Brain functional connectivity alterations associated with neuropsychological performance 6–9 months following SARS‐CoV ‐2 infection

Figures

• 

  FIGURE 1. Flowchart of the study

• 

  FIGURE 2. Intergroup comparisons for neuropsychological testing (after false discovery rate [FDR] correction). (a) Severe patients performed significantly more poorly than mild patients on the RL/RI 16—Sum of three free recalls. (b) Moderate patients had significantly higher interference scores than mild patients on the TMT B/A—Time

• 

  FIGURE 3. Patterns of significantly different functional connectivity in the intergroup comparison. Differences in functional connectivity between brain structures shown in a network representation on a glass brain when comparing severe versus mild (a), moderate versus mild (b), and severe versus moderate (c). Blue lines indicate a decrease in the connectivity measurement (mean decrease = −0.3), red lines indicate an increase in the connectivity measurement (mean increase = 0.3). Statistical significance was false discovery rate (FDR)‐corrected for multiple comparisons (p < .05 FDR). Networks: Cereb: cerebellum; DefaultB: default mode B; DorsAttnA and DorsAttnB: dorsal attention A and B; SalVentAttA: salience ventral attention A; SomMotA and SomMotB: somatosensory motor A and B; TemPar: temporoparietal; Regions: Aud: auditory cortex; ParMed: parietal medial; ParOcc: parietal occipital cortex; ParOper: parietal operculum; PostC: postcentral region; Temp: temporal region; TempOcc: temporo‐occipital region. Figures were created with the BrainNet Viewer (http://www.nitrc.org/projects/bnv/) (Xia et al., 2013)

• 

  FIGURE 4. Whole group partial least squares correlation (PLSC), multivariate latent component 1 (p = .022, 63.11% covariance). A large positive (or negative) weight indicates a larger contribution of the specific feature to the multivariate correlation pattern. (a) Loadings of the behavioral data. Dots represent samples from the bootstrap procedure, yellow highlights mark weights significantly different from zero, and error bars indicate the 95% confidence interval. (b) Bootstrap sampling ratio (BSR) of the functional connectivity. The network representation illustrates the neuroimaging pattern where red links imply a positive influence of the functional connectivity and darker colors indicate a higher number of connections involved in the pattern. (c) Networks and regions of the top 20 connections with higher impact (BSR) on the latent component. Yellow highlights indicate regions with statistically different connectivities in the group comparison

• 

  FIGURE 5. Group‐partial least squares correlation (PLSC), multivariate latent component 1 (p = .0019, 35.96% covariance). A large positive (or negative) weight indicates a larger contribution of the specific feature to the multivariate correlation pattern. (a) Loadings of the behavioral data. Dots represent samples from the bootstrap procedure, yellow highlights mark weights significantly different from zero, and error bars indicate the 95% confidence interval. (b) Bootstrap sampling ratio (BSR) of the functional connectivity. The network representation illustrates the neuroimaging pattern where red links imply a positive influence of the functional connectivity and darker colors indicate a higher number of connections involved in the pattern. (c) Network and regions of the top 20 connections with higher impact (BSR) on the latent component. Yellow highlights indicate regions with statistically different connectivities in the group comparison

• 

  FIGURE 6. Individual partial least squares correlation (PLSC) on mild patients, multivariate latent component 1 (p = .027, 55.99% covariance). A large positive (or negative) weight indicates a larger contribution of the specific feature to the multivariate correlation pattern. (a) Loadings of the behavioral data. Dots represent samples from the bootstrap procedure, yellow highlights mark weights significantly different from zero, and error bars indicate the 95% confidence interval. (b) Bootstrap sampling ratio (BSR) of the functional connectivity. The network representation illustrates the neuroimaging pattern where red links imply a positive influence of the functional connectivity and darker colors indicate a higher number of connections involved in the pattern. (c) Networks and regions of the top 20 connections with higher impact (BSR) on the latent component (left) or with the top 20 highest count in connections from the pattern (right). Yellow highlights indicate regions with statistically different connectivities in the group comparison