Reward Expectation Reduces Representational Drift in the Hippocampus

Abstract

SummarySpatial memory in the hippocampus involves dynamic neural patterns that change over days, termed representational drift. While drift may aid memory updating, excessive drift could impede retrieval. Memory retrieval is influenced by reward expectation during encoding, so we hypothesized that diminished reward expectation would exacerbate representational drift. We found that high reward expectation limited drift, with CA1 representations on one day gradually re-emerging over successive trials the following day. Conversely, the absence of reward expectation resulted in increased drift, as the gradual re-emergence of the previous day’s representation did not occur. At the single cell level, lowering reward expectation caused an immediate increase in the proportion of place-fields with low trial-to-trial reliability. These place fields were less likely to be reinstated the following day, underlying increased drift in this condition. In conclusion, heightened reward expectation improves memory encoding and retrieval by maintaining reliable place fields that are gradually reinstated across days, thereby minimizing representational drift.

Deep learning-based location decoding reveals that across-day representational drift is better predicted by rewarded experience than time

“…Importantly, this study examined drift on a shorter 4-day time-scale than the ∼ 30 days that we considered. Our study is nevertheless consistent with the results from 17 in the sense that both conclude that reward is a strong driver of drift dynamics.…”

“…Here, we distinguished between both different aspects of active experience and replayed experience. Another recent study using two-photon calcium imaging of hippocampal neural populations in rats performing a navigational task found that drift was slow when reward expectation was high, and lowering the animals' reward expectation sped up representational drift 17 . Importantly, this study examined drift on a shorter 4-day time-scale than the ∼ 30 days that we considered.…”

“…Though other research has been conducted which examined drift in local field potentials 31 , this work was focused on obtaining stable low dimensional representations of LFP, rather than quantifying the drift dynamics. Other studies which have quantified drift as some measure of the change of spiking behaviors of single units 2,12,16,17 . Our use of LFPs is complementary to these studies, as it captures the aggregate electrical activity of populations of neurons within a localized region.…”

“…Importantly, both these studies equated 'experience' with 'time spent on the maze'. This leaves open the question of whether different aspects of the animal's task experience (time doing task, number of task trials, rewards) may have different impacts on representational drift 17 .…”

Single field evolution rule governs the dynamics of representational drift in mouse hippocampal dorsal CA1 region

“…4) It should differentially affect place fields with distinct retained durations to result in the convergently growing retention probability (Figures 5C and 5D). Memory must be selected for long-term storage, with more valuable experiences, such as those related to rewards [66][67][68] , being more likely to be stabilized. From a computational perspective, RD delineates the process of neuronal representations optimizing toward an optimal solution 31 , which theoretically exhibits convergent dynamics 56 .…”