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.     

Categorical Discrimination of Human Facial Expressions by Pigeons: A Test of the Linear Feature Model

Pigeons were trained to discriminate human facial expressions, happiness and anger, in a go/ no-go discrimination procedure. Five pigeons learned to discriminate photographs of the happy and angry faces of 25 different people and showed high levels of transfer to novel faces expressing the training emotions. The pigeons directed their pecks predominantly to the mouth, eyes, or the area between these features. The pigeons were then tested with familiar stimuli in which the upper and lower parts of the face were manipulated separately by substitution or removal of facial features ('eyes-and-eyebrows' and 'mouth'). It was shown that the salience of particular features differed considerably among the birds, but that a linear feature model adequately accounted for discriminative performance of the birds with these stimuli. Furthermore, the discrimination was maintained when these features were inverted. Thus, the so-called Thatcher illusion did not occur. It is suggested that the discrimination is based not on a feature configuration or perceptual gestalt but on an additive integration of individual features.     

The ups and downs of face perception: evidence for holistic encoding of upright and inverted faces

A visual-search task was used to investigate the influence of facial organisation on discrimination of an internal facial feature. Participants searched for a downturned mouth in arrays of one to six faces that differed only in the target feature, with distractor faces containing an upturned mouth. Feature search was tested in four different face contexts: upright unaltered faces, inverted unaltered faces, upright faces in which the internal features were scrambled, or inverted scrambled faces. Normal face organisation facilitated feature search in upright faces, but slowed it in inverted faces. These findings demonstrate an interdependence of features and their configuration in the perceptual analysis of both upright and inverted faces.     

Multiple perceptual strategies used by macaque monkeys for face recognition

Successful integration of individuals in macaque societies suggests that monkeys use fast and efficient perceptual mechanisms to discriminate between conspecifics. Humans and great apes use primarily holistic and configural, but also feature-based, processing for face recognition. The relative contribution of these processes to face recognition in monkeys is not known. We measured face recognition in three monkeys performing a visual paired comparison task. Monkey and humans faces were (1) axially rotated, (2) inverted, (3) high-pass filtered, and (4) low-pass filtered to isolate different face processing strategies. The amount of time spent looking at the eyes, mouth, and other facial features was compared across monkey and human faces for each type of stimulus manipulation. For all monkeys, face recognition, expressed as novelty preference, was intact for monkey faces that were axially rotated or spatially filtered and was supported in general by preferential looking at the eyes, but was impaired for inverted faces in two of the three monkeys. Axially rotated, upright human faces with a full range of spatial frequencies were also recognized, however, the distribution of time spent exploring each facial feature was significantly different compared to monkey faces. No novelty preference, and hence no inferred recognition, was observed for inverted or low-pass filtered human faces. High-pass filtered human faces were recognized, however, the looking pattern on facial features deviated from the pattern observed for monkey faces. Taken together these results indicate large differences in recognition success and in perceptual strategies used by monkeys to recognize humans versus conspecifics. Monkeys use both second-order configural and feature-based processing to recognize the faces of conspecifics, but they use primarily feature-based strategies to recognize human faces.     

Configural features in the context of upright and inverted faces

When faces are turned upside down, recognition is known to be severely disrupted. This effect is thought to be due to disruption of configural processing. Recently, Leder and Bruce (2000, Quarterly Journal of Experimental Psychology A 53 513-536) argued that configural information in face processing consists at least partly of locally processed relations between facial elements. In three experiments we investigated whether a local relational feature (the interocular distance) is processed differently in upside-down versus upright faces. In experiment 1 participants decided in which of two sequentially presented photographic faces the interocular distance was larger. The decision was more difficult in upside-down presentation. Three different conditions were used in experiment 2 to investigate whether this deficit depends upon parts of the face beyond the eyes themselves; displays showed the eye region alone, the eyes and nose, or the eyes and nose and mouth. The availability of additional features did not interact with the inversion effect which was observed strongly even when the eyes were shown in isolation. In experiment 3 all eyes were turned upside down in the inverted face condition as in the Thatcher illusion (Thompson, 1980 Perception 9 483-484). In this case no inversion effect was found. These results are in accordance with an explanation of the face-inversion effect in which the disruption of configural facial information plays the critical role in memory for faces, and in which configural information corresponds to spatial information that is processed in a way which is sensitive to local properties of the facial features involved.     

