Poster #3 - How Does the Similarity of Related Experiences Impact Memory Representation Over Development?

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Attention, Learning, Memory

Integrative Statement

Previous research suggests that memories shift from being rigid to flexible across development. For example, the capacity to make decisions that span across separate but related experiences improves through adolescence (Schlichting et al., 2017). We propose that the refinement of flexible memory use results from parallel development in memory representation. In particular, we suggest that young adolescents and adults store memories for related experiences in fundamentally different ways due to the distinct mechanisms they engage during initial encoding. In accordance with prior work (Shohamy & Wagner, 2008; Zeithamova, Dominick & Preston, 2012), we hypothesize that when encoding two related associations—AB and AC, where ‘AB’ denotes a studied association between arbitrarily paired items A and B—adults will engage in integrative encoding. Namely, encountering AC triggers the reinstatement of the related AB memory, leading to integration across associations (AB→←AC) to emphasize their common elements—a process shown to rely on hippocampus and medial prefrontal cortex in adults (Schlichting & Preston, 2015). In contrast, we hypothesize that this same reactivation in adolescents drives memory differentiation, exaggerating the unique features across related AB and AC memories (AB←→AC), due to the relative immaturity of their brains’ “integration circuits” (Simmonds et al., 2014).
To assess developmental differences in the representation of related events, we modified the associative inference paradigm (Preston et al., 2004) to include conditions that differentially promote the reactivation of related events (Figure 1). Participants were healthy adolescents (11-13 years) and adults (25-35 years) who first learned overlapping AB and AC object-object pairs. We manipulated the similarity between AB and AC events by presenting half of the pairs on the same (high) or different (low) scene backgrounds. We hypothesized that high similarity would boost reactivation of AB during AC encoding, thereby exaggerating integration/differentiation. Participants then studied a new set of related DB associations. Our primary measure of representation was the nature of errors made during the critical DB test. We inferred integration from the tendency to erroneously endorse the related (C) as compared with the unrelated lure, and differentiation from the opposite tendency. Finally, participants were tested on their memory for AB/AC pairs (cued recall) as well as the scene backgrounds (old/new recognition).
Preliminary results (N=13 adolescents and 22 adults) revealed better memory for the overlapping AC associations in the high versus low similarity condition (F(1,43)=5.31, p=0.03; Figure 2A), suggesting that increased overlap has a facilitative effect at both developmental stages. Contrary to our predictions, we found significant evidence for differentiation across both ages groups and similarity conditions (all p<0.001; Figure 2B). We also observed qualitatively superior overall scene memory in adolescents versus adults (F(1,33)=2.74, p=0.11; Figure 2C), consistent with the notion that adolescents have a reduced ability to ignore “irrelevant” information. Taken together, our findings suggest that both age and cross-event similarity impact memory for related associations. Moreover, our results suggest that systematic differences in the distribution of cognitive resources during memory formation may give rise to important developmental differences in what is later remembered about experiences.

Authors

• Wangjing Yu, University of Toronto
  Presenting Author

• Katherine D. Duncan, University of Toronto
  Non-Presenting Author

• Meg Schlichting, University of Toronto
  Non-Presenting Author