Abstract
The visual system balances two complementary processes: temporal integration to construct stable representations, and segmentation to resolve change. This balance point can be captured by the ‘Temporal Integration Window’. If two stimuli fall within the same TIW, they are integrated; otherwise, segmented (VanRullen 2016). Previous work has suggested that temporal processing is slower in infants (Farzin et al. 2010). We tested this using a novel, gaze-based ‘pop-out’ search paradigm to measure TIWs in 5 to 13 month olds (N= 34; mean= 8.3 months, SD= 2.7 months). Trials consisted of a 4 s sequence of two alternating displays (ABAB…) containing search stimuli distributed in a 4x4 grid. Infants were tasked with finding a circle-shaped target - if fixated, the trial was coded as correct. On integration trials, display A consisted of 8 half-circles, and B, 9 half-circles. Alone, displays A and B contained no target, but if integrated, the circle target was formed by the alignment of two complementary half-circles. On segmentation trials, display A consisted of 15 half-circles, and B, 15 complementary half-circles, plus one full circle. Combined, this appeared as a field of undifferentiated circles, but if segmented, the single circle target was visible. Pace (SOA) was dependent on trial (integration: 33, 67, 133 ms; segmentation: 67, 133, 267 ms). Shorter SOAs facilitate integration performance, and longer, segmentation. TIWs were estimated by finding the intersection of the integration and segmentation performance functions for the group (Wutz et al. 2016). Data collection is ongoing, but preliminary analyses show infants have a TIW of 128 ms, much longer than our comparison group of adults (TIW= 55 ms), or our previous work with adults and 5-7 year-olds (TIWs of 73 and 68 ms, respectively) (Freschl et al. 2019). This suggests that infants are relatively slow; a conservative bias toward stability over speed.