Abstract
Reverse correlation with sinusoidal noise is a method useful for estimating the information used during facial emotion classifications. As yet, this method has not been used to test for developmental differences in processing strategy during face classification. We had adult and 7-10 year old participants classify a neutral face (either the Mona Lisa or a neutral face from the NimStim set of faces) overlaid with random sinusoidal noise as either happy or sad. Over relatively few trials, classification images and subsequent quantitative analyses revealed that the information necessary to turn a neutral face into an expressive one differs for adults and children, but not for both stimuli. For the Mona Lisa face, adults and children performed similarly, relying on mid-high spatial frequencies for their classifications. For the real neutral face, adults relied more on lower spatial frequencies, whereas children relied on higher spatial frequencies; this pattern matches data from more traditional experiments. Super-subject data bolster these results, and demonstrates the consistency between 2AFC and 4AFC tasks. For both adults and children, critical regions for classification were not only present around the corners of the mouth, but also on the cheeks, between the eyes, and around the corners of the head. Overall, results indicate that children between 7-10 years employ immature processing strategies within facial emotion processing, and that only real faces are capable of revealing these differences. More generally, reverse correlation is a useful technique for estimating perceptual templates in developmental populations.
Meeting abstract presented at VSS 2013