Episodic memory and self-reference via semantic autobiographical memory: insights from an fMRI study in younger and older adults

Abstract

Self-referential processing relies mainly on the medial prefrontal cortex (MPFC) and enhances memory encoding (i.e., Self-Reference Effect, SRE) as it improves the accuracy and richness of remembering in both young and older adults. However, studies on age-related changes in the neural correlates of the SRE on the subjective (i.e., autonoetic consciousness) and the objective (i.e., source memory) qualitative features of episodic memory are lacking. In the present fMRI study, we compared the effects of a self-related (semantic autobiographical memory task) and a non self-related (general semantic memory task) encoding condition on subsequent episodic memory retrieval. We investigated encoding-related activity during each condition in two groups of 19 younger and 16 older adults. Behaviorally, the SRE improved subjective memory performance in both groups but objective memory only in young adults. At the neural level, a direct comparison between self-related and non self-related conditions revealed that SRE mainly activated the cortical midline system, especially the MPFC, in both groups. Additionally, in older adults and regardless of the condition, greater activity was found in a fronto-parietal network. Overall, correlations were noted between source memory performance and activity in the MPFC (irrespective of age) and visual areas (mediated by age). Thus, the present findings expand evidence of the role of the MPFC in self-referential processing in the context of source memory benefit in both young and older adults using incidental encoding via semantic autobiographical memory. However, our finding suggests that its role is less effective in aging.

Keywords: aging; autobiographical memory; autonoetic consciousness; medial prefrontal cortex; self-reference effect; semantic memory; source memory.

Figure 1

Mean percentages (and standard deviations) of correctly recollected items (CRI) by group and encoding condition (effect of condition, ***p < 0.001, Bonferroni correction).

Figure 2

Mean percentages (and standard deviations) of correctly recalled sources (CRS) by group and encoding condition (effects of group and condition,* p < 0.05, ** p < 0.01, Bonferroni correction).

Figure 3

Results of the ANOVA with group (young-old) and condition (self-imagery) as factors. (A) Main effect of condition. All regions were more activated during self, compared to the imagery condition. (B) Main effect of group. All regions were more activated in the older, compared to the young group. Results are superimposed to a single subject rendering and are significant at a cluster level p(FWE) < 0.05.

Figure 4

Multiple regressions and mediation analysis. (A) The first column shows the regions showing a correlation with CRS scores: Cuneus and MPFC are shown in the upper and lower part of the figure respectively. Statistical maps are superimposed to an MNI T1 template. Statistical threshold was set at p < 0.001 (uncorrected) and survived SVC with a p(FDR) <0.05. The second column shows the graphical representation of the correlation between the CRS (y-axis) and the Cuneus activation (x-axis), and the CRS and the MPFC activation (x-axis) respectively on the upper and lower part of the figure. (B) Mediation analysis showing correlation between visual areas (Cuneus) and CRS before (green line, r = 0.38, p = 0.02) and after (gray dotted line, r = 0.24, p = 0.17) removing the effect of age in the upper part of the figure. In the lower part, the same mediation analysis shows that the correlation between MPFC and CRS (green line, r = 0.35, p = 0.04) remains significant after correcting for age (green dotted line, r = 0.35, p = 0.04). Red arrows represent negative correlations between Cuneus activation and age (r = −0.39, p = 0.02), and CRS and age (r = −0.49, p = 0.003).