Jessica Taubert

The role of holistic processing in face perception: evidence from the face inversion effect.

A large body of research supports the hypothesis that the human visual system does not process a face as a collection of separable facial features but as an integrated perceptual whole. One common assumption is that we quickly build holistic representations to extract useful second-order information provided by the variation between the faces of different individuals. An alternative account suggests holistic processing is a fast, early grouping process that first serves to distinguish faces from other competing objects. From this perspective, holistic processing is a quick initial response to the first-order information present in every face. To test this hypothesis we developed a novel paradigm for measuring the face inversion effect, a standard marker of holistic face processing, that measures the minimum exposure time required to discriminate between two stimuli. These new data demonstrate that holistic processing operates on whole upright faces, regardless of whether subjects are required to extract first- or second-level information. In light of this, we argue that holistic processing is a general mechanism that may occur at an earlier stage of face perception than individual discrimination to support the rapid detection of face stimuli in everyday visual scenes.

Are face representations viewpoint dependent? A stereo advantage for generalising across different views of faces

Almost all previous studies of face recognition have found that matching the same face depicted from different viewpoints incurs both reaction time and accuracy costs. This has been interpreted as evidence that the underlying neural representations of faces are viewpoint-specific, but such a conclusion depends on the experimental data being an accurate reflection of real-world viewpoint generalisation. An equally plausible explanation for poor viewpoint generalisation in experimental situations is that important information that is normally used to generalize across views in real-world settings is not available in the experiment. Stereoscopic information about the three-dimensional structure of the face is systematically misleading in nearly all previous investigations of face recognition, since a face depicted on a computer monitor contains explicit stereoscopic information that the face is flat. The current experiment demonstrates that viewpoint costs are reduced by depicting the face with stereoscopic three-dimensionality (compared to a synoptically presented face), raising the possibility that the viewpoint costs found in face recognition experiments might be a better reflection of the information that is typically unavailable in the experimental stimuli than of the underlying neural representation of facial identity.

Love at second sight: Sequential dependence of facial attractiveness in an on-line dating paradigm

Millions of people use online dating sites each day, scanning through streams of face images in search of an attractive mate. Face images, like most visual stimuli, undergo processes whereby the current percept is altered by exposure to previous visual input. Recent studies using rapid sequences of faces have found that perception of face identity is biased towards recently seen faces, promoting identity-invariance over time, and this has been extended to perceived face attractiveness. In this paper we adapt the rapid sequence task to ask a question about mate selection pertinent in the digital age. We designed a binary task mimicking the selection interface currently popular in online dating websites in which observers typically make binary decisions (attractive or unattractive) about each face in a sequence of unfamiliar faces. Our findings show that binary attractiveness decisions are not independent: we are more likely to rate a face as attractive when the preceding face was attractive than when it was unattractive.

Chimpanzee Faces are ‘Special’ to Humans

A longstanding proposal is that primates, including humans, might have an innate representation of face structure. But, if humans have such a representation, how broad is its form: limited to coding conspecifics, or general enough to cover related species? The results reported here show adult humans process faces of chimpanzees in a way previously assumed to be exclusive to human faces. The composite effect was used to provide the first direct test of so-called holistic processing. Despite no lifetime experience of chimpanzees, adult humans showed a substantial composite effect for chimpanzee faces, and also an inversion effect. There was no similar evidence of holistic processing for faces of species of greater phylogenic distance from humans, including gorillas, spider monkeys, sheep, chickens, and Jacky lizards; nor was there any effect for non-face objects. In contrast to the holistic processing results, discrimination of chimpanzee faces was, as expected, poor. In the context of evidence that poor discrimination of heterospecific faces arises from a process of perceptual narrowing in infancy, our results suggest that adults retain some aspects of a broader bandwidth present in neonates (holistic processing) but lose others (discrimination).

Different coding strategies for the perception of stable and changeable facial attributes

Perceptual systems face competing requirements: improving signal-to-noise ratios of noisy images, by integration; and maximising sensitivity to change, by differentiation. Both processes occur in human vision, under different circumstances: they have been termed priming, or serial dependencies, leading to positive sequential effects; and adaptation or habituation, which leads to negative sequential effects. We reasoned that for stable attributes, such as the identity and gender of faces, the system should integrate: while for changeable attributes like facial expression, it should also engage contrast mechanisms to maximise sensitivity to change. Subjects viewed a sequence of images varying simultaneously in gender and expression, and scored each as male or female, and happy or sad. We found strong and consistent positive serial dependencies for gender, and negative dependency for expression, showing that both processes can operate at the same time, on the same stimuli, depending on the attribute being judged. The results point to highly sophisticated mechanisms for optimizing use of past information, either by integration or differentiation, depending on the permanence of that attribute.

Identity Aftereffects, but Not Composite Effects, are Contingent on Contrast Polarity

People are very good at discriminating faces despite their gross similarity. Our ability to capitalize on the variance that exists between faces has been attributed to an adaptive face-coding system. Evidence from psychophysical adaptation paradigms has generally supported this view, although results from other paradigms have suggested alternative accounts. The composite face-effect (CFE), for example, has been used to argue that holistic processing supports face discrimination. The question addressed here is whether the notion of holistic processing can be integrated with the notion of an adaptive face-coding system. Experiment 1 clearly demonstrates that the CFE acts on contrast-reversed faces while experiment 2 reveals that face-space adaptation depends on contrast polarity, consistent with previous observations of contrast-reversal impairing discrimination performance. Our results suggest that two popular theoretical frame-works in the face-recognition literature can be integrated, with holistic processing occurring at an earlier stage than face discrimination.

