Evidence for holistic episodic recollection via hippocampal pattern completion

Abstract

Recollection is thought to be the hallmark of episodic memory. Here we provide evidence that the hippocampus binds together the diverse elements forming an event, allowing holistic recollection via pattern completion of all elements. Participants learn complex ‘events' from multiple overlapping pairs of elements, and are tested on all pairwise associations. At encoding, element ‘types' (locations, people and objects/animals) produce activation in distinct neocortical regions, while hippocampal activity predicts memory performance for all within-event pairs. When retrieving a pairwise association, neocortical activity corresponding to all event elements is reinstated, including those incidental to the task. Participant's degree of incidental reinstatement correlates with their hippocampal activity. Our results suggest that event elements, represented in distinct neocortical regions, are bound into coherent ‘event engrams' in the hippocampus that enable episodic recollection—the re-experiencing or holistic retrieval of all aspects of an event—via a process of hippocampal pattern completion and neocortical reinstatement.

Knowing what from where: Hippocampal connectivity with temporoparietal cortex at rest is linked to individual differences in semantic and topographic memory

“…Meta-analytic decoding linked the left hemisphere with terms such as semantic , language , words , while the right hemisphere was associated with attention and visual . These results support contemporary theories suggesting the hippocampus provides a mechanism for binding disparate representations in different cortical regions (Marr, 1971;Damasio, 1989;Teyler & Rudy, 2007;Horner et al, 2015) and learning meaningful configurations across domains including spatial and conceptual representations (Dusek & Eichenbaum, 1997;Eichenbaum & Cohen, 2004). Moreover, since semantic knowledge reflects more long lasting knowledge of the world around us, the association between anterior HC and these types of process is consistent with the observation that anterior HC maintains traces of prior episodic information for longer than does the posterior HC (Ritchey, Montchal, Yonelinas, & Ranganath, 2015).…”

“…Horner, Bisby, Bush, Lin, & Burgess, 2015). It is increasingly recognised that differences in connectivity along the posterior-anterior axis of the hippocampus could give rise to functional specialisation (Poppenk, Evensmoen, Moscovitch, & Nadel, 2013;Strange, Witter, Lein, & Moser, 2014;Ranganath & Ritchey, 2012;Chase et al, 2015).…”

Successful cuing of gender source memory does not improve location source memory

“…Even though in Experiment 2 we warned people that they would be tested on this information, this may not have prompted a thorough encoding approach that encourages ensemble encoding (e.g., Murnane & Phelps, 1993, 1994. To underscore this point, recent work by Burgess (2013, 2014; see also Horner, Bisby, Bush, Lin, & Burgess, 2015) demonstrated stochastic dependence among triplets of encoded information that were all mutually and focally encoded. In this work, people learned triplets such as Obama-Pencil-Supermarket and were asked to associate them with a considerable encoding time.…”

Eye-tracking the time‐course of novel word learning and lexical competition in adults and children

“…When listeners encounter incoming speech and relate it to their mental model, they are undoubtedly informed by the cortically stored lexical knowledge of the words that they hear. But nonetheless, speculatively, the generation of the mental model might depend much more heavily on the hippocampal episodic memory (e.g., Horner et al, 2015) of the novel words, alongside the short-term memory of the visual scene. Thus the discrepant results that we see between the VWP and response time tasks may also be a product of different weightings applied to hippocampal and neocortically mediated representations of newly learned words as a consequence of the memory demands of the task.…”