Michael A. Webster

Adaptation to natural facial categories

Face perception is fundamentally important for judging the characteristics of individuals, such as identification of their gender, age, ethnicity or expression. We asked how the perception of these characteristics is influenced by the set of faces that observers are exposed to. Previous studies have shown that the appearance of a face can be biased strongly after viewing an altered image of the face, and have suggested that these after-effects reflect response changes in the neural mechanisms underlying object or face perception. Here we show that these adaptation effects are pronounced for natural variations in faces and for natural categorical judgements about faces. This suggests that adaptation may routinely influence face perception in normal viewing, and could have an important role in calibrating properties of face perception according to the subset of faces populating an individuals environment.

Figural aftereffects in the perception of faces

We examined figural aftereffects in images of human faces, for which changes in configuration are highly discriminable. Observers either matched or rated faces before or after viewing distorted images of faces. Prior adaptation strongly biases face perception by causing the original face to appear distorted in a direction opposite to the adapting distortion. Aftereffects transferred across different faces and were similar for upright or inverted faces, but were weaker when the adapting and test faces had different orientations (e.g., adapt inverted and test upright). Thus the aftereffects depend on which images are distorted, and not simply on the type of distortion introduced. We further show that the aftereffects are asymmetric, for adapting to the original face has little effect on the perception of a distorted face. This asymmetry suggests that adaptation may play an important normalizing role in face perception. Our results suggest that in normal viewing, figural aftereffects may strongly influence form perception and could provide a novel method for probing properties of human face perception.

Visual adaptation and face perception

The appearance of faces can be strongly affected by the characteristics of faces viewed previously. These perceptual after-effects reflect processes of sensory adaptation that are found throughout the visual system, but which have been considered only relatively recently in the context of higher level perceptual judgements. In this review, we explore the consequences of adaptation for human face perception, and the implications of adaptation for understanding the neural-coding schemes underlying the visual representation of faces. The properties of face after-effects suggest that they, in part, reflect response changes at high and possibly face-specific levels of visual processing. Yet, the form of the after-effects and the norm-based codes that they point to show many parallels with the adaptations and functional organization that are thought to underlie the encoding of perceptual attributes like colour. The nature and basis for human colour vision have been studied extensively, and we draw on ideas and principles that have been developed to account for norms and normalization in colour vision to consider potential similarities and differences in the representation and adaptation of faces.

The influence of contrast adaptation on color appearance

Most models of color vision assume that signals from the three classes of cone receptor are recoded into only three independent post-receptoral channels: one that encodes luminance and two that encode color. Stimuli that are equated for their effects on two of the channels should be discriminable only to the remaining channel, and are thus assumed to isolate the responses of single channels. We used an asymmetric matching task to examine whether such models can account for changes in color appearance following adaptation to contrast--to temporal variations in luminance and chromaticity around a fixed mean luminance and chromaticity. The experiments extend to suprathreshold color appearance the threshold adaptation paradigm of Krauskopf, Williams and Heeley [(1982) Vision Research, 32, 1123-1131]. Adaptation changes the perceived color of chromatic test stimuli both by reducing their saturation (contrast) and by changing their hue (direction within the equiluminant plane). The saturation losses are largest for test stimuli that lie along the chromatic axis defining the adapting modulation, while the hue changes are rotations away from the adapting direction and toward an orthogonal direction within the S and L-M plane. Similar selective changes in both perceived color and perceived lightness occur following adaptation to stimuli that covary in luminance and chromaticity. The selectivity of the aftereffects for multiple directions within color-luminance space is inconsistent with sensitivity changes in only three independent channels. These aftereffects suggest instead that color appearance depends on channels that can be selectively tuned to any color-luminance direction, and that there are no directions that invariably isolate responses in only a single channel. We use the perceived color changes to examine the spectral sensitivities of the chromatic channels and to estimate the distribution of channels. We also examine how adaptation alters the contrast-response function, how it affects reaction times for luminance and chromatic contrast, the extent to which the aftereffects exhibit interocular transfer, and the way in which the perceived color changes differ from those induced by conventional light adaptation.

Adaptation and visual coding

Visual coding is a highly dynamic process and continuously adapting to the current viewing context. The perceptual changes that result from adaptation to recently viewed stimuli remain a powerful and popular tool for analyzing sensory mechanisms and plasticity. Over the last decade, the footprints of this adaptation have been tracked to both higher and lower levels of the visual pathway and over a wider range of timescales, revealing that visual processing is much more adaptable than previously thought. This work has also revealed that the pattern of aftereffects is similar across many stimulus dimensions, pointing to common coding principles in which adaptation plays a central role. However, why visual coding adapts has yet to be fully answered.

Adaptation and the color statistics of natural images.

