Conference on Cognitive Computational Neuroscience

Inspired by a seminal experimental work involving an odor-in-context association task conducted on rats (Panoz-Brown et al., 2016), we performed a computational study to investigate the interaction of long-term episodic processes with short-term dynamics (e.g., recency). In the task, rats were presented with odor pairs in two contexts (old vs. new), and rewarded for identifying the new item-in-context. To mechanistically explain the interplay of episodic memory with short-term memory processes, we built a spiking model consisting of two reciprocally connected networks that store contextual and odor information. Context-item coupling between the two networks was induced by Bayesian-Hebbian plasticity with eligibility traces to account for reward-based learning. By task design, old items were presented earlier than new items (old items featured low recency), putting episodic memory in conflict with recency effects. However, by simulating an alternative task at which old items were, instead, encoded later than new items (old items featured high recency), we confounded episodic memory with effects of recency. The model showed that high recency of old items enhanced item-in-context memory by boosting their sense of oldness. Overall, the study provided a framework for understanding the varying implications of different memory effects for observed behavior in episodic memory paradigms.