Fig. 2: Univariate activation during memory recall and future thinking.

a Whole-brain t-statistic maps of cortical parcels showing higher or lower activation while describing semantic memory about the world or other people (top) or about oneself (bottom), compared to describing the current state. b Whole-brain t-statistic maps of cortical parcels showing higher or lower activation while describing episodic memory (top) or future-oriented thoughts (bottom), compared to describing the current state. In both a and b, the t-statistic maps are displayed on the lateral (left) and medial (right) surfaces of the left hemisphere of the inflated fsaverage6 template brain. Parcels with significantly higher activation compared to the current state are shown in red, while those with significantly lower activation are shown in blue. The statistical significance of each contrast (two-tailed p < 0.05) was Bonferroni corrected across the 400 parcels in the Schaefer atlas³⁵. Supplementary Tables 4–7 provide the lists of suprathreshold parcels from both hemispheres. Similar activation maps for memory and future thinking were obtained after excluding thought boundary periods from the analysis (Supplementary Fig. 1; see “Neural responses at major thought transitions” in “Results”). c Mean blood oxygenation level-dependent (BOLD) signal for each thought category in the posterior medial cortex (PMC; top) and the hippocampus (bottom). Each colored dot represents an individual participant (N = 62, 75, 72, 73, and 72 for current, semantic-world, semantic-self, episodic, and future categories, respectively). Black circles indicate the mean across participants within each category. Error bars show the SEM across participants. Statistical significance indicates differences between thought categories based on two-tailed paired t-tests. Full statistics for individual comparisons, including exact p values, are reported in Supplementary Table 8. **p < 0.01, ***p < 0.001 (uncorrected).