Perceptual and decisional factors influencing the discrimination of inversion in the Thatcher illusion

The Thatcher illusion (Thompson, 1980) is considered to be a prototypical illustration of the notion that face perception is dependent on configural processes and representations. We explored this idea by examining the relative contributions of perceptual and decisional processes to the ability of observers to identify the orientation of two classes of forms-faces and churches-and a set of their component features. Observers were presented with upright and inverted images of faces and churches in which the components (eyes, mouth, windows, doors) were presented either upright or inverted. Observers first rated the subjective grotesqueness of all of the images and then performed a complete identification task in which they had to identify the orientation of the overall form and the orientation of each of the interior features. Grotesqueness ratings for both classes of image showed the standard modulation of rated grotesqueness as a function of orientation. The complete identification results revealed violations of both perceptual and decisional separability but failed to reveal any violations of within-stimulus (perceptual) independence. In addition, exploration of a simple bivariate Gaussian signal detection model of the relationship between identification performance and judged grotesqueness suggests that within-stimulus violations of perceptual independence on their own are insufficient for producing the illusion. This lack of evidence for within-stimulus configurality suggests the need for a critical reevaluation of the role of configural processing in the Thatcher illusion. (PsycINFO Database Record (c) 2011 APA, all rights reserved).     

How do rhesus monkeys (<Emphasis Type="Italic">Macaca mulatta</Emphasis>) scan faces in a visual paired comparison task?

When novel and familiar faces are viewed simultaneously, humans and monkeys show a preference for looking at the novel face. The facial features attended to in familiar and novel faces, were determined by analyzing the visual exploration patterns, or scanpaths, of four monkeys performing a visual paired comparison task. In this task, the viewer was first familiarized with an image and then it was presented simultaneously with a novel and the familiar image. A looking preference for the novel image indicated that the viewer recognized the familiar image and hence differentiates between the familiar and the novel images. Scanpaths and relative looking preference were compared for four types of images: (1) familiar and novel objects, (2) familiar and novel monkey faces with neutral expressions, (3) familiar and novel inverted monkey faces, and (4) faces from the same monkey with different facial expressions. Looking time was significantly longer for the novel face, whether it was neutral, expressing an emotion, or inverted. Monkeys did not show a preference, or an aversion, for looking at aggressive or affiliative facial expressions. The analysis of scanpaths indicated that the eyes were the most explored facial feature in all faces. When faces expressed emotions such as a fear grimace, then monkeys scanned features of the face, which contributed to the uniqueness of the expression. Inverted facial images were scanned similarly to upright images. Precise measurement of eye movements during the visual paired comparison task, allowed a novel and more quantitative assessment of the perceptual processes involved the spontaneous visual exploration of faces and facial expressions. These studies indicate that non-human primates carry out the visual analysis of complex images such as faces in a characteristic and quantifiable manner.     

The eyes or the mouth? Feature salience and unfamiliar face processing in Williams syndrome and autism

Using traditional face perception paradigms the current study explores unfamiliar face processing in two neurodevelopmental disorders. Previous research indicates that autism and Williams syndrome (WS) are both associated with atypical face processing strategies. The current research involves these groups in an exploration of feature salience for processing the eye and mouth regions of unfamiliar faces. The tasks specifically probe unfamiliar face matching by using (a) upper or lower face features, (b) the Thatcher illusion, and (c) featural and configural face modifications to the eye and mouth regions. Across tasks, individuals with WS mirror the typical pattern of performance, with greater accuracy for matching faces using the upper than using the lower features, susceptibility to the Thatcher illusion, and greater detection of eye than mouth modifications. Participants with autism show a generalized performance decrement alongside atypicalities, deficits for utilizing the eye region, and configural face cues to match unfamiliar faces. The results are discussed in terms of feature salience, structural encoding, and the phenotypes typically associated with these neurodevelopmental disorders.     

