Contrast Negation and the Importance of the Eye Region for Holistic
Representations of Facial Identity
Mladen Sormaz, Timothy J. Andrews, and Andrew W. Young
University of York
Reversing the luminance values of a face (contrast negation) is known to disrupt recognition. However,
the effects of contrast negation are attenuated in chimeric images, in which the eye region is returned to
positive contrast (S. Gilad, M. Meng, & P. Sinha, 2009, Role of ordinal contrast relationships in face
encoding, Proceedings of the National Academy of Sciences, USA, Vol. 106, pp. 5353–5358). Here, we
probe further the importance of the eye region for the representation of facial identity. In the first
experiment, we asked to what extent the chimeric benefit is specific to the eye region. Our results showed
a benefit for including a positive eye region in a contrast negated face, whereas chimeric faces in which
only the forehead, nose, or mouth regions were returned to positive contrast did not significantly improve
recognition. In Experiment 2, we confirmed that the presence of positive contrast eyes alone does not
account for the improved recognition of chimeric face images. Rather, it is the integration of information
from the positive contrast eye region and the surrounding negative contrast face that is essential for the
chimeric benefit. In Experiment 3, we demonstrated that the chimeric benefit is dependent on a holistic
representation of the face. Finally, in Experiment 4, we showed that the positive contrast eye region needs
to match the identity of the contrast negated part of the image for the chimera benefit to occur. Together,
these results show the importance of the eye region for holistic representations of facial identity.
Keywords: face recognition, eyes, photo negation, contrast chimera, contrast
Reversing the luminance values of an image (contrast negation)
disrupts the recognition of familiar faces (Galper, 1970; Galper &
Hochberg, 1971; Phillips, 1972; Johnston, Hill, & Carmen, 1992).
The effect of contrast negation provides an interesting counterex-
ample to the typical resistance of familiar face recognition to other
types of image degradation (Bruce & Young, 2012), and contrast
negation appears to disrupt face recognition to a greater degree
than recognition of other types of visual stimuli (Vuong, Peissig,
Harrison, & Tarr, 2005; Nederhouser, Yue, Mangini, & Bieder-
man, 2007). The effect of contrast negation is easily seen (by those
old enough to remember chemical photography) in the difficulty of
recognizing faces from photographic negatives. However, the neg-
ative of a normal color image reverses both its luminance and its
hue. A seminal study by Kemp, Pike, White, and Musselman
(1996) demonstrated that face images with negated luminance but
normal hue saw a reduction in recognition accuracy, whereas faces
with negated hue and normal luminance showed no recognition
accuracy decrease. It is therefore clear that information critical to
face recognition is somehow carried through luminance values.
The explanation for why contrast reversal has such a profound
effect on face recognition has been dominated by two different
ideas. The first is Kemp et al.’s (1996) suggestion that the diffi-
culty in recognizing contrast negated faces is due to the reversal of
three-dimensional (3D) shape from shading cues. Their underlying
premise is that disrupting depth cues through contrast reversal
interferes with the use of information about 3D shape. However,
more recent studies have tended to downplay the potential role of
3D information in face recognition and place more emphasis on
the importance of surface pigmentation (Bruce & Young, 2012).
For example Liu, Collin, and Chaudhuri (2000) found that recog-
nition was poor for faces with intact 3D information but missing
surface pigmentation. In line with such findings, Bruce and Lang-
ton (1994) developed the second main approach to the impact of
contrast negation by suggesting that it exerts its disruptive influ-
ence because it reverses the brightness of important pigmented
regions of the face, so that light regions of skin become dark, dark
hair becomes light, and so forth. Recent work expanding on this
hypothesis has been carried out by Russell, Sinha, Biederman, and
Nederhouser (2006), who propose that much of the contrast nega-
tion effect results from negating pigmentation cues that carry
substantial information about facial identity. This has been elabo-
rated by Sinha, Russell, and their colleagues (Sinha, Balas, Os-
trovsky, & Russell, 2006; Russell & Sinha, 2007) into a more
general claim that because pigmentation carries much information
about facial identity, any manipulation (e.g., contrast negation)
that disrupts pigmentation rather than shape will have deleterious
effects on recognition. Consistent with this hypothesis, further
work by Santos and Young (2008) shows that contrast negation
effects are not only confined to identity judgments but also impact
on a range of social inferences made to faces in which pigmenta-
tion cues are likely to play a substantial role. Importantly, too, a
study by White (2001) showed a degree of independence of
This article was published Online First May 13, 2013.
Mladen Sormaz, Timothy J. Andrews, and Andrew W. Young, Depart-
ment of Psychology, University of York, UK.
We thank Mel Irwin and Emily Lloyd, whose undergraduate project
provided pilot assessments for some of the experiments reported here,
though in the event none of their data were used.
Correspondence concerning this article should be addressed to Mladen
Sormaz, Department of Psychology, University of York, Heslington, York
YO10 5DD, UK. E-mail: ms930@york.ac.uk
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Experimental Psychology:
Human Perception and Performance
© 2013 American Psychological Association
2013, Vol. 39, No. 6, 1667–1677
0096-1523/13/$12.00 DOI: 10.1037/a0032449
1667