Purchase this article with an account.
Nichola Burton, Linda Jeffery, Andrew Calder, Gillian Rhodes; Adaptation to an average expression improves discrimination of facial expressions. Journal of Vision 2014;14(10):815. https://doi.org/10.1167/14.10.815.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Adaptation serves an important calibrating function in low-level vision, allowing efficient coding of a wide range of stimulus levels and enhancing discrimination around the adapted level. Adaptation is also present in high-level face perception processes such as expression perception, but the functional benefit of this adaptation has not been consistently demonstrated. Here we show that adaptation to an average expression improves discrimination around that average. We created a morphed expression trajectory that ran from fear, through an average expression, to anti-fear. Twenty-three participants were trained to recognize the two endpoint expressions and the average expression using arbitrary labels. They then categorized a range of briefly-presented expressions taken from this trajectory in three conditions: with no adaptation (baseline), after 160 s of adaptation to the average expression, and after 160 s of adaptation to the alternating endpoint expressions. A size change was included to minimize the effect of low-level adaptation. Thresholds were calculated for the detection of each endpoint expression; the distance between these thresholds on the trajectory served as a measure of expression discrimination. Following adaptation to the average expression, the inter-threshold distance narrowed significantly relative to baseline, indicating better discrimination around the average in this condition. In contrast, adaptation in the alternating condition did not change the inter-threshold distance relative to baseline, suggesting that our finding cannot be explained by a priming effect. Overall, our results demonstrate a functional role for adaptation in high-level expression perception.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only