Links between autobiographical memory richness and temporal discounting in older adults

When making choices between smaller, sooner rewards and larger, later ones, people tend to discount future outcomes. Individual differences in temporal discounting in older adults have been associated with episodic memory abilities and entorhinal cortical thickness. The cause of this association between better memory and more future-oriented choice remains unclear, however. One possibility is that people with perceptually richer recollections are more patient because they also imagine the future more vividly. Alternatively, perhaps people whose memories focus more on the meaning of events (i.e., are more “gist-based”) show reduced temporal discounting, since imagining the future depends on interactions between semantic and episodic memory. We examined which categories of episodic details – perception-based or gist-based – are associated with temporal discounting in older adults. Older adults whose autobiographical memories were richer in perception-based details showed reduced temporal discounting. Furthermore, in an exploratory neuroanatomical analysis, both discount rates and perception-based details correlated with entorhinal cortical thickness. Retrieving autobiographical memories before choice did not affect temporal discounting, however, suggesting that activating episodic memory circuitry at the time of choice is insufficient to alter discounting in older adults. These findings elucidate the role of episodic memory in decision making, which will inform interventions to nudge intertemporal choices.

the IRI Perspective-Taking subscale, as this was the one we expected to be most strongly related to temporal discounting based on previous research. The LOT-R tests for optimism, which tends to be elevated in older adults 2 , and may be related to future-oriented decision-making 3 . The GDS was included as a screening tool, because symptoms of depression are associated with deficits in memory ability, especially in positive memory recall 4 . Therefore, anyone with a GDS score of 9 or above (out of 15), indicating moderate or severe depression, was excluded. Finally, the VVIQ instructs participants to imagine different scenarios in order to measure individual differences in self-reported imagery vividness. VVIQ scores have been shown to be correlated with temporal discounting 5 . We examined Spearman correlations between scores on these questionnaires and age, as well as with the size of the effect of the positive memory recall manipulation in our study.
Outliers (scores that were more than 2.5 SD from the mean) were removed. There was 1 outlying VVIQ score (3.16 SD below the mean), and 1 outlying IRI-perspective-taking score (3.17 SD below the mean), leaving n = 33 for those analyses.
List of cues used to elicit positive memory recall on Day 1

Participant ratings of memories were not associated with temporal discounting
Although objective assessments of autobiographical memories were associated with temporal discounting, none of the participants' subjective ratings of their memories were significant predictors of temporal discounting: there was no association with the Day 1 rating of "similarity between past and present self" (ρ = 0.08; p = 0.661), the Day 1 rating of "feeling

Participant ratings of memories did not predict change in choice following recall of memories
We investigated whether any subjective characteristics of the memories themselves, as rated by the participants, could predict the extent to which retrieving them was effective in reducing temporal discounting rate. We conducted a series of mixed-effects logistic regressions to see which ratings could predict choice of delayed reward on a trial-by-trial basis, controlling for the subjective value of rewards computed assuming the Control condition discount rate. None of the ratings were significant predictors of choice: there was no effect of the Day 1 rating of

Self-reported perspective-taking is associated with reduction in discounting following memory recall across participants
The VVIQ, which measures individual differences in self-reported imagery vividness, was not correlated with temporal discounting rate (n = 33, 1 outlier excluded; ρ = -0.09; p = 0.627) or with the effect of memory recall on temporal discounting (ρ = -0.03; p = 0.883).
However, VVIQ was associated with age (ρ = -0.71; p < 0.001) in the expected direction, with younger age associated with more vivid mental imagery. With respect to autobiographical memory details, VVIQ was associated with the overall number of internal details (ρ = 0.37; p = 0.033), but not the perception-based detail ratio score (ρ = -0.02; p = 0.903).
The LOT-R, which measures optimism, was associated with the effect of memory recall