Interconnections within the rrAD420 Functional Network Atlas for Older Adults
Poster No:
2210
Submission Type:
Abstract Submission
Authors:
Norman Scheel1, Zachary Fernandez1, Seyedeh-Rezvan Farahibozorg2, Jeffrey Keller3, Ellen Binder4, Eric Vidoni5, Jeffrey Burns5, Ann Stowe6, Diana Kerwin7, Wanpen Vongpatanasin8, C. Munro Cullum8, Rong Zhang8, David Zhu1
Institutions:
1Michigan State University, East Lansing, MI, 2University of Oxford, Oxford, Non-US/Other, 3Pennington Biomedical Research Center, Baton Rouge, LA, 4Washington University School of Medicine, St. Louis, MO, 5University of Kansas Alzheimer's Disease Center, Fairway, KS, 6University of Kentucky, Lexington, KY, 7Texas Health Presbyterian Hospital, Dallas, TX, 8University of Texas Southwestern Medical Center, Dallas, TX
First Author:
Co-Author(s):
Introduction:
Risk Reduction for Alzheimer's Disease (rrAD) is a recently completed randomized clinical trial designed to investigate effects of improved cardiovascular health on neurocognitive function, brain structure, and brain network functional connectivity (Szabo-Reed et al., 2019). rrAD enrolled 513 hypertensive older adults (60 to 84 years, 68.8±5.9) who had a family history of dementia or subjective cognitive decline. Study participants underwent aerobic exercise training and/or intensive pharmacological interventions for 2 years. Of 513 participants, 420 completed anatomical and resting-state functional MRI (rs-fMRI) scans at baseline and after 2-years of interventions. MRI scans were performed on 5 different 3T scanners. Most of the currently used rs-fMRI atlases are based on young and healthy subjects which may not be applicable in older adults who had brain atrophy and/or changes in brain resting-sate network (RSN) connectivity. The aim of this study was to use the rrAD baseline rs-fMRI data to create a robust rs-fMRI atlas, which is suitable for studying older adults, without the limitation of spatial independence, and allowing for spatially overlapping temporal configurations (modes) to facilitate the investigation of functional interconnections of RSNs.
Methods:
Data acquisition protocols on all five 3T scanners were harmonized to ensure comparability across sites. Using SPM12's DARTEL registration we created a cohort-specific MNI-adjacent anatomical template space, namely rrAD420. The fMRI data were preprocessed using slice-timing correction, T1 co-registration and motion correction, spatial blurring of 4mm, aggressive ICA-AROMA, and normalization into the rrAD420 space. Details on scanning and preprocessing parameters were presented in Scheel et al., 2022. After extensive manual quality assessment and control, we used a data-driven probabilistic functional mode decomposition (PROFUMO) approach (Farahibozorg et al., 2021; Harrison et al., 2015) to create a rs-fMRI atlas that includes different modes of major networks. With sampling multiple dimensions (30,50, and 80) for the PROFUMO decomposition, we assessed RSN representation, splitting, grouping, reproducibility, and identifiability, through cross-referencing with diverse brain atlases and functional parcellations (Damoiseaux et al., 2006; Yeo et al., 2011; Shirer et al., 2012; Pruim et al., 2015), to determine which decomposition provides the best overall RSN fit.
Results:
For network representation, we found that all PROFUMO decompositions captured the RSNs of the reference functional atlases. However, the 50-mode parcellation ranked highest for RSN splitting and grouping metrics. For the 50-dimensional PROFUMO, we found 42 functional modes attributed to neural activity and 8 components attributed to noise. Following consensus on RSN taxonomy (Uddin et al., 2023), we grouped the resulting 42 modes into 13 singular networks and 6 combinatory networks, and sorted these by their reproducibility (see Table 1).
Conclusions:
Using PROFUMO decomposition, we created a rs-fMRI network atlas specifically for an older population. Additionally, we subdivided these networks into spatiotemporally overlapping modes, providing more insight into the temporal organization of resting-state networks in older subjects. The resulting combinatory networks of the salience and attention, as well as the language and default modes (see Figure 1), are especially interesting as they are consistent with recent findings (Gordon et al., 2020). Thus, this new atlas may be able to give more nuanced insights into cognitive processes leading to cognitive decline, previously difficult to disentangle using common Independent Component Analysis (ICA) approaches. We expect that the rrAD420 rs-fMRI atlas will be applicable to study the rs-fMRI connectivity and cognition of older populations in general, which could lead to more reliable biomarker development and implementation.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)
Lifespan Development:
Aging 2
Modeling and Analysis Methods:
fMRI Connectivity and Network Modeling
Task-Independent and Resting-State Analysis
Neuroinformatics and Data Sharing:
Brain Atlases 1
Keywords:
Aging
Degenerative Disease
FUNCTIONAL MRI
Language
Modeling
MRI
Other - Resting-State, Networks, Atlas
1|2Indicates the priority used for review
Provide references using author date format
Farahibozorg, S. R., Bijsterbosch, J. D., Gong, W., Jbabdi, S., Smith, S. M., Harrison, S. J., & Woolrich, M. W. (2021). Hierarchical modelling of functional brain networks in population and individuals from big fMRI data. NeuroImage, 243. https://doi.org/10.1016/j.neuroimage.2021.118513
Gordon, E. M., Laumann, T. O., Marek, S., Raut, R. V., Gratton, C., Newbold, D. J., Greene, D. J., Coalson, R. S., Snyder, A. Z., Schlaggar, B. L., Petersen, S. E., Dosenbach, N. U. F., & Nelson, S. M. (2020). Default-mode network streams for coupling to language and control systems. Proceedings of the National Academy of Sciences, 117(29), 17308–17319. https://doi.org/10.1073/pnas.2005238117
Harrison, S. J., Woolrich, M. W., Robinson, E. C., Glasser, M. F., Beckmann, C. F., Jenkinson, M., & Smith, S. M. (2015). Large-scale probabilistic functional modes from resting state fMRI. NeuroImage, 109, 217–231. https://doi.org/10.1016/j.neuroimage.2015.01.013
Pruim, R. H. R., Mennes, M., Buitelaar, J. K., & Beckmann, C. F. (2015). Evaluation of ICA-AROMA and alternative strategies for motion artifact removal in resting state fMRI. NeuroImage, 112, 278–287. https://doi.org/10.1016/j.neuroimage.2015.02.063
Scheel, N., Keller, J. N., Binder, E. F., Vidoni, E. D., Burns, J. M., Thomas, B. P., Stowe, A. M., Hynan, L. S., Kerwin, D. R., Vongpatanasin, W., Rossetti, H., Cullum, C. M., Zhang, R., & Zhu, D. C. (2022). Evaluation of noise regression techniques in resting-state fMRI studies using data of 434 older adults. Frontiers in Neuroscience, 16. https://doi.org/10.3389/fnins.2022.1006056
Shirer, W. R., Ryali, S., Rykhlevskaia, E., Menon, V., & Greicius, M. D. (2012). Decoding subject-driven cognitive states with whole-brain connectivity patterns. Cerebral Cortex (New York, N.Y. : 1991), 22(1), 158–165. https://doi.org/10.1093/cercor/bhr099
Szabo-Reed, A. N., Vidoni, E., Binder, E. F., Burns, J., Cullum, C. M., Gahan, W. P., Gupta, A., Hynan, L. S., Kerwin, D. R., Rossetti, H., Stowe, A. M., Vongpatanasin, W., Zhu, D. C., Zhang, R., & Keller, J. N. (2019). Rationale and methods for a multicenter clinical trial assessing exercise and intensive vascular risk reduction in preventing dementia (rrAD Study).