THEORY AND MEASUREMENT OF VISUAL MECHANISMS

Abstract

Comparison of monocular and binocular critical flash intensities for recognition of flicker, using a centrally fixated square image subtending ca. 6.13° on a side (white light), shows that for the cone segment of the response contour the inflection point of the probability integral correlating flash frequency F (for symmetrical flicker) and log mean critical flash intensity Im is with the binocular measurements exactly intermediate between those for each eye separately. This does not mean that in general the data are intermediate; they are not; the binocular asymptotic Fmax. agrees with or lies above the greater one of the two uniocular curves. The entire contour must be considered for valid intercomparisons, as is also true if homologous curves for different observers are to be compared. For the measurements in the predominantly rod region the binocular data are more or less intermediate. The rod curves result, however, from the integrative interplay of rod and cone effects for which the intrinsic curves overlap. The resultant rod curve as measured is determined by the partial inhibition of rod effects by cone effects, and by the summation of the remaining rod contributions with those labelled cone in origin. It is pointed out that in this respect, as in others, it is desirable to consider the rôles of retinal area, and location, from the standpoint of integration of neural effects. These phenomena are essentially independent of the light-time fraction and of the spectral (λ) quality of the light used. For binocular, as for uniocular excitation, the normal probability summation provides an efficient general description, under diverse conditions of size and location of retinal image, wave-length composition of light, light-time cycle-fraction, and kind of animal. It is pointed out that this is the only function abstractly likely to exhibit this kind of efficiency. That a summation of veritable effects independently generated by simultaneous, symmetrical uniocular excitation does occur in the recognition of flicker is specifically demonstrated by the fact that for a given mean critical flash intensity the associated variation is lower for binocular than for either or the average of the single-eyed presentations,—and in the ratio not statistically different from 1:1.41; the relative scatter of the binocular indices of dispersion is also reduced below the uniocular. Since the mean variation of the critical intensity is statistically in a constant ratio to Im, in appropriately homogeneous series, independent for example of the brightness level and of the light-time fraction, this signifies an essential doubling of the effectiveness (potential) of each of the elements concerned in the discrimination of flicker when binocular excitation is concerned, although the total number of these elements is only slightly or not at all affected. The potential in question is not exclusively correlated with subjective brightness-at-fusion, which is, however, increased with binocular regard.