Abstract:

Updating outdated memories with newly acquired information is an important ability. To ensure less interference from outdated memories, one either trims old memory traces or differentiates between novel and old memories. Although the age-related decline in episodic memory has been well-established, aging effect on episodic memory updating and its underlying cognitive mechanisms are less understood. Memory differentiation and memory interference may act as two different mechanisms that underlie episodic memory updating. To bridge this research gap, we investigated the different updating memory patterns in young and older adults. Furthermore, we intended to clarify the updating differences in item and source memory, which are the two essential elements embedded in episodic memory. By implementing an adapted AB-AC memory updating paradigm, we posited that young adults would comparably attain A-B and A-C memory. In contrast, older adults would experience C memory intrusions in the A-B memory test, indicating memory interference.

We examined episodic memory updating for a sample of 30 young and 30 older adults randomly selected from the university and neighboring community. On Day 1, participants learned 36 naturalistic A-B pairings and completed the encoding test immediately after learning. On Day 2, participants first reviewed the A-B pairs before being introduced to 36 novel A-C pairings (where C’s items or sources were different from those in B), following which they completed the Day 2 encoding test. Finally, all the participants returned to the laboratory and completed the A-B and A-C tests on Days 3 and 5. To elucidate the updating memory patterns between the two age groups, we classified their responses into target, competitor, and lure categories to test the group-level differences in memory updating.

We ran a 2 (age: young adults and older adults) × 2 (association: A-B and A-C) × 3 (category: target, competitor, and lure) × 2 (time: Days 3 and 5) repeated analysis of variance to examine the different memory updating pattern between the two age groups. Expectedly, there were significant age × category and age × association × category interaction effects in both item and source memory tasks (Table 1). More precisely, young adults performed comparatively well in A-B and A-C memory tests (Item memory: t(57) = −1.98, p = 0.053; Source memory: t(57) = −0.24, p = 0.812) and showed no significant memory intrusions (Item memory: t(57) = −0.31, p = 0.756; Source memory: t(57) = −0.735, p = 0.465) and distortions (Item memory: t(57) = 4.24, p < 0.001; Source memory: t(57) = 2.05, p = 0.045). In comparison, older adults performed better in the A-C memory test than in the A-B memory test (Item memory: t(57) = −6.75, p < 0.001; Source memory: t(57)= −3.88, p < 0.001) and showed B intrusions in the A-C memory test (Item memory: t(57) = 5.65, p < 0.001; Source memory: t(57) = 4.55, p < 0.001) but no C intrusion in the A-B test (Item memory: t(57) = −0.31, p = 0.756; Source memory: t(57) = −0.88, p = 0.382).

We found that young adults could strengthen their novel memories without inhibiting outdated ones and simultaneously maintained A-B and A-C memory traces. However, older adults updated their memories by overwriting their previously-stored A-B memory traces. The findings illuminate episodic memory updating and its underlying cognitive mechanisms, whereby young adults update memories via differentiating old with new memory traces. However, older adults inhibit previous memory traces when confronted with two pieces of competing information. Notably, separate mechanisms apply to item and source memories and show a long-term effect.

Key words: memory updating, episodic memory, aging, retrograde interference