Accessing Real-Life Episodic Information from Minutes versus Hours Earlier Modulates Hippocampal and High-Order Cortical Dynamics

Abstract

It is well known that formation of new episodic memories depends on hippocampus, but in real-life settings (e.g., conversation), hippocampal amnesics can utilize information from several minutes earlier. What neural systems outside hippocampus might support this minutes-long retention? In this study, subjects viewed an audiovisual movie continuously for 25 min; another group viewed the movie in 2 parts separated by a 1-day delay. Understanding Part 2 depended on retrieving information from Part 1, and thus hippocampus was required in the day-delay condition. But is hippocampus equally recruited to access the same information from minutes earlier? We show that accessing memories from a few minutes prior elicited less interaction between hippocampus and default mode network (DMN) cortical regions than accessing day-old memories of identical events, suggesting that recent information was available with less reliance on hippocampal retrieval. Moreover, the 2 groups evinced reliable but distinct DMN activity timecourses, reflecting differences in information carried in these regions when Part 1 was recent versus distant. The timecourses converged after 4 min, suggesting a time frame over which the continuous-viewing group may have relied less on hippocampal retrieval. We propose that cortical default mode regions can intrinsically retain real-life episodic information for several minutes.

Keywords: declarative memory; functional connectivity; intersubject correlation; naturalistic; timescales.

Figure 1.

Task diagram and behavioral results. (a) We divided a 25-min movie into 2 sections. The RM group viewed the entire movie without breaks, the DM group viewed Part 1 (Minutes 1–15) 1 day prior to viewing Part 2 (Minutes 16–25), and the NM group viewed Part 2 without ever seeing Part 1. (b) After viewing Part 2, subjects took a written comprehension test. Subjects in the NM group were less accurate than the other groups for questions probing events from Minutes 1–3 of Part 2. (c) In a separate experiment, subjects in the same 3 conditions answered comprehension questions during viewing of Part 2, with the video pausing for each question (average 4 times per minute). RM and DM were more accurate than NM throughout Part 2. No significant differences were found between RM and DM in any minute.

Figure 2.

Processing timescales and the DMN. (a) Map of processing timescales. Subjects viewed a ∼5-min movie clip either temporally scrambled at a fine timescale (<2 s), scrambled coarsely (7–22 s), or intact. For each clip, we calculated ISC at each voxel. Voxels were classified as “Long Timescale” if they responded reliably (i.e., above-threshold ISC) only when the movie was intact (blue); as “Medium Timescale” if they responded reliably during both the intact and coarsely scrambled movie (green); and as “Short Timescale” if they responded reliably in all 3 conditions (red). (b) Map of voxels positively correlated with posterior cingulate during the intact movie, the “default mode network.” Note the similarity between this connectivity map and the “Long Timescale” regions (blue) in Figure 1a. The 2 analyses are independent methods that produce convergent maps.

Figure 3.

Reliable responses in medium- and long-timescale regions reflect memory-based story comprehension during the beginning of Part 2 (Minutes 1–2). (a) Map of voxels with reliable responses (ISC > 0.1) in the RM1 and DM conditions, but not in the NM condition. While short-timescale voxels were reliable in all conditions, ISC was lower for NM than the other groups in medium- and long-timescale voxels. The NM group had impaired story comprehension during Part 2 of the movie (see Fig. 1). (b) The number of voxels exceeding threshold in each condition in the short-timescale, medium-timescale, and long-timescale ROIs, at multiple thresholds, during Minutes 1–2 of Part 2 of the movie. The dotted line indicates the bootstrapped statistical threshold, which was between 0.060 and 0.065 for all groups. Differences between groups were greatest in Minutes 1–2 of Part 2 (Group × Time interaction, F_12,188 = 2.60, P < 0.005; see Supplementary Figure 1 for all time windows). See Supplementary Figure 2 for internal replication.

Figure 4.

