Linking appearance to neural activity through the study of the perception of lightness in naturalistic contexts

The present paper deals with the classical question how a psychological experience, in this case apparent lightness, is linked by intervening neural processing to physical variables. We address two methodological issues: (a) how does one know the appropriate physical variable (what is the right x?) to look at, and (b) how can behavioral measurements be used to probe the internal transformation that leads to psychological experience. We measured so-called lightness transfer functions (LTFs), that is the functions that describe the mapping between retinal luminance and perceived lightness for naturalistic checkerboard stimuli. The LTFs were measured for different illumination situations: plain view, a cast shadow, and an intervening transparent medium. Observers adjusted the luminance of a comparison patch such that it had the same lightness as each of the test patches. When the data were plotted in luminance–luminance space, we found qualitative differences between mapping functions in different contexts. These differences were greatly diminished when the data were plotted in terms of contrast. On contrast–contrast coordinates, the data were compatible with a single linear generative model. This result is an indication that, for the naturalistic scenes used here, lightness perception depends mostly on local contrast. We further discuss that, in addition to the mean adjustments, one may find it useful to consider also the variability of an observer’s adjustments in order to infer the true luminance-to-lightness mapping function.

Perceptual Organization and White's Illusion

The apparent lightness of a surface can be strongly modulated by the spatial context in which it is embedded. Early theories of such context dependence emphasized the role of low-level mechanisms that sense border contrast, whereas a number of recent authors have emphasized the role of perceptual organization in determining perceived lightness. One of the simplest and most theoretically challenging lightness illusions was described by White. This illusion has been explained with a variety of different models, ranging from low-level filter outputs to computations underlying the extraction of mid-level representations of surfaces. Here, I present a new method for determining the organizational forces that shape this illusion. I show that the spatial context of White's pattern not only transforms the apparent lightness of homogeneous target patches, but can also induce dramatic inversions of figure–ground relationships of textured target regions. These phenomena provide new evidence for the role of scission in causing the lightness illusion experienced in White's effect.