A Snapshot seminar hosted by the School of Physiology, Pharmacology and Neuroscience

Host: Petra Fischer

The natural environment is filled with many potential sources of sustenance, each with its own richness. A foraging animal, such as a mouse, must engage in behavioural strategies to learn about and select amongst these concurrently available sources. Yet studies in the laboratory have often chosen to reduce the foraging problem to a choice between two, often non-concurrent, sources of reward. By studying foraging decisions in the context of many (>2) options, it becomes possible to investigate these underexplored strategies. In addition, studying multi-option foraging decisions allows for greater differentiation between models that would otherwise make very similar predictions in binary choice. To understand the behavioural strategies adopted by naïve mice when foraging, we developed a paradigm in which freely moving animals sampled from six reward ports arranged around the walls of a large (~2.5m) arena. Each port was concurrently rewarded at a deterministic interval, ranging from 30s to 2400s, and mice were free to sample ports in any order and frequency across sessions lasting 3 hours. Naive mice quickly (< 20 minutes) learned the relative value of all six ports within a single session and were able to collect the majority of available rewards by the end of the first session. When port schedule was subsequently varied across days, mice again rapidly learned the new port values and updated their sampling of ports appropriately. To understand the strategy underlying the mice’s ability to learn and update their behaviour in this task context, we explored a range of decision models of varying sophistication. We found that mice used the history of visits at each port to learn its respective quality and thus match their visit rate to the port in question. Intriguingly, models in which mice updated a ‘global’ model of all ports with experience were more consistent with behaviour than models in which each option was updated independently. Finally, we also recorded dopamine levels in the nucleus accumbens core or dorsal striatum using dLight as mice behaved in the arena, and found that environment context – a change in the global arrangement of ports – was also reflected in levels of dopamine release.

Join Zoom Meeting: https://bristol-ac-uk.zoom.us/j/99114489384?pwd=Y0NZZlUweDI3ZFA1b3F2Ym15U1pFZz09, Meeting ID: 991 1448 9384, Passcode: 329871