Abstract
How do infants segregate objects from patterned backgrounds? We showed that 5.5-month-old infants can segregate a novel object from a patterned background when the background changes during habituation but not when it remains constant (Salvagio, et al. VSS 2014). There, we measured infants’ relative looking time (LT) at two novel objects on an unpatterned background during a single pre-test trial. Then, infants habituated to one of the novel objects, accompanied by a verbal phrase, on either a changing or constant background. At post-test, infants viewed the pre-test display while hearing the habituation phrase. If infants segmented the object during habituation, their post-test LTs at the habituation object should be longer than at the other object. Infants in the changing background group showed evidence of segmentation (LT at habituation object pre- and post-test = 0.44 & 0.55), p < 0.05; infants in the constant background group did not (LT pre- and post-test = 0.55 & 0.50), p > 0.46. This year we asked whether a changing background was sufficient for segmentation or whether experience with the pre-test display established a memory for a segmented object that operated when infants viewed the habituation displays. Perhaps, in the changing background group stronger weights were assigned to this interpretation with each presentation, whereas, weights for individual unified representations remained lower. In the constant background group a single unified representation garnered stronger weights over time. Here, we tested the changing background condition without pre-test. No evidence of segmentation was observed (post-test LT = 0.51), p > 0.46. These results may support the memory interpretation. Alternatively, the novelty of the test displays may have caused exploration behavior that overpowered preference for the habituation object at post-test. This second interpretation raises an important methodological concern for many infant studies. An experiment to distinguish these possibilities is underway.
Meeting abstract presented at VSS 2015