Comparing the Shape of the Contrast Sensitivity Functions of Normal and Low Vision

Abstract

Purpose: The contrast sensitivity function (CSF) provides a comprehensive description of an individual’s spatial-pattern detection capability. Because most low vision patients suffer from impaired acuity and contrast sensitivity, their CSFs will differ from those of people with normal vision. The goal of this study was to test the hypothesis that the CSFs of people with low vision differ from a “normal” CSF only in their horizontal and vertical positions along the spatial frequency (SF) and the contrast sensitivity (CS) axes.

Methods: Contrast sensitivity for detecting the presence of a horizontal sinewave grating was measured with a two temporal-interval forced-choice staircase procedure, for a range of SFs spanning 5-6 octaves. CSFs were measured for 23 low-vision eyes (11 with AMD, 5 with Stargardt’s disease and 7 with other pathologies). CSFs were also measured for five adults with normal vision, and the aggregate data were fit with an asymmetric parabolic function. This “normal template”, with the width parameters (shape) of the function constrained, was shifted horizontally and vertically along the SF and CS axes (both in log coordinates) to find the best fit for each of the 23 low-vision CSFs. The low-vision CSFs were also directly fit with best-fitting asymmetric parabolic functions (“free-fit”).

Results: A comparison of the peak CS, the SF at which peak CS occurs (SF_peak) and the high-frequency cut-off derived from the two fitting methods (template vs. free-fit) reveals that the values are highly correlated (r=0.77 to 0.98) and in good agreement (Bland-Altman analysis) with one another, suggesting that the template fit is comparable with the free-fit method in estimating these parameters. The width of the left-half of the low-vision CSFs was 1.72±0.64× [mean±SD] that of the normal’s (p<0.0001) while the width of the right-half of the low-vision CSFs was 0.90±0.15× that of the normal’s (p=0.005), implying that the low-vision CSFs are wider on the left and slightly narrower on the right than a normal CSF.

Conclusions: Although the low-vision CSFs are wider than that of a normal CSF, a normal CSF template predicts the peak CS, SF_peak and the high-frequency cut-off of low-vision CSFs reasonably well. Our results suggest that a normal template provides an approximate description of the spatial-pattern detection capability of low vision patients, especially if the emphasis is on the high-SF half of the function.