Binding of Episodic Memories in the Rat

Abstract

Summary People remember an event as a coherent scene [1-4]. Memory of such an episode is thought to reflect binding of a fully integrated representation, rather than memory of unconnected features [4-7]. However, it is not known if rodents form bound representations. Here we show that rats remember episodes as bound representations. Rats were presented with multiple features of unique episodes at memory encoding: what (food flavor), where (maze location), source (self-generated food seeking–running to the food site– or experimenter-generated food seeking –placement by the experimenter at the food site), and context (spatial cues in the room where the event occurred). After a delay, the trial continued with a memory assessment in which one flavor replenished at the self-generated- but not at experimenter-generated-locations. We presented rats with multiple overlapping features, in rapid succession, to ensure that successful memory retrieval required them to disambiguate multiple study episodes (using two rooms). We found that binding is resistant to interference from highly similar episodes and survives long retention intervals (~1 week). Our results suggest that multiple episodic memories are each structured as bound representations, which suggests that nonhumans represent episodic memories using a structure similar to that of people. This finding enhances the translational potential for utilizing animal models of episodic memory to explore the biological mechanisms of memory and validate therapeutic approaches for treating disorders of memory.

Animal models of episodic memory

“…Rats revisited the replenishment location at a higher rate than the nonreplenishment location while avoiding revisits to chow locations ( Figure 3B). These data are consistent with the hypothesis that rats remember the source of encoded information (Crystal & Alford, 2014;Crystal & Smith, 2014;Smith, Slivicki, et al, 2017;Smith et al, 2016). Moreover, source memory is quite long-lasting ( Figure 3C), surviving retention-interval challenges of at least 1 week (Crystal & Alford, 2014;Crystal & Smith, 2014).…”

“…The rats revisited the replenishing chocolate location in the memory assessment at a higher rate than the nonreplenishment chocolate location when we used a memory load of two rooms, at a level of proficiency similar to that observed when the memory load was one room (Crystal & Smith, 2014). Moreover, source-memory performance was resistant to interference from highly similar episodes and survived long retention intervals (about 1 week; Crystal & Smith, 2014).…”

“…Moreover, source-memory performance was resistant to interference from highly similar episodes and survived long retention intervals (about 1 week; Crystal & Smith, 2014). These results suggest that multiple episodic memories are each structured as bound representations.…”

“…We developed an animal model of source memory (Crystal & Alford, 2014;Crystal, Alford, Zhou, & Hohmann, 2013;Crystal & Smith, 2014;Smith, Dalecki, & Crystal, 2017;Smith, Slivicki, Hohmann, & Crystal, 2017;Smith et al, 2016;reviewed in Crystal, 2016a; see Basile & Hampton, 2017, for an example of source memory in rhesus monkeys). Source memory is memory for the origin of episodic memories (Janowsky, Shimamura, & Squire, 1989;Johnson, Hashtroudi, & Lindsay, 1993;Mitchell & Johnson, 2009).…”

“…We used our source memory approach to test the hypothesis that rats remember episodic memories as bound representations (Crystal & Smith, 2014). The binding hypothesis proposes that the source memory for the event is stored with the remaining elements of the episodic event in an integrated manner.…”