Abstract
A patterned annulus that fails to activate a cell on its own may reduce activity when the cell is excited by a center patterned disc. Most models assume lateral interactions among low-level mechanisms (pooled gain control) are responsible, and psychophysical models further assume that the interactions are reciprocal between center and surround. To complicate these simple models, there is some physiological evidence that the apparent size of a cell's excitatory zone is not fixed; it appears to decrease with increasing contrast. Here, we a) tested the reciprocity assumption for fine spatial discriminations and b) psychophysically measured summation areas at three contrast levels. In both experiments, observers discriminated between two highly similar grating patches (4 cpd) in each block of eighty trials. Half discriminated on the basis of orientation; half on spatial frequency. Each condition of each experiment was run in a separate session, using a signal detection rating paradigm, and was replicated five times following extensive practice and individual determination of “threshold” (d'=1.5). In the reciprocity experiments, observers made decisions either on a center circular patch of grating with and without an annular masking surround or on an annular surround ring with and without a center masking disk. Four test/mask size ratios and six contrasts were used. In the summation experiment, discrimination performance was measured for center discs alone as a function of size (5 levels) and contrast (3 levels). We confirmed that summation areas decreased with increasing contrast when measured psychophysically. Furthermore, interactions were not reciprocal. Surrounds masked centers, but centers had no effect on surrounds. These results imply that current psychophysical models of lateral interactions are too simple, and that mechanisms of figure-ground organization may play in role in mediating these lateral interactions.
This research was supported by NIH Grant EY13953