Conference on Cognitive Computational Neuroscience

In principle, animals should vary their motivation-state over time to take advantage of changes in their general environment. This is difficult to study, however, because motivation-states are latent phenomena that must be inferred from behaviour. Here, we implement a hybrid Hidden Markov Model General Linear Model to show evidence for two meaningfully distinct and autocorrelated motivation-states in macaque monkeys performing a decision-making task. We show that changes in motivation-state are correlated with neural activity in dorsal raphe nucleus (DRN), and that minimally invasive and reversible disruption of DRN reduces the frequency of such changes. We therefore suggest that it is both possible to quantify changes in persistent behavioural states like motivation, and that DRN has a fundamental role in making these changes happen.