Divergent Strategies for Learning in Males and Females.

A frequent assumption in value-based decision-making tasks is that agents make decisions based on the feature dimension that reward probabilities vary on. However, in complex, multidimensional environments, stimuli can vary on multiple dimensions at once, meaning that the feature deserving the most credit for outcomes is not always obvious. As a result, individuals may vary in the strategies used to sample stimuli across dimensions, and these strategies may have an unrecognized influence on decision-making. Sex is a proxy for multiple genetic and endocrine influences on behavior, including how environments are sampled. In this study, we examined the strategies adopted by female and male mice as they learned the value of stimuli that varied in both image and location in a visually cued two-armed bandit, allowing two possible dimensions to learn about. Female mice acquired the correct image-value associations more quickly than male mice, preferring a fundamentally different strategy. Female mice were more likely to constrain their decision-space early in learning by preferentially sampling one location over which images varied. Conversely, male mice were more likely to be inconsistent, changing their choice frequently and responding to the immediate experience of stochastic rewards. Individual strategies were related to sex-biased changes in neuronal activation in early learning. Together, we find that in mice, sex is associated with divergent strategies for sampling and learning about the world, revealing substantial unrecognized variability in the approaches implemented during value-based decision making.