We aim to study how populations of neurons in the hippocampus support navigational learning and decision- making. During awake replay, hippocampal neurons are active in speeded-up sequences that reflect spatial trajectories. These trajectories are highly non-local, in the sense that they activate representations of space far from the current location and throughout an environment. Thus, replay is a uniquely attractive mechanism, at least within current neuroscience understanding, for taking information about something that is happening locally, such as the discovery of a reward, or the discovery of a blocked route, and broadcasting that information out through the brain's representation of the rest of the environment. This is the specific hypothesis we aim to test in this proposal. First, we want to understand the circuit mechanisms that could lead to reward having an effect on hippocampal replay. We previously showed that reward increases lead to increases in the rate at which replays occur, and specifically for reverse replay. Now we propose to test the circuitry behind this directly, by suppressing the activity of dopaminergic neurons in the ventral tegmental area, since these neurons are known to signal an important reward-related quantity called reward prediction error. Models have proposed that this should regulate replay and our preliminary data support this hypothesis. Second, it remains unknown how replay processes changes to the structure of the environment, such as the blocking of a path. We will test the effect on replay of small changes to the environment. The innovation is that each change is designed to have either a small or a big effect on navigation, depending on how easy it is to circumvent. All the changes experienced by the animal will need to be remembered, so that if replay is mainly a memory consolidation mechanism, then it should not distinguish the two circumstances. However, only the globally significant changes require the broadcasting of navigational information out through the environment. Our preliminary data show that globally important changes cause an increase in replay-associated events while others do not. Third, it remains unknown how replay supports planning in more complex environments, that are more realistic than the small arenas or tracks usually studied. Theoretical studies suggest that planning at multiple spatial scales can support navigation in more complex environments. More ventral parts of the hippocampus represent space at increasingly large scales, but replay outside of dorsal has not been studied. Our preliminary data replicate increased place field size in intermediate hippocampus. Moreover, we demonstrate replay there for the first time. We expect to establish whether and how replay along the longitudinal axis of the hippocampus supports planning at multiple spatial scales in the more complex task. Altogether, our proposal will provide new insight into replay function through the lens of navigational learning.
Despite the prevalence of memory disorders associated with damage to the hippocampus, from such conditions as Alzheimer's disease, aging, stroke and epilepsy, we lack a mechanistic understanding of how hippocampal neurons encode and retrieve memories, which could help us address these problems. However, recently developed techniques for recording from large numbers of neurons simultaneously in awake, freely behaving animals, have begun to describe 'replay' of patterns of activity across hippocampal neurons that may provide a model of memory processing. This study will take these results further by investigating mechanisms that initiate replays selectively, and the ways in which replay functions in tasks incorporating realistic challenges, including the need to respond flexibly to changes, and the need to plan over long distances and at multiple spatial scales.
