Color Discrimination Task Using Pseudoisochromatic Stimulus: Luminance Noise Variation Provides Better Sensitivity Than Noise Mean Luminance

Abstract

Purpose: : To study the role of amplitude and mean luminance noise for color discrimination task using pseudoisochromatic stimulus

Methods: : Ten female subjects (22.7 ± 3 years old) were monocularly tested using the Cambridge Color. All subjects were normal trichomats and had 20/20 corrected visual acuity. A pseudoisochromatic stimulus composed by C-Landolt surrounded by a background that differed only in chromaticity. The chromaticity of the background in the CIE1976 color space was u’= 0.197, v’= 0.468. The chromaticity of the target varied along eight chromatic vectors. The subjects had to find the direction of the C-Landolt gap while the chromatic vector was decreased. The final result of the subject’s test was the minimum length for the eight chromatic vectors. The color discrimination task was performed against luminance noise at seven noise mean luminance (7, 10, 13, 16, 19, 22, and 25 cd/m²). The luminance noise was composed by patches with luminance ranging between 7 cd/m² above and below stimulus mean luminance (constant delta protocol) or by luminance ranging between 70% above and below stimulus mean luminance (constant proportion protocol). Power functions fitted the data and the model exponent was used to quantify the vector length decrement rate as a function of noise mean luminance. Two-way ANOVA was used to compare subject performance for both luminance noise conditions.

Results: : The two protocols provided different results. In the constant delta protocol, vector length was high (poor color discrimination) when the mean luminance of luminance noise was low and it decreased when the mean luminance of the luminance noise was raised. In the constant proportion protocol, vector length varied around a mean value at all tested noise mean luminance. There were statistical differences for the power function exponent estimated from both protocols at four chromatic vectors, with higher exponents for constant delta protocol (p < 0.05)

Conclusions: : The vector length was dependent of the amplitude of luminance proportion above or below the mean luminance of the luminance noise. The performance of the subjects was the same when the similar luminance proportions were used nevertheless the mean luminance of luminance noise. Color discrimination performances using pseudoisochromatic stimulus obeyed the Weber’s law

Keywords: perception • color vision • color appearance/constancy