Abstract
Infants' psychophysical functions measured by variants of Forced Choice Preferential Looking (FPL) rarely asymptote at one. This suggests that infants orient randomly on an unknown fraction of trials even to stimuli well above threshold. Using a simple regression technique described by Dannemiller (2003) and Lufti (1995) we re-evaluated data from N=166 infants whose abilities to discriminate changes in heading direction had been studied using a modified version of FPL. As expected, average sensitivity to the display depicting a changing heading angle was larger as the size of the change increased. In contrast, there was little evidence that infants responded to changes in the spatial position of the Focus of Expansion (FOE), nor was there a systematic left or right looking bias. For all measures, however, there were large differences between individuals and between different visits by the same individual. Taken together, these data confirm previous findings that prelocomotor infants detect heading direction changes of 20 deg or more, but cast doubt on the possibility that infants respond systematically to left or right shifts in the position of the FOE. More important, the data illustrate that some infants ignore even highly salient display differences, adopting instead a systematically biased pattern of responding, while others respond randomly a large fraction of the time. Whether individual differences in the degree of stimulus-driven, systematically biased, or random responding have developmental consequences is not yet known.
Supported by the National Science Foundation (BCS-0092452).