Comparison of fMRI Features for Brain-Behavior Prediction
Poster No:
1116
Submission Type:
Abstract Submission
Authors:
Mikkel Schöttner1, Thomas Bolton1, Jagruti Patel1, Patric Hagmann2
Institutions:
1University of Lausanne, Lausanne, Vaud, 2Lausanne University Hospital and University of Lausanne (CHUV-UNIL), Lausanne, Vaud
First Author:
Co-Author(s):
Introduction:
Brain-behavior prediction enables the development of robust models of how the brain gives rise to our behavior, and is the first step in the development of neuroimaging biomarkers. However, many methodological choices need to be made in the processing pipeline. An important consideration is the choice of feature that is used to predict the behavior at hand. Functional connectivity (FC), is an established choice (He et al., 2020; Ooi et al., 2022), but regional measures of functional magnetic resonance imaging (fMRI) and those derived using graph signal processing (GSP), have also shown potential in modeling inter-individual differences (Brahim & Farrugia, 2020; Griffa et al., 2022). In addition to evaluating its predictive performance, varying the sample size and scan time available to train a model for a given feature can give an indication of potential performance reserves. At the same time, the choice between acquiring more subjects or scanning each subject for longer is important to explore in the context of prospective brain-behavior prediction studies (Ooi et al., 2024).
Methods:
We compared the prediction performance of 10 fMRI measures on 979 subjects from the Human Connectome Project Young Adult dataset (Van Essen et al., 2012), predicting cognition, age, and gender. We explored the comparative performance of FC, regional measures, and measures using GSP. The region-wise measures were mean, standard deviation, mean square successive difference (MSSD, Garrett et al., 2013), and fractional amplitude of low-frequency fluctuations (fALFF, Zou et al., 2008). The GSP-based measures were structural decoupling index (SDI, Preti & Van De Ville, 2019), power spectral density (PSD), and coupled and decoupled FC (Griffa et al., 2022). To test whether any differences in performance was due to the dimensionality of the feature, the first 274 principal components of FC were also included as a feature. Depending on the dimensionality of the feature and whether the target was continuous or discrete, we used elastic-net regression, kernel-ridge regression, elastic-net classifier, or support vector machine as our model. The model was evaluated in a nested cross-validation with 10 random splits. The scaling properties were explored by fitting the same models, varying the size of the training set and the number of scanning sessions.
Results:
Figure 1 shows the results of the feature comparison. FC-based measures came out best for all three targets. Standard deviation and MSSD also showed potential for predicting cognition, while PSD and SDI did relatively well in the age prediction. Coupled FC outperformed FC when predicting gender. Figure 2 shows the results of the scaling experiment. Both sample size and scan time were important for predicting each target. Prediction performance leveled off for a given scan time for cognition and gender, indicating that it places a limit on how well a model is able to perform.
Conclusions:
Our results confirm that FC represents a robust feature for brain-behavior prediction. They further illustrate the potential of GSP-based features and measures of variability for this task. Both scan time and sample size should be considered when planning data acquisition for brain-behavior prediction studies, with the former placing a limit on the maximum performance that can be achieved.
Higher Cognitive Functions:
Higher Cognitive Functions Other
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Connectivity (eg. functional, effective, structural) 2
fMRI Connectivity and Network Modeling
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Cognition
Data analysis
FUNCTIONAL MRI
Machine Learning
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.
Garrett, D. D., Samanez-Larkin, G. R., MacDonald, S. W. S., Lindenberger, U., McIntosh, A. R., & Grady, C. L. (2013). Moment-to-moment brain signal variability: A next frontier in human brain mapping? Neuroscience & Biobehavioral Reviews, 37(4), 610–624.
Griffa, A., Amico, E., Liégeois, R., Van De Ville, D., & Preti, M. G. (2022). Brain structure-function coupling provides signatures for task decoding and individual fingerprinting. NeuroImage, 250, 118970. https://doi.org/10.1016/j.neuroimage.2022.118970
He, T., Kong, R., Holmes, A. J., Nguyen, M., Sabuncu, M. R., Eickhoff, S. B., Bzdok, D., Feng, J., & Yeo, B. T. T. (2020). Deep neural networks and kernel regression achieve comparable accuracies for functional connectivity prediction of behavior and demographics. NeuroImage, 206, 116276.
Ooi, L. Q. R., Chen, J., Zhang, S., Kong, R., Tam, A., Li, J., Dhamala, E., Zhou, J. H., Holmes, A. J., & Yeo, B. T. T. (2022). Comparison of individualized behavioral predictions across anatomical, diffusion and functional connectivity MRI. NeuroImage, 263, 119636.
Ooi, L. Q. R., Orban, C., Nichols, T. E., Zhang, S., Tan, T. W. K., Kong, R., Marek, S., Dosenbach, N. U. F., Laumann, T., Gordon, E. M., Zhou, J. H., Bzdok, D., Eickhoff, S. B., Holmes, A. J., & Yeo, B. T. T. (2024). MRI economics: Balancing sample size and scan duration in brain wide association studies (p. 2024.02.16.580448). bioRxiv.
Preti, M. G., & Van De Ville, D. (2019). Decoupling of brain function from structure reveals regional behavioral specialization in humans. Nature Communications, 10(1), Article 1.
Van Essen, D. C., Ugurbil, K., Auerbach, E., Barch, D., Behrens, T. E. J., Bucholz, R., Chang, A., Chen, L., Corbetta, M., Curtiss, S. W., Della Penna, S., Feinberg, D., Glasser, M. F., Harel, N., Heath, A. C., Larson-Prior, L., Marcus, D., Michalareas, G., Moeller, S., … WU-Minn HCP Consortium. (2012). The Human Connectome Project: A data acquisition perspective. NeuroImage, 62(4), 2222–2231.
Zou, Q.-H., Zhu, C.-Z., Yang, Y., Zuo, X.-N., Long, X.-Y., Cao, Q.-J., Wang, Y.-F., & Zang, Y.-F. (2008). An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: Fractional ALFF. Journal of Neuroscience Methods, 172(1), 137–141.