Hippocampal contribution to self-referential encoding in children

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

Abstract

Background: Information that is encoded in relation to the self tends to be better remembered. This self-referential effect of memory has been previously shown in adults and in children, for both recognition memory and source memory. Previous fMRI studies in adults suggest that regions in the default network such as the medial prefrontal cortex (MPFC) support successful self-referential encoding (Leshikar and Duarte, 2013). How different regions in the episodic memory system support self-referential encoding in children is less well understood. We hypothesized that children may show less engagement of the medial frontal cortex and rely more on the hippocampus during successful self-referential encoding.

Methods: Our study investigated the development of neural correlates of self-referential encoding in a group of children (7-11 years of age, N = 28) and adults (18-35 years of age, N= 41) using a source memory paradigm. Participants viewed pictures of common color objects displayed on a background image while inside the MRI scanner and answered either a self-referential question (Do you like the object?) or a semantic question (Is the object a living thing?) for each object. After the scan, participants were tested on their recognition memory of the objects, and on their source memory about the background image associated with the object and the encoding question (self or semantic).

Results: For both the adult and children groups, self-referential encoding resulted in better recognition memory accuracy (regardless of source correctness) (ps < .001, Figure 1A) and higher numbers of trials with both source information correct (ps < .001). In children, self-referential encoding resulted in a higher proportion of trials with correct source out of the remembered trials compared to semantic encoding condition (p = .03; Figure 1B). Adults had higher numbers of trials with correct source compared to children in the semantic condition (p = .018) but not in the self-referential condition (p = .5). These results suggest that self-referential encoding facilitates source memory in children and narrows performance gap with adults.

Across all participants and both encoding conditions, we observed the expected pattern of subsequent memory regions including the bilateral hippocampus, fusiform gyrus, dorsal lateral prefrontal cortex and MPFC (Figure 2A). Successful self-referential encoding compared to semantic encoding activated regions in the default network including MPFC, posterior cingulate cortex, left inferior parietal cortex and left hippocampus (Figure 2B). Between-group comparison revealed higher activations in children in bilateral hippocampus for self > semantic (Figure 2C). ROI analysis showed that children had higher activations in the left hippocampus than adults for trials with correct source during self-referential encoding (p =.006), but not during semantic encoding (p =.4) (Figure 2C). MPFC and other cortical regions did not show differences between children and adults during either encoding conditions.

Conclusions: These results suggest that self-referential encoding facilitates source memory in children and can be an effective learning strategy. Children compared to adults rely preferentially on the hippocampus during self-referential encoding. These findings add novel insights into the role of the hippocampus during self-related memory processes in children.

Authors

• Hilary Sweatman, Montreal Neurological Institute, McGill University
  Presenting Author

• Ross Lawrence, Johns Hopkins University
  Non-Presenting Author

• Xiaoqian Chai, McGill University
  Non-Presenting Author