Fig. 4: Neural responses at thought boundaries.
From: Neural dynamics of spontaneous memory recall and future thinking in the continuous flow of thoughts
a Whole-brain t-statistic map of the univariate contrast between strong boundary (boundary agreement = 1) and non-boundary periods. Parcels with significantly higher activation during strong boundary periods compared to non-boundary periods are shown in red, while those with significantly lower activation are shown in blue. Statistical significance (two-tailed p < 0.05) was Bonferroni corrected across all parcels. White outlines indicate the auditory cortex and the posterior medial cortex (PMC), respectively. b Mean PMC (left) and hippocampus (right) activation time courses aligned at different types of thought boundaries, with different colors indicating the boundary types (strong boundary = red; topic change = orange; category change = yellow; non-boundary = gray). Time zero for the non-boundary condition represents the middle of thoughts longer than 15 s. For other conditions, time zero represents the offset of the pre-boundary thought. Solid lines indicate the mean across participants (N = 75 for all conditions). Shaded areas indicate the SEM across participants. Asterisks above the x-axis indicate time points where activation for strong boundaries is significantly higher than non-boundaries, as determined by two-tailed paired t-tests with Bonferroni correction (p < 0.05). Full statistics for individual time points, including exact p values, are reported in Supplementary Table 14. c Boundary pattern similarity analysis. For each region, we computed the mean activation pattern of between-movie boundaries from the movie-watching phase of our prior study7. This template pattern was correlated with the mean activation patterns of strong thought boundaries (red bars) and non-boundary periods (gray bars) during think-aloud. d Whole-brain t-statistic map of boundary-specific pattern similarity. Parcels are shown in red if their between-movie boundary patterns were more similar to their strong thought boundary patterns than to the non-boundary patterns. The map is masked to only include areas that showed positive correlations between the between-movie boundary patterns and the strong thought boundary patterns. Statistical significance (two-tailed p < 0.05) was Bonferroni corrected across all parcels. e Boundary pattern similarity in PMC. The think-aloud strong thought boundary and non-boundary patterns were correlated with the mean activation patterns of between-movie boundary periods (left panel) or silent periods (right panel) from the movie watching phase7. Each colored dot represents an individual participant (N = 75 for all conditions). Black circles indicate the mean across participants. Error bars show the SEM across participants. Statistical significance relative to zero was assessed using two-tailed one-sample t-tests, while differences between conditions were evaluated using two-tailed paired t-tests.
