home
| Color Appearance and Chromatic Induction |
The appearance of two physically identical lights can appear quite different. The figure below shows two sets of backgrounds. The top two backgrounds are composed of circles alternating between 'lime' and 'purple' colors while the bottom two backgrounds are composed of uniform fields at either component chromaticity. The four central test rings in each background are physically identical but appear different due to context. The color shift observed with patterned backgrounds (top pair) is much larger than observed with the unifrom backgrounds (bottom pair). |
 |
These shifts in color were investigated in a series of papers with Steven Shevell. First, we demonstrated that the conspicuous shifts observed with patterned backgrounds can not be accounted for by optical factors alone (Monnier & Shevell, 2003). Instead, we proposed a type of neuron with a center-surrrond antagonistic receptive-field organization as a neural substrate. Quantitative implementation of the receptive field idea accounted for quite a few measurements (Monnier & Shevell, 2004). Most recently, we tested the receptive field model further by measuring color apperance as a function of the chromatic contrast of the background. The figure below shows a series of patterns with decreasing chromatic contrast. Measurements show that the magnitude of the color shifts is linearly related to the chromatic contrast within the inducing pattern (Shevell & Monnier, 2005). |
 |
In the same paper, we tested another prediction made by the saptial structure of the receptive field model: spatial frequency selectivity, which predicts the magnitude of the color shift should peak at an intermediate, moderate pattern-spatial frequency. This prediction results from the bandpass spatial modulation transfer function of the receptive field. The figure below shows some of the stimuli used to test the spatial frequency prediction. |
 |
|