connectivity MRI. Cereb Cortex, 28(9), 3095-3114. https://doi.org/10.1093/cercor/bhx179
Suarez, L. E., Markello, R. D., Betzel, R. F., & Misic, B. (2020, Apr). Linking Structure and Function in Macroscale Brain Networks. Trends Cogn Sci, 24(4), 302-315. https://doi.org/10.1016/j.tics.2020.01.008
Van Essen, D. C., Smith, S. M., Barch, D. M., Behrens, T. E., Yacoub, E., Ugurbil, K., & Consortium, W. U.-M. H. (2013, Oct 15). The WU-Minn Human Connectome Project: an overview. Neuroimage, 80, 62-79. https://doi.org/10.1016/j.neuroimage.2013.05.041

UNESCO Institute of Statistics and World Bank Waiver Form

I attest that I currently live, work, or study in a country on the UNESCO Institute of Statistics and World Bank List of Low and Middle Income Countries list provided.