The color appearance of stimuli detected via short-wavelength-sensitive cones: comparisons with visual adaptation and visual field data for peri- or post-menopausal women under 70 years of age

Abstract

Dynamics of foveal light adaptation for vision mediated via short-wavelength-sensitive (SWS) cones were compared for two groups of healthy amenorrheic (peri- or post-menopausal) women not using hormonal medication. Each subject was assigned to a group based on the color name-"lavender" ( approximately 2/3 of all subjects) or "white" (approximately 1/3 of all subjects)-chosen in a forced-response paradigm to best describe a threshold-level 440-nm test presented on a larger 3.6 log td 580-nm background that had been viewed for approximately 5 min. During the first 20-30s after this 3.6 log td background abruptly replaced a much dimmer background, the threshold elevations (relative to the steady-state levels measured at approximately 5 min) were significantly greater for the lavender-naming subjects than for the white-naming subjects. However, exponential rates of recovery were indistinguishable for the two groups. A viable interpretation is that the gain of the visual response at background onset is greater for lavender-naming subjects than for white-naming subjects at or distal to a site where responses from middle-wavelength-sensitive and long-wavelength-sensitive (MWS and LWS) cones oppose responses from SWS cones. In addition, the color names derived from foveal testing were related systematically to extrafoveal sensitivities measured with Short Wavelength Automated Perimetry (SWAP), in a manner suggesting that response gain and/or response speed may be greater for lavender-naming subjects in the direction of increased SWS response also. Evidence from other subject populations suggests that the choice of color name and the dynamics of visual response each can be affected by alterations (particularly reductions) of estrogen synthesis and response.

Figure 1

Threshold elevations (log units) at 20 sec (interpolated values) after onset of the 3.6 log td, 580-nm adapting background field. Filled circles (●) represent subjects who chose “lavender” to best describe the threshold-level 440-nm test stimulus, unfilled circles (○) represent subjects who chose “white”, and the asterisk (*) represents the subject who chose “blue”. Threshold elevations are computed relative to the final asymptotic thresholds, measured ~5 minutes after background onset.

Figure 2

Threshold elevations (log units) as a function of time at 10-second interpolated intervals after onset of the 3.6 log td, 580-nm adapting background field. Filled circles (●) represent the median values from “lavender-naming” subjects, and unfilled circles (○) represent the median values from “white-naming” subjects. The horizontal dashed line corresponds to no threshold elevation. Connecting lines have no theoretical significance. The between-group differences were significant (p < .05) at 20 and 30 seconds. Other details as in Fig. 1

Figure 3

Figure 3 (left). Threshold elevation factor (linear units, converted from the data in Fig. 2) plotted as a function of time after onset of the 3.6 log td, 580-nm adapting background field. Horizontal dashed line (threshold elevation factor = 1) corresponds to no threshold elevation. The superimposed exponential curves are calculated based on the data from 30–100 sec, as described in the text and using the equation also provided in the legend to Fig. 3 (right). For reference, the interquartile range of threshold elevation factors at 20 sec spanned values of 5.9 to 17.9 for lavender-naming subjects; thus this range contains the value of the exponential curve at 20 sec. For white–naming subjects, the corresponding range spanned values of 3.4 to 5.7. Same symbols as Fig. 2. Figure 3 (right). Data from Fig. 3 (left) transformed so as to convert an exponential curve of the form y=1+Kexp(−t/t₀) to a straight line having a slope equal to the inverse negative time constant (−1/t₀) of the exponential curve. The transformation function was f(y) = ln(y-1) = lnK −t/t₀. Same symbols as Fig. 2. For lavender-naming subjects, t₀ =35.0 and K=10.1; for white-naming subjects, t₀ =35.7 and K=5.62.

Figure 4

Data represent differences (in dB units) between Full Threshold SWAP and SITA Standard white-on-white visual field “total deviations”. The abscissa represents each subject’s total deviation difference averaged across the 4 points in the innermost portion of the visual field, at 4.2° (ring 1, see Methods). The ordinate represents the each subject’s total deviation difference (SWAP – white-on-white) averaged across ring 2 minus the corresponding value for ring 3. (This difference of differences is the “eccentricity factor”, see Methods, also Eisner et al. (2006)). The ordinate is oriented so that larger reductions of SWAP sensitivity with eccentricity are represented in the downward direction. Same symbols as Fig. 1.