Intrinsic Phase-Coupled Variations of Global Signal in Human Resting State Functional Connectivity
Poster No:
1493
Submission Type:
Abstract Submission
Authors:
Theodore LaGrow1, Harrison Watters2, Nmachi Anumba3, Lauren Daley3, Vaibhavi Itkyal3, Lisa Meyer-Baese3, Wen-Ju Pan3, Shella Keilholz3
Institutions:
1Georgia Institute of Technology, Atlanta, GA, 2Emory University, Decatur, GA, 3Georgia Institute of Technology and Emory University, Atlanta, GA
First Author:
Co-Author(s):
Introduction:
This study explores the complex relationship between global signal (GS) and organization of brain states derived by co-activation patterns (CAPs) in human resting-state fMRI data, highlighting the intrinsic phase-coupled relationship between brain states and GS. While CAPs are beneficial in analysis (Bolton et al. 2020; Liu et al. 2018), their effectiveness is influenced by methodological choices such as seed regions and clustering methods (Liu et al. 2018; Iraji et al. 2022). Our focus is on CAPs' correlation with the GS, typically considered a nuisance but shown indicative of brain dynamics (Gutierrez-Barragan et al. 2019; Bolt et al. 2022). Extending previous rodent studies (Gutierrez-Barragan et al. 2019), we assess these relationships using two human datasets to include a broader range of clinical data. This approach facilitates understanding the GS's role in brain function, particularly in different age groups, underscoring its potential in exploring neuropathological and clinical contexts.
Methods:
This study uses two datasets: (1) HCP (Van Essen et al. 2013) consisting of 20 subjects (10F/10M; 4 scans/subject; 80 scans total; avg. age 29.6±3.8) at TR=0.72s; and (2) ADNI (Mueller et al. 2005) consisting of 4 normal cognitive subjects (2F/2M; 16/14 scans; 30 scans total, avg. age 71.5±4.3) at TR=3.0s. Preprocessing follows the standard C-PAC pipeline (Craddock et al. 2013) to best remove non-neuronal noise. Both datasets are parcellated with the Brainnetome atlas (Fan et al. 2016). For HCP, the dataset was bandpass filtered with the Slow-5 range (0.01-0.027 Hz) for specific replication. CAPs analysis followed standard protocol where the top 15% of the DMN was utilized for the HCP data (Liu et al. 2018; Iraji et al. 2022). For ADNI, the dataset was bandpass filtered with Infraslow range (0.01-0.1 Hz). Further, 100% of the dataset was used for CAPs analysis for the ADNI dataset, feasibility demonstrated by Maltbie et al. 2022, to demonstrate the intrinsic nature of the phase-coupled brain states. An ANOVA analysis was performed to test the separability of the brain states to instantaneous phase (Hilbert transform) of the GS
Results:
We find similar variations in the number of brain states for both datasets observing at least two distinct phase-coupled brain states. Understanding GS is usually linked to noise distribution needing removal, organization of nonuniform brain state groups phase-coupled with GS furthers the utility of GS in fMRI analysis. Replicating Gutierrez et al.'s approach, we first examined brain states with 6 components, k=6. By visual inspection, we can a see distinct separation of the GS phase. Fig 1 shows the replication of functional brain states and the relationship to the GS phase with at least two distinct phase-coupled brain state groups of the GS (ANVOA 3-way test: p<1.40973e-17). Fig 2 extends this work and validates two distinct phase-coupled brain state groups with the GS utilizing the Infraslow signal (ANOVA 2-way: p<0.1278; ANOVA 3-way: p<2.57435e-07). These results further our investigation extending to clinical populations. Noting several brain states group at specific GS phases in both rodents and humans, we simplified the analysis by considering 2 brain states, k=2. The separation of these states is highlighted by a clear distinction in the absolute phase values of the GS. Finally, with 3 brain states, k=3, and post ANOVA we confirm the separation of functional brain states based on the phase of the GS. The separation of unique phase-couple brain states into 2 groups is significant for both datasets, however there is less separation for the clinical population. There is clear evidence of separation validating this signal of brain states is intrinsic phase-coupled to the GS.
·Global Signal Phase-Coupled Brain States Demonstration in Human rs-fMRI with HCP dataset
·Global Signal Phase-Coupled Brain States Demonstration in Human rs-fMRI with ADNI dataset.
Conclusions:
Our study both replicates previous findings from rodents to clinical populations and reveals new insights into the phase-coupled relationship in human rs-fMRI data, emphasizing the need for further GS analysis across clinical cohorts.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
fMRI Connectivity and Network Modeling 2
Keywords:
Other - Global Signal, rs-fMRI, Open-Source Data, Co-Activation Patterns, Brain States, Phase-Coupled, HCP, ADNI
1|2Indicates the priority used for review
Provide references using author date format
Bolton, T.A., Tuleasca, C., Wotruba, D., Rey, G., Dhanis, H., Gauthier, B., Delavari, F., Morgenroth, E., Gaviria, J., Blondiaux, E. and Smigielski, L. (2020). TbCAPs: A toolbox for co-activation pattern analysis. Neuroimage, 211, p.116621.
Craddock, C., Sikka, S., Cheung, B., Khanuja, R., Ghosh, S.S., Yan, C., Li, Q., Lurie, D., Vogelstein, J., Burns, R. and Colcombe, S., 2013. Towards automated analysis of connectomes: The configurable pipeline for the analysis of connectomes (c-pac). Front Neuroinform, 42(10.3389).
Fan, L., Li, H., Zhuo, J., Zhang, Y., Wang, J., Chen, L., ... & Jiang, T. (2016). The human brainnetome atlas: a new brain atlas based on connectional architecture. Cerebral cortex, 26(8), 3508-3526.
Gutierrez-Barragan, D., Basson, M. A., Panzeri, S., & Gozzi, A. (2019). Infraslow state fluctuations govern spontaneous fMRI network dynamics. Current Biology, 29(14), 2295-2306.
Iraji, A., Faghiri, A., Fu, Z., Kochunov, P., Adhikari, B.M., Belger, A., Ford, J.M., McEwen, S., Mathalon, D.H., Pearlson, G.D. and Potkin, S.G. (2022). Moving beyond the ‘CAP’of the Iceberg: Intrinsic connectivity networks in fMRI are continuously engaging and overlapping. Neuroimage, 251, p.119013.
Maltbie, E., Yousefi, B., Zhang, X., Kashyap, A. and Keilholz, S., 2022. Comparison of Resting-State Functional MRI Methods for Characterizing Brain Dynamics. Frontiers in Neural Circuits, 16, p.681544.
Mueller, S.G., Weiner, M.W., Thal, L.J., Petersen, R.C., Jack, C., Jagust, W., Trojanowski, J.Q., Toga, A.W. and Beckett, L., 2005. The Alzheimer's disease neuroimaging initiative. Neuroimaging Clinics, 15(4), pp.869-877.
Van Essen, D.C., Smith, S.M., Barch, D.M., Behrens, T.E., Yacoub, E., Ugurbil, K. and Wu-Minn HCP Consortium, 2013. The WU-Minn human connectome project: an overview. Neuroimage, 80, pp.62-79.
Zhang, N., Chang, C., & Duyn, J. H. (2018). Co-activation patterns in resting-state fMRI signals. Neuroimage, 180, 485-494.