Abstract
Purpose. We showed that collinear flankers facilitate contrast discrimination at low contrasts but impede it at high (Chen & Tyler, 2001, Proc. Roy. Soc. B.). Physiological evidence suggests that the long range interaction is orientation and location specific. Thus, noncollinear flankers may have qualitatively different effects from collinear ones.
Methods. We measured contrast thresholds for vertical Gabor targets superimposed on a pedestal in the presence of flankers. All patterns were 4 c/deg Gabor patches. The pedestal orientation was vertical and the flankers were either vertical or horizontal, placed 1o from the target. “Aligned” flankers were placed in a vertical line, “side” flankers in a horizontal line through the target and “cross” flankers were the sum of the line and the side flankers.
Results. Without flankers, the target threshold vs. pedestal contrast (TvC) function showed a dipper shape: facilitation at low and masking at high contrast. The vertical line flankers increased both facilitation and masking effect. Compared with the vertical aligned flankers, the horizontal aligned flankers had less effect at low contrast, but had similar effect at high contrasts. The vertical side flankers had less effect at both low and high contrasts. The cross flankers had the same effect as the vertical aligned flankers, suggesting the flanker effects did not sum linearly across space.
Conclusions. We fit a sensitivity modulation model to the data. This model suggests that the target detector received multiplication inputs from other mechanisms in different hypercolumns that modulate its sensitivity to both excitatory and the divisive inhibitory elements of the response function. The best fit parameters reveal that there are two different long-range interactions: one is orientation and location specific and the other is broadly tuned.