Abstract
Continuous motion tracking has been recently proposed to estimate visual sensitivity. Bonnent et al (2015) has shown that the new method was more efficient and easier to conduct in estimating position sensitivities, compared to traditional psychophysical methods. But it remains unclear whether the method can be also adopted to measure contrast sensitivity. Here, we applied continuous motion tracking to measure chromatic contrast sensitivity of both red-green and blue-yellow color (L-M, and S axes in DKL color space) in both color-deficient observers (red-green “color-blind”, N = 8) and normal observers (N = 10). The observer moved a mouse cursor to track an isoluminant color target as it moved in random walk among a field of dynamic luminance noise. The dynamic luminance noise ensured that the color target cannot be segemented by residual luminance artifacts. Each tracking session lasted 60 seconds, and there were 2 sessions for each contrast (7 levels) in each color (2 colors: red-green, blue-yellow). Tracking performance was fitted by a Kalman filter model, using Gibbs sampling to estimate the posterior distributions of perceptual noise. We found that tracking performance for both colors increased as contrast increased, which can be well described by a hyperbolic ratio function (i.e., Naka-Rushton function). The contrast response estimation is highly reliable across the two testing sessions (each 60s) with a correlation of r = 0.995. Furthermore, at lower contrasts of red-green color (15% and 28%), we can use the tracking performance to separate color-deficiencies from normal controls at 100% accuracy (AUC = 1 with ROC analysis). By analyzing data with an even shorter duration, we found that mere 30-second tracking is sufficient to achieve the acruracy of 100% in classification. Taken together, the continuous motion-tracking task can be used to measure contrast sensitivity rapidly, and to identify color deficiency reliably.