Abstract
We showed that curve adaptation can propagate along the visual hierarchy to influence high-level facial expression judgment, and furthermore this effect is highly local (Xu et al., 2008). The question remains whether adapting to an incomplete curve generates the same aftereffect as adapting to a complete curve. In the current study, the adapting stimuli are a concave curve, and a set of 8 bisected concave curves with a variable central gap, displayed in separate conditions. The test stimuli are a set of curves varying from convex to concave. In each trial, observers viewed the adapting curve for 4 s, and after a 500 ms inter-stimulus interval viewed a test curve for 100 ms. Observers judged the curvature of the test curve (convex or concave) via a key press. A baseline condition without adaptation was also conducted. We measured the curvature aftereffects for all the adaptors as shifts of the psychometric curves from the baseline condition. We found that the curvature aftereffect produced by the complete-curve adaptor is the largest among all the adaptors (n = 5, p = .014), as expected. Interestingly, we found a significant curvature aftereffect (p = .038) when the gap of the incomplete curve adaptor is small (1/10 of the complete-curve length). The aftereffect remains significant as the gap increases to half the length of the complete curve (p = .013). When the gap increases further, the aftereffect gradually decreases to zero. This result suggests that we do not need the entire curve to generate curvature aftereffect. The observed aftereffect for incomplete-curve adaptation may have two possible explanations: 1) adaptation to the two ends of the incomplete curve, indicating the low-level root of visual adaptation; 2) perceptual filling-in for the missing part, indicating the top-down influence along the cortical hierarchy.
Meeting abstract presented at VSS 2012