Abstract
One hallmark of perceptual expertise is holistic processing, wherein objects are processed as wholes rather than by individual features. It has been suggested that holistic face processing could arise from a strategy of attending to all face parts (Curby et al., 2012, Richler et al., 2011). In a training study with two novel-race faces, we reported that holistic face processing depends on experience attending to diagnostic face parts, and is found even for combinations of face parts never seen together (Chua et al., VSS2013; under revision). We extend this paradigm to two categories of novel objects (Greebles) that should not be processed holistically before training. Fifty-one subjects learned to individuate objects from two visually distinct Greeble categories, where diagnostic information was in complementary halves (e.g., if diagnostic information was in the top half for one Greeble category, then it would be in the bottom for the other category). We then measured holistic processing in the composite task with new Greebles made up of aligned or misaligned diagnostic or non-diagnostic parts. A control group (n=38) with no prior exposure to Greebles also performed the composite task. In a within-subjects analysis of the composite task for trained subjects, we found no difference in holistic processing elicited by composites made of diagnostic vs. non-diagnostic Greeble parts (diagnosticity x alignment x congruency interaction in d'), F(1,50) = 1.33, p = 0.26, ηp2=.03. However, Greeble training led to increased holistic processing relative to the control group, F(1,87) = 7.45, p <0.01, ηp2=.08, showing that individuation training increased holistic processing of non-face objects. One explanation for the difference between these results and those we obtained with novel-race faces is that the two Greeble categories may not have been sufficiently distinctive, being distinguished by the shape of the non-diagnostic contour rather than the diagnostic parts themselves.
Meeting abstract presented at VSS 2014