Abstract
Determining where objects are located in space is a fundamental function of the visual system. When objects are moving, position information can be combined with visual motion signals to improve localization accuracy and compensate for delays in sensory-motor pathways. Previous studies have shown that adults integrate position and motion signals optimally or nearly so (Kwon et al, 2015), but how and when this ability develops is unknown. Up to ~8-11 years, children do not optimally account for uncertainty when integrating different sensory cues. Does this extend to position-motion integration for predictive object tracking? To test this, we measured motion-induced position shifts (MIPS) in adults and children 6- to 10- years old. Participants judged the relative heights of two stimuli left and right of fixation, consisting of 1/f luminance noise presented within a Gaussian contrast envelope. The noise pattern could either drift coherently upward or downward, inducing a perceived position-shift. We varied eccentricity, presentation duration, and speed of the internal pattern motion. To assess precision of position judgments, we included trials in which the internal pattern varied with no coherent direction. We found that even after taking attention lapses into account, children exhibited much larger variability in position judgments (just noticeable differences in children <8 years were ~4 times larger than in adults), indicating larger positional uncertainty. However they showed only a trend for a small increase in MIPS magnitude, smaller than predicted by optimal visual tracking models (Kwon et al, 2015). Taken together, these findings suggests that children do not fully account for their positional uncertainty when combining position and motion signals. Similar to other sensory integration processes (e.g., Dekker et al, 2015), mechanisms underlying position-motion integration might undergo a prolonged developmental trajectory during childhood, and may contribute to improvements in object interception.