Poster #16 - Neural evidence for reinstatement of associative memories in children and adults

• In Event: 1-115 - Poster Session 03
  In Poster Session: PS 03 Section - Biological Processes: Neuroscience and Genetics

Integrative Statement

The ability to retrieve memories to guide behavior during a new experience undergoes pronounced development across middle childhood. Adults can retrieve information from degraded or partially overlapping cues during new experiences, whereas younger children require nearly exact overlap with the original experience to effectively cue retrieval of the memory (Ackerman, 1982). This age-related improvement in so-called ‘retrieval flexibility’ is related to stronger hippocampal recruitment across development (DeMaster, Coughlin, Ghetti, 2016). In adults, functional magnetic resonance imaging (fMRI) evidence has further shown that hippocampal patterns are reinstated during memory retrieval (Mack & Preston, 2016) and may drive reactivation of the initial experience in sensory regions (Polyn et al., 2005). Reinstatement of these perceptual features additionally supports the recollective quality associated with episodic memory (Tulving, 1972; Yonelinas, 1994). Hence, one possible explanation for differences in episodic memory retrieval across development is that children fail to reinstate the sensory aspects of experience, or that the degree of sensory reinstatement is less robust relative to adults.

In the present research, we used pattern analysis of fMRI activation to test how reinstatement of specific memory elements influences subsequent decisions about those experiences at different ages. Children (7-10 years) and adults (18-30 years) learned a series of object-scene and object-face associations (Figure 1A). Following learning, memory retrieval of the target associate was measured with fMRI (Figure 1B). On each retrieval trial, participants were cued with an object, followed by a 9s delay. During the delay, participants were instructed to hold the associated face or scene in their mind in preparation for the upcoming retrieval decision. A probe item then appeared and participants judged whether it was the same (match) or different (mismatch) as the associate they learned previously.

Preliminary findings (N=12 children and 4 adults) show that overall decision accuracy was high in both children (M proportion correct = 0.84, SD = 0.12) and adults (M proportion correct = 0.96, SD = .03), suggesting that both groups were able to retrieve and maintain the memories in preparation for judging the match/mismatch decision probe. To decode reinstatement of the paired associate prior to the decision probe, we trained a classifier to discriminate patterns of fMRI activation associated with perception of faces and scenes for each participant using a separate visual localizer task. Classifier accuracy was comparable (Means = 0.81) and reliably above chance (ps < .001) in children and adults, meaning that our classifier was similarly able to discriminate perception of faces from scenes in both age groups. We then applied the classifier to the retrieval task. Across the delay, we quantified evidence of reinstatement of the target category (i.e., matching the category of the cued item) versus nontarget category (e.g., face if target was a scene). We observed significantly greater reinstatement of the target relative to the nontarget category in ventral temporal cortex in both adults and children (Figure 2). These findings extend previous research with adults and provide initial support that children can reinstate perceptual features of their prior experiences in preparation for upcoming decisions.

Authors

• Nicole Varga, University of Texas, Austin
  Presenting Author

• Hannah Roome, University of Texas at Austin
  Non-Presenting Author

• Alison R Preston, The University of Texas at Austin
  Non-Presenting Author

• Meg Schlichting, University of Toronto
  Non-Presenting Author