Evidence for different hippocampal codes yet similar neocortical reinstatement of memories across development

• In Event: Functional development of episodic memory in the hippocampus

Abstract

Episodic memory entails the ability to form and retrieve representations of specific past events. Such memories depend on a network of brain regions, including cortical areas that can encode and store rich representations of the initial experience. Hippocampus binds these cortical traces together (Marr, 1971) to ultimately reflect the unique features of even overlapping events. Those hippocampal codes are thought to drive neocortical reinstatement during later memory retrieval (Dickerson & Eichenbaum, 2009)—possibly giving rise to the sense of “reliving” attributed to episodic memory (Tulving, 1972). Functional magnetic resonance imaging (fMRI) evidence in adults supports this proposal, showing that hippocampus works in concert with neocortical regions to guide reinstatement of specific memories (Mack & Preston, 2016). However, how this happens in children is less clear. Hippocampus, neocortex, and their connections undergo a protracted course of development (Calabro, Murty, Jalbrzikowski, Tervo-Clemmens, & Luna, 2019; DeMaster & Ghetti, 2013), suggesting that memory retrieval may be characterized by age-related differences in the nature of hippocampal organization, the profile of neocortical reinstatement, and/or the translation of information between these regions. We assessed these possibilities in the present research.
Children (7-10 years; N = 27) and adults (18-30 years; N = 25) learned a set of object-face and object-scene associations to criterion (Figure 1A). The same face or scene was paired with multiple objects during learning, enabling assessment of how hippocampus represents overlapping events. Following learning, participants were cued with an object, followed by a delay in which they were instructed to hold the associated face or scene in mind in preparation for an upcoming decision. We measured the degree to which hippocampal representations differentiated overlapping memories during presentation of the object retrieval cue, as well as the specificity and degree of ensuing cortical reinstatement of the associated face or scene during retrieval (Figure 1B).
Decision accuracy was high in both groups (Figure 2A), suggesting that participants were able to retrieve the memories in service of an upcoming decision. Interrogating representation of overlapping cues in hippocampus, we observed significant differentiation in children but not adults (Figure 2B)—such that related memories (objects paired with the same face) were stored as less similar than unrelated memories (objects paired with different faces). Such a representational strategy may be especially important earlier in development as a means of resolving mnemonic interference among related events. Moreover, participants of all ages reinstated the related memories in neocortical regions, as measured at both the specific item (Peter Pan vs. other faces; ps < .07) and category (faces vs. scenes; ps < .01) levels. These detailed neocortical representations were deployed to support retrieval in both groups, such that item reinstatement in parietal cortex and category reinstatement in ventral temporal cortex (VTC) predicted trial-by-trial retrieval success (Figure 2C). Greater hippocampal differentiation during the cue in children was also associated with greater neocortical reinstatement during the delay (Figure 2D). These findings indicate that neocortical signatures of memory reinstatement show developmental continuity, whereas the hippocampal representations that govern how memories are organized change with development.

Authors

• Nicole Varga, University of Texas at Austin
  Presenting Author

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

• Robert Molitor, University of Oregon
  Non-Presenting Author

• Lucia Martinez
  Non-Presenting Author

• Elizabeth Hipskind
  Non-Presenting Author

• Michael Mack, University of Toronto
  Non-Presenting Author

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

• Meg Schlichting, University of Toronto
  Non-Presenting Author