The effect of face inversion for neurons inside and outside fMRI-defined face-selective cortical regions

It is widely believed that face processing in the primate brain occurs in a network of category-selective cortical regions. Combined functional MRI (fMRI)-single-cell recording studies in macaques have identified high concentrations of neurons that respond more to faces than objects within face-selective patches. However, cells with a preference for faces over objects are also found scattered throughout inferior temporal (IT) cortex, raising the question whether face-selective cells inside and outside of the face patches differ functionally. Here, we compare the properties of face-selective cells inside and outside of face-selective patches in the IT cortex by means of an image manipulation that reliably disrupts behavior toward face processing: inversion. We recorded IT neurons from two fMRI-defined face-patches (ML and AL) and a region outside of the face patches (herein labeled OUT) during upright and inverted face stimulation. Overall, turning faces upside down reduced the firing rate of face-selective cells. However, there were differences among the recording regions. First, the reduced neuronal response for inverted faces was independent of stimulus position, relative to fixation, in the face-selective patches (ML and AL) only. Additionally, the effect of inversion for face-selective cells in ML, but not those in AL or OUT, was impervious to whether the neurons were initially searched for using upright or inverted stimuli. Collectively, these results show that face-selective cells differ in their functional characteristics depending on their anatomicofunctional location, suggesting that upright faces are preferably coded by face-selective cells inside but not outside of the fMRI-defined face-selective regions of the posterior IT cortex.

Effect of familiarity and viewpoint on face recognition in chimpanzees.

Numerous studies have shown that familiarity strongly influences how well humans recognize faces. This is particularly true when faces are encountered across a change in viewpoint. In this situation, recognition may be accomplished by matching partial or incomplete information about a face to a stored representation of the known individual, whereas such representations are not available for unknown faces. Chimpanzees, our closest living relatives, share many of the same behavioral specializations for face processing as humans, but the influence of familiarity and viewpoint have never been compared in the same study. Here, we examined the ability of chimpanzees to match the faces of familiar and unfamiliar conspecifics in their frontal and 3/4 views using a computerized task. Results showed that, while chimpanzees were able to accurately match both familiar and unfamiliar faces in their frontal orientations, performance was significantly impaired only when unfamiliar faces were presented across a change in viewpoint. Therefore, like in humans, face processing in chimpanzees appears to be sensitive to individual familiarity. We propose that familiarization is a robust mechanism for strengthening the representation of faces and has been conserved in primates to achieve efficient individual recognition over a range of natural viewing conditions.

How the Thatcher illusion reveals evolutionary differences in the face processing of primates

Face recognition in humans is a complex cognitive skill that requires sensitivity to unique configurations of eyes, mouth, and other facial features. The Thatcher illusion has been used to demonstrate the importance of orientation when processing configural information within faces. Transforming an upright face so that the eyes and mouth are inverted renders the face grotesque; however, when this “Thatcherized” face is inverted, the effect disappears. Due to the use of primate models in social cognition research, it is important to determine the extent to which specialized cognitive functions like face processing occur across species. To date, the Thatcher illusion has been explored in only a few species with mixed results. Here, we used computerized tasks to examine whether nonhuman primates perceive the Thatcher illusion. Chimpanzees and rhesus monkeys were required to discriminate between Thatcherized and unaltered faces presented upright and inverted. Our results confirm that chimpanzees perceived the Thatcher illusion, but rhesus monkeys did not, suggesting species differences in the importance of configural information in face processing. Three further experiments were conducted to understand why our results differed from previously published accounts of the Thatcher illusion in rhesus monkeys.

Geometric distortions affect face recognition in chimpanzees (Pan troglodytes) and monkeys (Macaca mulatta)

All primates can recognize faces and do so by analyzing the subtle variation that exists between faces. Through a series of three experiments, we attempted to clarify the nature of second-order information processing in nonhuman primates. Experiment one showed that both chimpanzees (Pan troglodytes) and rhesus monkeys (Macaca mulatta) tolerate geometric distortions along the vertical axis, suggesting that information about absolute position of features does not contribute to accurate face recognition. Chimpanzees differed from monkeys, however, in that they were more sensitive to distortions along the horizontal axis, suggesting that when building a global representation of facial identity, horizontal relations between features are more diagnostic of identity than vertical relations. Two further experiments were performed to determine whether the monkeys were simply less sensitive to horizontal relations compared to chimpanzees or were instead relying on local features. The results of these experiments confirm that monkeys can utilize a holistic strategy when discriminating between faces regardless of familiarity. In contrast, our data show that chimpanzees, like humans, use a combination of holistic and local features when the faces are unfamiliar, but primarily holistic information when the faces become familiar. We argue that our comparative approach to the study of face recognition reveals the impact that individual experience and social organization has on visual cognition.