Effects of Healthy and Neuropathological Aging on Autobiographical Memory: A Meta-Analysis of Studies Using the Autobiographical Interview

Abstract

Objective A meta-analytic review was conducted to assess the effects of healthy aging, amnestic Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD) on naturalistic autobiographical memory using the Autobiographical Interview, a widely used, standardized assessment that derives measures of internal (episodic) and external (non-episodic) details from freely recalled autobiographical narratives. Method A comprehensive literature search identified 21 aging, 6 MCI, and 7 AD studies (total N =1556 participants). Summary statistics for internal and external details for each comparison (younger vs. older or MCI/AD vs. age-matched comparison groups) and effect size statistics were extracted and summarized using Hedges’ g (random effects model) and adjusted for the presence of publication bias. Results The pattern of reduced internal and elevated external details in aging was robust and consistent across nearly all 21 studies. MCI and – to a greater extent – AD were associated with reduced internal details, whereas the external detail elevation faded with MCI and AD. Although there was evidence of publication bias on reporting of internal detail effects, these effects remained robust after correction. Discussion The canonical changes to episodic memory observed in aging and neurodegenerative disease are mirrored in the free recall of real-life events. Our findings indicate that the onset of neuropathology overwhelms the capacity of older adults to draw upon distributed neural systems to elaborate on past experiences, including both episodic details specific to identified events and non-episodic content characteristic of healthy older adults’ autobiographical narratives.

Retrieving autobiographical memories in autobiographical contexts: Are age-related differences in narrated episodic specificity present outside of the laboratory?

“…To determine whether the magnitudes of effects in the virtual Autobiographical Interview were similar to the findings of a recent meta-analysis (Simpson et al, 2023), we calculated Cohen's d for age differences in frequency of internal and external details, and we asked whether the 95% confidence interval for these effects spanned the Hedge's g reported in Simpson and colleagues (Simpson et al, 2023). As is common among laboratory-based research, we followed up with an independent samples t-test of episodic specificity proportional scores (i.e., internal: total), which accounts for any subtle group variability in overall detail generation.…”

“…One notable difference is in the specificity with which autobiographical events are orally retold, with younger adults tending to provide more event-specific detail relative to older adults, and older adults often providing more background (e.g. semantic), meaning-based information (Levine et al, 2002;Simpson et al, 2023;St. Jacques & Levine, 2007).…”

“…Autobiographical episodic specificity, and its age-associated reduction, is commonly experimentally assessed with the well-known Autobiographical Interview (Levine et al, 2002;Simpson et al, 2023). In the Autobiographical Interview, specific details are characterized as "internal" and include references to the episodic or unique features of the event.…”

“…Although the Autobiographical Interview has undergone several experimental adaptions (e.g., different types of cueing, emotional content of memories, past or future in autobiographical orientation), a consistent finding across tasks is that older adults tend to provide less internal detail relative to young adults, whereas external detail is either statistically equivalent or elevated with older age, which is considered a classic profile of reduced autobiographical episodic specificity (Acevedo-Molina, Matijevic, et al, 2020;Addis et al, 2010;Levine et al, 2002;Madore et al, 2014;Peters & Sheldon, 2020;Wank et al, 2021). A recent meta-analysis including 21 studies comparing young to healthy older adults found large, statistically significant effects for both an age-associated reduction in internal detail (Hedges' g = 1.007) and an inflation in external detail (Hedges' g = -0.799) (Simpson et al, 2023).…”

What about “space” is important for episodic memory?

“…For example, the original Autobiographical Interview protocol elicits one event from each of five lifetime periods (Levine et al, 2002). This practice promotes test standardization and detection of group‐level differences in older adults and clinical groups—especially when medial temporal lobe damage is present (Simpson et al, 2022). Yet the most accessible event from a given life period is—by definition—highly rehearsed and therefore unlikely to be sensitive to inter‐individual differences in healthy adults' episodic autobiographical recollection relative to unrehearsed events (e.g., staged events as in Armson et al, 2021).…”