ISC in the hippocampus and long-timescale regions. (a) Volume view of ISC map. Voxels throughout the hippocampus were statistically reliable in all groups (RM1 pictured, see Supplementary Fig. 3 for all groups) when ISC was calculated across the entirety of Part 2 (Minutes 1–10). Differences between groups were greatest in Minutes 1–2 of Part 2: threshold of R > 0.1, Group × Time interaction, F_12,188 = 3.04, P < 0.001; effect of Group during Minutes 1–2 (F_3,47 = 4.25, P < 0.01, other time windows P > 0.1). Anatomically defined hippocampus outlined in blue. See also Supplementary Figure 3. (b) The number of voxels exceeding threshold in each group in hippocampus, at multiple thresholds, during Minutes 1–2 of Part 2. More hippocampal voxels exceeded threshold in the DM group than in the other groups. Threshold of R > 0.1: F_3,47 = 4.25, P < 0.01; post hoc t-tests of DM versus other groups, P < 0.05. (c) ISC in hippocampus using all voxels in the ROI during Minutes 1–2 of Part 2. Results are presented at the ROI level (voxel timecourses averaged within-ROI before calculating ISC) and voxel level (ISC calculated at each voxel before averaging across the ROI). Note that ROI-level and voxel-level values are closely related. In both cases, hippocampal ISC was significantly higher in the DM group than the other groups during Minutes 1–2. Voxel-level, F_3,47 = 4.77, P < 0.01; ROI-level ISC, F_3,47 = 3.54, P < 0.05; t-tests of NM versus other groups, P < 0.05. (d) ISC in long-timescale voxels at the ROI level and voxel level during Minutes 1–2 of Part 2. Note that ROI-level and voxel-level values are closely related. In both cases, ISC was significantly lower in the NM group than the other groups during Minutes 1–2. Voxel level: F_3,47 = 9.23, P < 0.0001; ROI level, F_3,47 = 5.85, P < 0.005; t-tests of NM versus other groups, P < 0.0005. Hippocampus and the long-timescale ROI demonstrated different response patterns during the 1–2 Min window: Group × ROI interaction, F_3,47= 6.18, P < 0.005.

Figure 5.

Stimulus-locked correlations between the hippocampus and the rest of the brain during Minutes 1–2 of Part 2. (a) ISFC (see Materials and Methods) within each group for the beginning of Part 2 (Minutes 1–2), using the hippocampus as the seed region. Hippocampal activity was significantly correlated with activity in areas with long timescales, including voxels in posterior cingulate, retrosplenial, and medial prefrontal cortex, in the DM group. Hippocampus ROI shown in gray. (b) On average, correlations between the hippocampus and voxels in long-timescale regions were higher in the DM group than those in the NM or RM groups (F_3,47 = 5.63, P < 0.005; post hoc t-tests, P < 0.05). (c) Correlations between the hippocampus and voxels in posterior cingulate cortex (PCC) and medial prefrontal cortex (mPFC), 2 regions in the DMN defined from an external atlas, were significantly greater in the DM group compared with the other groups. PCC: F_3,47 = 3.70, P < 0.05; t-tests, P < 0.05; mPFC: F_3,47 = 4.21, P < 0.05; t-tests, P < 0.05. See also Supplementary Figure 4.

Figure 6.

Timecourses of RM and DM groups differ in long-timescale voxels during Part 2 of the movie. We created regions of interest composed of voxels with short, medium, and long timescales from the map shown in Figure 2a. At every voxel in each region, we calculated the correlation between activity in DM and RM (DM∼RM2, purple) using a sliding window across the duration of Part 2, then averaged within-region to create the plotted timecourses. To estimate the maximum possible similarity between groups at every time window, we calculated the correlation between 2 independent groups in the RM condition (RM1∼RM2, blue) using the same procedure. The gap between DM∼RM2 and RM1∼RM2 indicates differing neural dynamics when subjects had a 1-day break (DM) versus no break (RM) between Part 1 and Part 2 of the movie. DM and RM were most dissimilar at the beginning of Part 2 and became gradually more aligned over time. Neural effects of the 1-day break lasted until the window ending at 3 min in long-timescale voxels. The sliding window was 120 s wide and center-plotted, i.e., the first value in the trace corresponds to time window 0–120 s. Diamonds indicate q < 0.05, crosses indicate q < 0.1, FDR corrected. See also Supplementary Figures 5 and 6.

Figure 7.

Voxels most sensitive to a 1-day break before Part 2 of the movie are found in long-timescale cortical areas. (a) Dissimilarity between DM and RM groups (two-tailed t-test of [DM∼RM2] versus [RM1∼RM2]) during Minutes 1–2 of Part 2. The voxels with the greatest dissimilarity values were found in long-timescale cortical regions, including posterior medial cortex and posterior parietal cortex. (b) Dissimilarity was strongly correlated with Timescale Index (sensitivity to scrambling in the Timescale Localizer) during Minutes 1–2 of Part 2 (R = 0.32, P < 0.0001). Colors represent the same voxel categories shown in Figure 2a.