Color perception depends profoundly on adaptation processes that adjust sensitivity in response to the prevailing pattern of stimulation. We examined how color sensitivity and appearance might be influenced by adaptation to the color distributions characteristic of natural images. Color distributions were measured for natural scenes by sampling an array of locations within each scene with a spectroradiometer, or by recording each scene with a digital camera successively through 31 interference filters. The images were used to reconstruct the L, M and S cone excitation at each spatial location, and the contrasts along three post-receptoral axes [L + M, L - M or S - (L + M)]. Individual scenes varied substantially in their mean chromaticity and luminance, in the principal color-luminance axes of their distributions, and in the range of contrasts in their distributions. Chromatic contrasts were biased along a relatively narrow range of bluish to yellowish-green angles, lying roughly between the S - (L + M) axis (which was more characteristic of scenes with lush vegetation and little sky) and a unique blue-yellow axis (which was more typical of arid scenes). For many scenes L - M and S - (L + M) signals were highly correlated, with weaker correlations between luminance and chromaticity. We use a two-stage model (von Kries scaling followed by decorrelation) to show how the appearance of colors may be altered by light adaptation to the mean of the distributions and by contrast adaptation to the contrast range and principal axes of the distributions; and we show that such adjustments are qualitatively consistent with empirical measurements of asymmetric color matches obtained after adaptation to successive random samples drawn from natural distributions of chromaticities and lightnesses. Such adaptation effects define the natural range of operating states of the visual system.

Orientation and spatial-frequency discrimination for luminance and chromatic gratings

We have examined the accuracy of orientation and spatial-frequency discrimination for sine-wave gratings that vary in either luminance or color. The equiluminant chromatic gratings were modulated along either a tritanopic confusion axis (so that they were detectable on the basis of activity in only the short-wavelength-sensitive cones) or an axis of constant short-wavelength-sensitive cone excitation (so that they could be detected on the basis of opposing activity in only the long- and medium-wavelength-sensitive cones). Grating contrasts ranged from the detection threshold to the highest levels that we could produce; the contrasts of the luminance and color patterns were equated for equal multiples of their respective detection thresholds. Discrimination thresholds for all patterns showed a similar dependence on stimulus contrast, rising sharply at low contrasts and becoming nearly asymptotic at moderate contrasts. However, even at threshold contrasts, observers could still reliably discriminate sufficiently large differences in the orientation or spatial frequency of all patterns, and they could also reliably identify the type of variation (luminance or which color) defining the grafting. For most conditions the discrimination thresholds did not differ from the two types of color grafting and reached values as low as 1 deg (orientation) or 4% (spatial frequency). Thus observers were able to make accurate spatial judgments on the basis of either type of chromatic information. However, these thresholds were slightly but consistently higher than the thresholds for comparable luminance graftings. This difference in the color and luminance discrimination thresholds may reflect somewhat coarser orientation and spatial-frequency selectivity in the mechanisms encoding the chromatic patterns.

Variations in normal color vision. II. Unique hues

We examined individual differences in the color appearance of nonspectral lights and asked how they might be related to individual differences in sensitivity to chromatic stimuli. Observers set unique hues for moderately saturated equiluminant stimuli by varying their hue angle within a plane defined by the LvsM and SvsLM cone-opponent axes that are thought to characterize early postreceptoral color coding. Unique red settings were close to the +L pole of the LvsM axis, while green, blue, and yellow settings clustered along directions intermediate to the LvsM and SvsLM axes and thus corresponded to particular ratios of LvsM to SvsLM activity. Interobserver differences in the unique hues were substantial. However, no relationship was found between hue settings and relative sensitivity to the LvsM and SvsLM axes. Moreover, interobserver variations in different unique hues were uncorrelated and were thus inconsistent with a common underlying factor such as relative sensitivity or changes in the spectral sensitivities of the cones. Thus for the moderately saturated lights we tested, the unique hues appear largely unconstrained by normal individual differences in the cone-opponent axes. In turn, this suggests that the perceived hue for these stimuli does not depend on fixed (common) physiological weightings of the cone-opponent axes or on fixed (common) color signals in the environment.

Temporal properties of brightness and color induction.

With a matching procedure, we studied the temporal properties of direct brightness (or lightness) and chromatic changes (produced by modulation of the region being matched) and induced brightness and chromatic changes (produced by modulation of the surround of the region being matched). The amount of direct brightness and color change was found to vary only slightly with temporal frequency over the 0.5-8 Hz range studied, whereas induced changes were found to occur only at low temporal frequencies, below about 2.5 Hz. With high temporal-frequency modulation of the surround, the induced patterns appeared to flicker but not to change in brightness or color. Despite the fact that chrominance and luminance temporal contrast sensitivity functions are very different, the temporal induction curves for color and brightness were very similar. However, brightness induction was found to increase approximately linearly with increasing surround modulation up to very high levels, whereas the amount of color induction was much less dependent on the modulation depth of the surround.

Stimulus selectivity of figural aftereffects for faces

Viewing a distorted face induces large aftereffects in the appearance of an undistorted face. The authors examined the processes underlying this adaptation by comparing how selective the aftereffects are for different dimensions of the images including size, spatial frequency content, contrast, and color. Face aftereffects had weaker selectivity for changes in the size, contrast, or color of the images and stronger selectivity for changes in contrast polarity or spatial frequency. This pattern could arise if the adaptation is contingent on the perceived similarity of the stimuli as faces. Consistent with this, changing contrast polarity or spatial frequency had larger effects on the perceived identity of a face, and aftereffects were also selective for different individual faces. These results suggest that part of the sensitivity changes underlying the adaptation may arise at visual levels closely associated with the representation of faces.