Relating Chromatic Antagonism in Pi-5 to Red/Green Hue Cancellation.

Abstract

Recent theories of two-color increment threshold detection have sought to explain W. S. Stiles' (pi)-mechanisms in terms of pathways with multiple gain sites. In these pathway models, the first gain site is presumed to be the (single) cone class in which the signal orginates. In pathways with a second gain site, other classes of cones may influence the threshold of a signal they do not detect. Failures of field additivity have provided strong evidence for such a second gain site. In the case of a detection signal originating in the long-wavelength sensitive ((gamma)) cones, addition of short- and medium-wavelength light to a long-wavelength background can actually lower threshold. Such field antagonism needs to be connected with supra-threshold color opponency. Following a successful model of (pi)(,1) detection through the yellow/blue opponent pathway, it was proposed that the field antagonism evident in detection of (gamma) cone signals is due to their use of the opponent pathway also responsible for red/green hue cancellation. Furthermore, since second site gain is known to decrease for fields strongly red polarized the "equilibrium hypothesis" was advanced. This asserted that second site gain would be maximized by fields in perceptual red/green equilibrium. Testing the equilibrium hypothesis required decomposing observed gain (threshold) into a first and second site component. This was accomplished through the use of two test stimuli. A 10 msec, 1 degree disk was used to measure first site adaptation, and a "low frequency" test was used to measure adaptation in the red/green pathway. This low frequency test had a temporal profile which was one trough-to-trough period of a 2 Hz. sine wave and a spatial profile which was a 3 degree disk with blurred edges. Thresholds for these two stimuli were measured on mixture backgrounds which were slightly red, slightly green, or in red/green equilibrium. These backgrounds were always composed of a fixed 650 nm part (with intensity 10.5 log quanta/(sec)(sqr.deg)) and a greenish component of wavelength 495, 510, 525, or 540 nm and with intensity chosen to satisfy the above perceptual criterion. The results contradicted the equilibrium hypothesis. Red/green neutral backgrounds produced more second site gain than backgrounds which were slightly red, but did not maximize second site gain. Backgrounds slightly greener than equilibrium produced more second site gain than those precisely in equilibrium. In fact, other tests concluded that there is no steady-state desensitization due to polarization of the red/green pathway in the green direction. Although the equilibrium hypothesis was not upheld, greens with the same cancellation green content did have about the same effect at the second site. Therefore there is reason to believe that the field antagonism evident in threshold signals originating in the (gamma) cones shares part of the physiological substrate responsible for red/green hue cancellation.