The transformation of memory from a system that stores discrete events to one that codes the commonalities across related events is critical to higher-order cognition, enabling individuals to extend beyond direct experience to support flexible behaviors such as reasoning, optimal decision making, and prospection. Explicating how relational memory representation emerges across development not only has important implications for memory behaviors (i.e., retrieval of past events), but impairments in memory formation are also associated with academic outcome and several mental health disorders that onset before adulthood, including depression, autism, and schizophrenia. Yet despite the critical need to elucidate how the neural mechanisms underlying the formation of integrated memory traces emerges across development, measuring the formation of integrated memories at the representational level has proved challenging. Consequently, no study to date has examined the neural mechanisms driving the formation of integrated memories in development. The proposed project will directly address this gap. It builds on theory and empirical work with adults suggesting that the formation of integrated memories is supported by the anterior hippocampus (aHPC) and medial prefrontal cortex (mPFC), as well as research showing that this aHPC-mPFC circuit continues to develop through the third decade of life, thus underscoring the critical need to examine how changes in these structures relates to the protracted emergence of memory integration. We predict that memory integration and reasoning in childhood may therefore rely on a more effortful strategy supported by the posterior HPC (pHPC) and lateral PFC (lPFC), wherein memories are represented separately from one another and then recombined when necessary. In contrast to children, we hypothesize that adolescents will begin to rely on the aHPC?mPFC network, but immature connection and rapid changes in development of these regions will lead to less reliable memory integration, such that traces continue to be represented discretely. Through completion of one experiment, children (7-8 years), adolescents (13-14 years), and adults (25-30 years) will complete a paired associate inference task assessing memory integration (scanned) and subsequent reasoning (not scanned). Multimodal neuroimaging techniques (structural MRI, high-resolution) and advanced analysis methods (multi-voxel pattern analysis, relational similarity analysis) will be used to test the neural mechanisms supporting memory integration (Aim 1) and how these mechanisms relate to representational change with age (Aim 2). Results from this project would provide a substantial contribution to theory on the neurocognitive development of representational capacities, and have the potential to provide critical insight into interventions designed to improve complex behaviors that rely on memory integration, especially within populations with deficits in this learning ability (i.e., individuals with autism, depression, or schizophrenia).
By representing the connections among events, memory integration enables extension of knowledge beyond direct experience to support not only learning, but other complex behaviors that operate over memory representations, including reasoning, decision making, and prospection. A growing body of empirical work has shown that difficulties with memory and reasoning in development are associated with poor academic outcomes, as well as with several mental health disorders with developmental onsets, including depression, autism, and schizophrenia. Isolating the neural mechanisms that underlie changes in memory integration from childhood to adulthood may thus help identify markers of atypical development and contribute to interventions that may improve cognitive and academic outcomes.
