Abstract
Facilitation of contrast detection by collinear flankers (Polat & Sagi, 93) has been broadly researched and modeled. Here we report new experimental evidence that would provide further insights into this effect. The stimuli were either collinear Gabor target and flankers or in-phase D6 target and butterfly-shaped sinusoidal grating flankers. Both configurations produce significant facilitation at appropriate target-flanker separations in foveal vision. (1) When a random Gaussian noise was added to the target, facilitation diminished at a low noise level 2–3 times of the noise threshold. The threshold vs. noise functions under baseline and flanker conditions showed mainly a reduction of equivalent internal noise but no change of efficiency, suggesting some neural process, rather than uncertainty reduction, is responsible for collinear facilitation. (2) Collinear facilitation was evident at a stimulus duration as brief as 8 ms, a timing not possible for cortical feedback. (3) We examined the possibility that collinear facilitation is a pedestal effect as a result of receptive fields responding to both the target and flankers. By adding high contrast random noise between the target and flankers to disturb the responses of such receptive fields, we found no disruption of collinear facilitation. Collinear facilitation was still tuned to spatial frequency and thus was irrelevant to potential spatial cuing by the noise. These data question the pedestal-effect explanation of collinear facilitation and favor a longer-range spatial interaction theory. (4) Collinear flanker effects were reported previously to be suppressive in the periphery, which we believe was a result of inappropriate spatial scaling of the peripheral stimuli. On the basis of earlier peripheral scaling data (Yu & Essock, 96), we scaled the target-flanker separation and the length of the flankers separately. At 10 deg temporal periphery, collinear facilitation was evident at low flanker contrasts.