Design and validation of a chart-based measure of the limits of spatial contrast sensitivity.

Abstract

Listen

Translate article

Current chart-based tests of spatial contrast sensitivity (SCS) with fixed or narrow frequency ranges (≤18cycles/°) cannot characterise the limits of spatial contrast vision. Here we present the design and validation of a chart-based measure of the spatial contrast envelope. Following the principles of the standard visual acuity (Bailey-Lovie) and contrast sensitivity (Pelli-Robson) charts, a combined spatial-contrast and visual acuity chart was designed using a language-independent triangular symbol for a four-alternative forced-choice procedure plus chart rotation. Symbol frequencies ranged between 0.38 and 60cycles/° spaced along 10radial axes (0.55%-100% contrast). The chart was validated with reference to the Bailey-Lovie and Pelli-Robson charts; its reliability and sensitivity to changes in illumination, simulated cataract and blur was evaluated in healthy adults. The photopic SCS function could be measured in 5.5±0.5min; thresholding around the spatial contrast resolution limit reduced completion times to ~2min. There was good agreement with high-contrast visual acuity (difference=0.08±0.02logMAR) and contrast-sensitivity at 1.5 cycles/° (0.13±0.06logCS). Test-retest reliability was excellent at all spatial frequencies (ICC=0.99). Mesopic illumination or simulated cataract caused a generalised SCS loss; myopic blur reduced high-frequency sensitivity. Spatial contrast sensitivity was independent of radial axis orientation (cardinal or oblique). The chart provides a time-efficient, reliable and inexpensive measure of SCS with applications in research and clinic for detecting subtle deficits in early stages of ocular and neurological conditions that often manifest at higher frequencies. It is sensitive to vision changes occurring in dim lighting and with simulated cataract and blur. The chart is available open-access for self-printing; contrast variation in print can be controlled through user calibration and/or establishing normative SCS functions using the theoretical values.