Figure 1.

Connectivity matrices of macaque cortex. A, Structural connectivity matrix derived from the CoCoMac database. Connections are shown in white. B, Example of a “static” functional connectivity matrix (top) constructed from an entire resting-state BOLD-fMRI scan lasting 600 s from one animal. Examples of “dynamic” functional connectivity matrices constructed using different window sizes from the same scan (bottom). Functional connectivity was computed as Pearson correlations of ROI time series. Matrices are organized generally from anterior to posterior regions, for left then right hemispheres. Labels and ordering of regions within each hemisphere are specified in Table 1. C, Cosine similarity between rsFC and structural connectivity increases with increasing FC window size. Similarity was computed between the structural connectivity matrix and each rsFC matrix at each time point for all window sizes and averaged (± SEM) across time points for each window size. For comparison with previous studies, performing this analysis using Pearson correlations resulted in a similar increase in correspondence between the structural and functional networks as window size increased (r range: 0.081–0.226; Kruskal–Wallis, p < 0.001). Gray circles represent the average ± SEM of cosine similarity between windowed FC networks and 1000 null models with fixed degree distributions.