Combining Fast and Slow fMRI Sampling Rates Enhances Predictive Power in Resting-State Data
Poster No:
1142
Submission Type:
Abstract Submission
Authors:
Joanne Wardell1, Kseniya Solovyeva2, David Danks3, Niko Huotari4, Vesa Kiviniemi4, Vesa Korhonen4, Thomas DeRamus2, Godfrey Pearlson5, Vince Calhoun6, Sergey Plis1
Institutions:
1Georgia State University, Atlanta, GA, 2Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Atlanta, GA, 3Halıcıoğlu Data Science Institute and Department of Philosophy, University of California, San Diego, CA, 4Oulu Functional NeuroImaging Group, University of Oulu, Oulu, Finland, 5Yale University, New Haven, CT, 6GSU/GATech/Emory, Atlanta, GA
First Author:
Co-Author(s):
Kseniya Solovyeva
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
David Danks
Halıcıoğlu Data Science Institute and Department of Philosophy, University of California
San Diego, CA
Halıcıoğlu Data Science Institute and Department of Philosophy, University of California
San Diego, CA
Thomas DeRamus
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Introduction:
Data collection technology in functional MRI (fMRI) is rapidly advancing, resulting in continuous improvements in spatio-temporal resolution. Understanding brain dynamics is crucial for deciphering brain function and dysfunction, which has driven innovation in this field. However, current limits on the frequency of data measurement restrict the types of questions that can be addressed. Large volumes of fMRI data collected at the temporal resolutions available during earlier studies remain underutilized, as a common assumption persists: higher measurement frequencies are the only way to obtain more informative data about brain dynamics. This has led to a tendency to discard older datasets in favor of faster acquisitions, underestimating the potential of integrating slower sampling rates. Furthermore, it results in suboptimal use of the capabilities of modern MRI technology, where only the fastest available rate is often employed. Recent theoretical work suggests that combining high-frequency data with deliberately slower-sampled data can reveal more information about brain dynamics (Solovyeva et al., 2023). This study aims to empirically test this hypothesis in the context of resting-state fMRI.
Methods:
We analyzed resting-state fMRI data from the OULU dataset (Huotari et al., 2019), comprising 10 subjects scanned at slow (TR = 2125ms) and fast (TR = 100ms) sampling rates, and the HCP dataset (Vanessen et al., 2013), featuring 1200 subjects scanned at TR = 720ms. Preprocessing included motion correction, slice timing correction, spatial smoothing, and ICA-based component extraction (Du et al., 2020). For the OULU dataset, features were extracted from combined windows of slow and fast timecourses, while the HCP data were undersampled and recombined to simulate coprime sampling conditions. Noise was generated by using the correlation matrices of permuted ICA timecoures from FBIRN (Keator et al., 2016) subjects and COBRE (Çetin et al., 2014) subjects. The Cholesky decomposition of these matrices were used to transform white noise to colored noise of this structure. The noise was then added to the ICA timecourses of the OULU and HCP data at varying signal to noise ratio (SNR) levels. Classification tasks to distinguish noise-corrupted from non-corrupted timecourses were conducted using machine learning models, including Naive Bayes, SVM, Logistic Regression, and Multi-Layer Perceptron, with cross-validation to ensure robust evaluation.
·Figure 2: Brain networks extracted using constrained ICA for the fast rate (TR=100ms) and slow rate (TR=2150ms) OULU subjects.
Results:
In the OULU dataset, models trained on combined features from slow and fast sampling rates outperformed those trained on single-rate data, demonstrating enhanced predictive power. In contrast, for the HCP dataset, no significant advantage was observed when combining rates, potentially reflecting the limitations of simulated coprime conditions. These results align with theoretical predictions, validating that the combination of distinct sampling rates can yield more informative features under appropriate experimental conditions.
·Figure 1: The mean value of the ROCAUC with shading intervals denoting the standard error are plotted across SNR levels for each classifier.
Conclusions:
This study provides empirical evidence supporting the hypothesis that combining slow and fast sampling rates enhances the informativeness of fMRI data. These findings underscore the value of integrating legacy datasets with modern acquisitions and advocate for deliberate undersampling strategies to optimize data collection. By leveraging combined sampling rates, researchers can gain deeper insights into brain dynamics without exclusively relying on faster acquisition technologies. This approach has implications for both the design of future fMRI studies and the utilization of existing data resources.
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Connectivity (eg. functional, effective, structural) 2
Keywords:
Computing
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?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
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.
Calhoun, V. D. (2014). Thalamus and posterior temporal lobe show greater inter-network connectivity at rest and
across sensory paradigms in schizophrenia. NeuroImage, 97, 117–126. https://doi.org/10.1016/j.neuroimage.2014.
04.009
Du, Y., Fu, Z., Sui, J., Gao, S., Xing, Y., Lin, D., Salman, M., Abrol, A., Rahaman, M. A., Chen, J., Hong, L. E., Kochunov,
P., Osuch, E. A., & Calhoun, V. D. (2020). Neuromark: An automated and adaptive ica based pipeline to identify
reproducible fmri markers of brain disorders. NeuroImage: Clinical, 28, 102375. https://doi.org/https://doi.org/10.
1016/j.nicl.2020.102375
Huotari, N., Raitamaa, L., Helakari, H., Kananen, J., Raatikainen, V., Rasila, A., Tuovinen, T., Kantola, J., Borchardt,
V., Kiviniemi, V. J., & Korhonen, V. O. (2019). Sampling rate effects on resting state fmri metrics. Frontiers in
Neuroscience, 13, 279. https://doi.org/10.3389/fnins.2019.00279
Keator, D. B., van Erp, T. G. M., Turner, J. A., Glover, G. H., Mueller, B. A., Liu, T. T., Voyvodic, J. T., Rasmussen,
J., Calhoun, V. D., Lee, H. J., Toga, A. W., McEwen, S., Ford, J. M., Mathalon, D. H., Diaz, M., O’Leary, D. S.,
Bockholt, H. J., Gadde, S., Preda, A., . . . Potkin, S. G. (2016). The function biomedical informatics research network
data repository. NeuroImage, 124 (Pt B), 1074–1079. https://doi.org/10.1016/j.neuroimage.2015.09.003
Solovyeva, K., Danks, D., Abavisani, M., & Plis, S. (2023, November). Causal learning through deliberate undersampling.
In M. van der Schaar, C. Zhang, & D. Janzing (Eds.), Proceedings of the second conference on causal learning and
reasoning (pp. 518–530, Vol. 213). PMLR. https://proceedings.mlr.press/v213/solovyeva23a.html
Van Essen, D. C., Smith, S. M., Barch, D. M., Behrens, T. E. J., Yacoub, E., & Ugurbil, K. (2013). The wu-minn human
connectome project: An overview. NeuroImage, 80, 62–79. https://doi.org/10.1016/j.neuroimage.2013.05.041