Within and between subject variability of effective connectivity in (small) resting state networks: A spectral dynamic causal modeling study

Abstract

Event Abstract Back to Event Within and between subject variability of effective connectivity in (small) resting state networks: A spectral dynamic causal modeling study Hannes Almgren1*, Frederik Van De Steen1, Simone Kühn2, 3, Adeel Razi4, 5 and Daniele Marinazzo1 1 Ghent University, Data analysis, Belgium 2 Max Planck Institute for Human Development, Center for Lifespan Psychology, Germany 3 University Clinic Hamburg-Eppendorf, Clinic and Polyclinic for Psychiatry and Psychotherapy, Germany 4 University College London, The Wellcome Trust Centre for Neuroimaging, United Kingdom 5 NED University of Engineering and Technology, Department of Electronic Engineering, Pakistan Spectral dynamic causal modeling (spDCM; Friston et al, 2014) is a method developed to infer effective connectivity within resting state networks (RSNs). In the present study we assessed the within and between subject variability of effective connectivity in small-scale RSNs. Three extensive longitudinal datasets were analyzed in the present study. The first dataset included 104 resting state sessions (518 to 600 volumes for each session) from a single subject (male; see, Laumann et al., 2015; myconnectome.org/wp), the second dataset included 158 resting state sessions (200 volumes per session) from one subject (male; see, Choe et al., 2015), the third dataset included 11 to 50 resting state sessions (150 volumes per session) from eight subjects (2 male; see, Pannunzi et al., 2016). All (pre)processing and further analyses were done using spm12 and DCM version 6801. The following preprocessing steps were performed: (1) removal of first five volumes, (2) slice-time correction to middle slice, (3) spatial realignment to first volume, (4) coregistration to anatomical scan acquired during the first session, (5) normalization of both anatomical and functional volumes to MNI space, and (6) spatial smoothing using a Gaussian kernel (FWHM = 6mm). RSNs and regions-of-interest (ROIs) were derived from independent component maps acquired by Smith and colleagues (2009; www.brainmap.org). Eight networks were chosen for the present analyses: occipital and lateral visual, auditory, left and right frontoparietal, somatomotor, default mode, and salience network. The former five networks consisted of two regions, the latter three consisted of three, four, and five regions respectively. To identify peak-voxels of low frequency fluctuations, a GLM was specified and inverted including a discrete cosine basis set with frequencies from 0.0078 to 0.1Hz as principal regressors. Confound regressors included motion (six parameters), WM, and CSF signals. After estimation of the GLM, an F-contrast was specified across the cosine basis set, and the resulting SPM was masked with ROI images (spheres with 10mm radius centered around peak ICA values). The principal eigenvariate from each sphere with radius 8mm, centered on the peak low frequency voxel within each mask, was then computed and imported as time-series in the DCM. Finally, DCMs were constructed and inverted to yield estimates of connectivity. All DCMs were specified and inverted independently for each resting state session. Estimated DCMs that didn’t meet quality standards (e.g., low fit) were not considered for further analyses. For the present analyses, we assessed stability of the sign (i.e., inhibitory versus excitatory) of effective connections in RSNs. Connections were considered as stable if at least 75% of sessions yielded the same sign (positive = excitatory, negative = inhibitory) for a given subject. Across all subjects and networks, 35.21% of connections reached such stability. This proportion was negatively related to the size of the networks: the proportion of stable connections for two, three, four, and five-region networks was 63.00%, 51.67%, 30.83%, and 19.00%, respectively. Furthermore, interhemispheric connections showed more stability compared to intrahemispheric and mediolateral connections across all networks (46.67% versus 32.50% and 30.00% stable connections, respectively). To assess between-subject variability, we only took into account connections that were considered as stable within subjects, using the same standards as above. Two-region networks often showed more pronounced influence in