Functional brain modularity as a metric of cognitive function in neuroPASC
Poster No:
1568
Submission Type:
Abstract Submission
Authors:
David O'Connor1, Shams Rashid1, Jacqueline Becker1, Sera Saju1, Claudia Kirsch1, Yael Jacob1, Laurel Morris1, Alan Seifert1, Priti Balchandani1
Institutions:
1Icahn School of Medicine at Mount Sinai, New York, NY
First Author:
Co-Author(s):
Introduction:
Post-acute sequelae of COVID-19 (PASC) refers to a group of persistent symptoms occurring after acute SARS-CoV-2 infection (Proal & VanElzakker, 2021). Neurological manifestations of PASC include cerebrovascular, neurodegenerative, and mental health symptoms, as well as impaired cognitive function (Douaud et al., 2022). The potential burden of neuroPASC on the health system is significant. Our understanding of how COVID-19 infections lead to neuroPASC, and subsequent symptoms is poor. In this study we attempt to characterize the post-acute effects of COVID-19 on neural networks and cognitive function. A well validated functional MRI (fMRI) marker of cognitive function is the modularity of functional connectivity matrices. We sought to investigate the relationship between functional brain modularity and cognitive functioning in a long covid cohort.
Methods:
7T MR imaging of the brain was performed in 49 subjects (35 neuroPASC [24F, 11M], 14 controls [7F, 7M]). Multi echo resting state fMRI (rs-fMRI) and anatomical MP2RAGE were collected. Cognitive measures assessed all domains . MR data were preprocessed using multi-echo independent component analysis (MEICA) implemented in AFNI (Kundu et al., 2013). Functional connectivity matrices were generated using the Desiken-Killiany atlas (Desikan et al., 2006). Graph theory metrics including community membership and modularity were generated from the FC matrices using the Brain Connectivity Toolbox (Rubinov & Sporns, 2010). These metrics were then related to the cognitive functioning scores using linear regression, correcting for age and sex. Following this analysis, cooccurrence of brain regions in communities across subjects was generated for each subgroup.
Results:
Functional brain modularity was marginally higher in controls vs neuroPASC (median 0.34 vs 0.33 , Mann Whitney U = 227, p = 0.65). Semantic fluency (Mean Z = -0.21 vs -0.55) and number span backwards (Mean Z = 0.34 vs 0.3) were also marginally higher in controls. Despite the non-significant relative differences, modularity exhibited a significant association with semantic fluency (ß = 2.7, 95% CI [0.3, 5.1], p = 0.03) and working memory (ß = 2.96, 95% CI [0.04, 5.88], p = 0.047), which attenuated with correction for age and sex (Semantic fluency: ß = 2.24, 95% CI [-0.28, 4.77], p = 0.08, working memory: ß = 2.71, 95% CI [-0.28, 5.71], p = 0.075), as shown in Figure 1. Other cognitive domains including memory, attention, executive functioning, and processing speed were non-significant. A within subgroup analysis of community membership of brain regions revealed the most common divergences between neuroPASC and health controls. Figure 2 shows the regions which differentially associated more often (>35% of the time) in neuroPASC (red), and in healthy controls (blue). The regions which differentially associated most often (>40% of the time) in healthy controls were the left thalamus and right putamen, right thalamus and left putamen, and the left inferior temporal cortex and right inferior temporal cortex. In neuroPASC the regions which associated more often were the left caudal middle frontal area and left inferior temporal area, the left medial temporal orbitofrontal cortex, and left pars triangularis, and the left frontal pole and left pars triangularis.
·Figure 1
·Figure 2
Conclusions:
Our results suggest that while there were non-significant differences in functional brain modularity and cognitive measures in our cohort, perhaps due to sample size, an association was exhibited between modularity and each of semantic fluency and working memory. These results suggest that functional brain modularity may be sensitive to cognitive function and could help shed light on the brain regions whose function are most impacted in neuroPASC.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 2
Learning and Memory:
Working Memory
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
Univariate Modeling
Keywords:
Cognition
FUNCTIONAL MRI
HIGH FIELD MR
1|2Indicates the priority used for review
Provide references using author date format
Douaud, G., Lee, S., Alfaro-Almagro, F., Arthofer, C., Wang, C., McCarthy, P., Lange, F., Andersson, J. L. R., Griffanti, L., Duff, E., Jbabdi, S., Taschler, B., Keating, P., Winkler, A. M., Collins, R., Matthews, P. M., Allen, N., Miller, K. L., Nichols, T. E., & Smith, S. M. (2022). SARS-CoV-2 is associated with changes in brain structure in UK Biobank. Nature, 604(7907), Article 7907. https://doi.org/10.1038/s41586-022-04569-5
Kundu, P., Brenowitz, N. D., Voon, V., Worbe, Y., Vértes, P. E., Inati, S. J., Saad, Z. S., Bandettini, P. A., & Bullmore, E. T. (2013). Integrated strategy for improving functional connectivity mapping using multiecho fMRI. Proceedings of the National Academy of Sciences of the United States of America, 110(40), 16187–16192. https://doi.org/10.1073/pnas.1301725110
Proal, A. D., & VanElzakker, M. B. (2021). Long COVID or Post-acute Sequelae of COVID-19 (PASC): An Overview of Biological Factors That May Contribute to Persistent Symptoms. Frontiers in Microbiology, 12. https://www.frontiersin.org/articles/10.3389/fmicb.2021.698169
Rubinov, M., & Sporns, O. (2010). Complex network measures of brain connectivity: Uses and interpretations. NeuroImage, 52(3), 1059–1069. https://doi.org/10.1016/J.NEUROIMAGE.2009.10.003