Figure 3.

Intraindividual and interindividual variability of the parcellations. A) The parcellations generated using the resting-state sessions demonstrated high intraindividual reliability and high interindividual variability. When comparing two resting-state sessions of the same participant, on average, 69–78% of the STC vertices were assigned to the same networks. Between a given participant and any other participant, on average, only 57–68% of the STC vertices were assigned to the same networks. For each STC parcellation, the consistency of network membership was significantly higher within than between participants (p<0.001 for all conditions). The overlap between the resting-state parcellations within participants was higher than that between the resting-state and task parcellations (task vs. rest: 62–74%, rest vs. rest: 69–78%, p<0.05) for all parcellations expect the left hemispheric 6-network parcellation. This suggests that auditory cortex functional networks can be shaped by task. B) The correspondence between task and resting-state parcellations of a given participant (62–74%) was higher than the correspondence between task parcellations of the participant and any other participant (55–66%; p< 0.001), suggesting than the parcellations are most similar within participants regardless of the task. The differences between these three conditions were estimated with the Friedman test and pairwise Wilcoxon signed rank tests. The results were corrected for multiple comparisons using Benjamini-Hochberg procedure [55]. Error bars indicate standard errors of the mean (SEM).