one direction compared to the other. For interhemispheric two-region networks, the direction of the dominant influence varied between-subjects for both the auditory and occipital visual network (50% and 71.43% of participants showed left to right dominance, respectively). However, for the lateral visual network, the connection from left to right hemisphere was dominant in all subjects. Within intrahemispheric two-region networks (i.e., left and right frontoparietal) dominance of one connection over the other also varied between subjects (66.66% and 57.14% of participants showed dominance from posterior to anterior regions, respectively), but was less pronounced compared to interhemispheric networks. For larger networks, we found that the same connections were stable within and between subjects (see, figure 1 for the default mode network), disregarding lateralization. Hemispheric dominance again varied between subjects: within the default mode network 80% of subjects showed more stable projections arising from left hemisphere, within the somatomotor network 70% of subjects showed more stable projections from right hemisphere, and within the salience network 80% of subjects showed more stable projections arising from right insula compared to left insula. Conclusion: We found that approximately 35% of effective connections showed notable stability, which depended on the size of the network and the type of connection (e.g.,, inter- versus intrahemispheric connections). Importantly, connections that were considered stable within subjects were also found stable between subjects (disregarding lateralization). Hemispheric dominance varied between networks and subjects. Finally, the lateral visual network showed consistently left to right hemispheric dominance. Our future research will focus on (1) the effect of preprocessing steps on stability and estimated effective connectivity, and (2) the relation between psychological, emotional, and physiological measures and variation in effective connectivity, (3) the relation between behavioral dominance (e.g., handedness) and hemispheric dominance in effective connectivity. Figure 1 References Choe, A.S., Jones, C.K, Joel, S.E., Muschelli, J., Belegu, V., Caffo, B.S., et al. (2015). Reproducibility and temporal structure in weekly resting-state fMRI over a period of 3.5 years. PLOS one 10, 1-29. Friston, K. J., Kahan, J, Biswal, B, Razi, A. (2014). A DCM for resting state fMRI. NeuroImage 94, 396-407. Laumann, T.O., Gordon, E.M., Adeyemo, B., Snyder, A.Z., Joo, S.J., Chen, M. Y., et al. (2015). Functional system and areal organization of a highly sampled individual human brain. Neuron 87, 657-670. Pannunzi, M., Hindriks, R., Bettinardi, R.G., Wenger, E., Lisofsky, N., Martensson, J., et al. (2016). Resting-state fMRI correlations: from link-wise unreliability to whole brain stability. BioRxiv 2016, [dx.doi.org/10.1101/081976] Smith, S.M., Fox, P.T., Miller, K.L, Glahn, D.C., Fox, P.M., Mackay, C.E., et al. (2009). Correspondence of the brain’s functional architecture during activation and rest. PNAS 106, 13040-13045. Keywords: Dynamic causal modeling (DCM), effective connectivity, stability analysis, fMRI methods, Resting-state fMRI Conference: 12th National Congress of the Belgian Society for Neuroscience, Gent, Belgium, 22 May - 22 May, 2017. Presentation Type: Poster Presentation Topic: Cognition and Behavior Citation: Almgren H, Van De Steen F, Kühn S, Razi A and Marinazzo D (2019). Within and between subject variability of effective connectivity in (small) resting state networks: A spectral dynamic causal modeling study. Front. Neurosci. Conference Abstract: 12th National Congress of the Belgian Society for Neuroscience. doi: 10.3389/conf.fnins.2017.94.00010 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 02 May 2017; Published Online: 25 Jan 2019. * Correspondence: Mr. Hannes Almgren, Ghent University, Data analysis, Gent, Oost-Vlaanderen, 9000, Belgium, Hannes.almgren@ugent.be Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Hannes Almgren Frederik Van De Steen Simone Kühn Adeel Razi Daniele Marinazzo Google Hannes Almgren Frederik Van De Steen Simone Kühn Adeel Razi Daniele Marinazzo Google Scholar Hannes Almgren Frederik Van De Steen Simone Kühn Adeel Razi Daniele Marinazzo PubMed Hannes Almgren Frederik Van De Steen Simone Kühn Adeel Razi Daniele Marinazzo Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.