Despite their apparent importance for the acquisition of full-fledged human intelligence, mechanisms of intrinsically motivated autonomous learning are poorly understood. How do humans identify useful sources of knowledge and decide which learning situations to approach in the absence of external rewards? While the recognition of this important problem has grown in psychological sciences over the recent years, an intriguing proposition for the possible mechanism comes from artificial intelligence, where efficient autonomous learning is achieved by programming agents to follow the heuristic of maximizing learning progress (LP) during exploration. In this study, we set out to examine the empirical evidence for this idea. Using computational modeling, we demonstrate that humans show signs of following LP while they freely explore and practice a set of multiple learning activities of varying difficulty, including an activity that is impossible to learn. Different approaches to operationalizing the notion of LP and their plausibility in light of empirical data are also discussed. We also show that models combining several types of intrinsic rewards fit better human exploration data than single component models considered so far in theoretical accounts.

Humans are immensely curious and motivated to reduce uncertainty, but little is known about the neural mechanisms that generate curiosity. Curiosity is inversely associated with confidence, suggesting that it is triggered by states of low confidence (subjective uncertainty), but the neural mechanisms of this link, have been little investigated. Inspired by studies of sensory uncertainty, we hypothesized that visual areas provide multivariate representations of uncertainty, which are read out by higher-order structures to generate signals of confidence and, ultimately, curiosity. We scanned participants (17 female, 15 male) using fMRI while they performed a new task in which they rated their confidence in identifying distorted images of animals and objects and their curiosity to see the clear image. We measured the activity evoked by each image in the occipitotemporal cortex (OTC) and devised a new metric of "OTC Certainty" indicating the strength of evidence this activity conveys about the animal versus object categories. We show that, perceptual curiosity peaked at low confidence and OTC Certainty negatively correlated with curiosity, establishing a link between curiosity and a multivariate representation of sensory uncertainty. Moreover, univariate (average) activity in two frontal areas-vmPFC and ACC-correlated positively with confidence and negatively with curiosity, and the vmPFC mediated the relationship between OTC Certainty and curiosity. The results reveal novel mechanisms through which uncertainty about an event generates curiosity about that event.