Abstract
Recent research (Kahn, Mattar, & Aguirre, SfN 2012) has suggested that the gamut of a stimulus set affects the form of neural encoding, as measured by neural adaptation or norm-based effects. In addition, it has been long argued that the number of exemplars (divisions) along a single dimension plays a role in perception, regardless of gamut (Miller, 1956; Shiffrin & Nosofsky, 1994). We sought to investigate the influence of both gamut and division on neural encoding effects related to face similarity. 8 pilot subjects viewed subsets of a linear face morph space presented using a carry-over design (Aguirre, 2007) while event-related potentials were measured. In three different experiments (with counter-balanced order across subjects) the stimuli used either densely sampled the whole space (9-stimuli, full gamut), densely sampled part of the space (5-stimuli, narrow gamut), or sparsely sampled the whole space (5-stimuli, full gamut). Our preliminary findings replicate prior work (Kahn, Harris, Wolk, & Aguirre, 2010), demonstrating a linear recovery from adaptation within the P200 component of the evoked response proportional to the metric similarity of a 5-step, full-gamut morph of two face identities. We expected that as the number of morphed steps increased between the two face identities, that the degree of adaptation resulting from small stimulus transitions should demonstrate a floor effect. Our measure of adaptation within the P200 appears to exhibit this floor effect when the number of face morph exemplars reaches 9 across the full morph gamut. Interestingly, the pattern of adaptive effects resulting from the 5 stimulus narrow-gamut space does not readily correspond to the pattern of either of the other two, suggesting a more complex model of early perceptual encoding may be necessary.
Meeting abstract presented at VSS 2013