Stronger effects of motion priors on sensitivity and bias in children

According to Bayesian inference models, perception is affected by incoming input (i.e., likelihood) and by experience (i.e., priors). How perceptual inference is modified during development remains unclear. This gap is especially important when considering the two fundamental dimensions of perception-sensitivity and bias, which may evolve differently in children. In adults, a natural prior for cardinal directions affects both dimensions, resulting in lower motion direction thresholds and larger estimation biases for cardinal versus oblique directions. Here, we investigate how this natural prior for cardinal directions influences sensitivity and bias in the perception of motion direction in adults (n = 28) and children (n = 25, age range = 7-8) with identical stimuli presentation durations, and in an additional group of children with longer stimuli presentation durations to examine whether providing more time to integrate sensory information affects children's reliance on priors (long duration: n = 25, age range = 6.5-10.5). First, we tested individual coherence discrimination thresholds for near oblique versus near horizontal directions. Then, participants moved a mouse to estimate the at-threshold motion directions. As expected, children had higher overall thresholds compared to adults. All groups showed a prior to horizontal - significantly lower thresholds and larger estimation biases for near horizontal motion compared to near oblique motion. Critically, these differences, between oblique and horizontal, were significantly more pronounced in both groups of children compared to the adult group. These results reveal a larger impact of priors in children and highlight the developmental changes in sensitivity and bias, underscoring the importance of studying perceptual inferences in children.