The eyes or the mouth? Feature salience and unfamiliar face processing in Williams syndrome and autism

Using traditional face perception paradigms the current study explores unfamiliar face processing in two neurodevelopmental disorders. Previous research indicates that autism and Williams syndrome (WS) are both associated with atypical face processing strategies. The current research involves these groups in an exploration of feature salience for processing the eye and mouth regions of unfamiliar faces. The tasks specifically probe unfamiliar face matching by using (a) upper or lower face features, (b) the Thatcher illusion, and (c) featural and configural face modifications to the eye and mouth regions. Across tasks, individuals with WS mirror the typical pattern of performance, with greater accuracy for matching faces using the upper than using the lower features, susceptibility to the Thatcher illusion, and greater detection of eye than mouth modifications. Participants with autism show a generalized performance decrement alongside atypicalities, deficits for utilizing the eye region, and configural face cues to match unfamiliar faces. The results are discussed in terms of feature salience, structural encoding, and the phenotypes typically associated with these neurodevelopmental disorders.     

Features and their configuration in face recognition

Tanaka and Farah (1993) have proposed a holistic approach to face recognition in which information about the features of a face and their configuration are combined together in the face representation. An implication of the holistic hypothesis is that alterations in facial configuration should interfere with retrieval of features. In four experiments, the effect of configuration on feature recognition was investigated by creating two configurations of a face, one with eyes close together and one with eyes far apart. After subjects studied faces presented in one of the two configurations (eyes-close or eyes-far), they were tested for their recognition of features shown in isolation, in a new face configuration, and in the old face configuration. It was found that subjects recognized features best when presented in the old configuration, next best in the new configuration, and poorest in isolation. Moreover, subjects were not sensitive to configural information in inverted faces (Experiment 2) or nonface stimuli (i.e., houses; Experiments 3 and 4). Importantly, for normal faces, altering the spatial location of the eyes not only impaired subjects’ recognition of the eye features but also impaired their recognition of the nose and mouth features—features whose spatial locations were not directly altered. These findings emphasize the interdependency of featural and configural information in a holistic face representation.     

Face processing in cotton-top tamarins (Saguinus oedipus)

Current research on face processing in primates has focused on a few species, mostly macaques and chimpanzees; to date, only one New World monkey, the squirrel monkey, has been tested. We explored face processing, and the inversion effect in particular, in a New World primate species, the cotton-top tamarin (Saguinus oedipus). In phase 1 of our study, we trained subjects to discriminate between two faces and two scrambled faces; we then presented the tamarins with a series of novel probes in order to determine the features underlying classification. Results showed that the tamarins relied on the external contour of the face for discrimination more than the internal features and their configuration. Statistical analyses revealed no differences in accuracy or response times to upright versus inverted stimuli, and thus no inversion effect. In phase 2, we provided subjects with additional training on the face versus scrambled face discrimination task in order to focus their attention on the configuration of the internal features. Accuracy data revealed individual differences in how tamarins classified these stimuli, even though each subject was trained in the same way. In phase 3, we tested for generalization to a new set of face stimuli, as well as for the capacity to show an inversion effect. For one subject who attended to the configuration of internal features, we found significant evidence of generalization, but no evidence for an inversion effect. Electronic Publication     

Face processing in humans and new world monkeys: the influence of experiential and ecological factors

This study tests whether the face-processing system of humans and a nonhuman primate species share characteristics that would allow for early and quick processing of socially salient stimuli: a sensitivity toward conspecific faces, a sensitivity toward highly practiced face stimuli, and an ability to generalize changes in the face that do not suggest a new identity, such as a face differently oriented. The look rates by adult tamarins and humans toward conspecific and other primate faces were examined to determine if these characteristics are shared. A visual paired comparison (VPC) task presented subjects with either a human face, chimpanzee face, tamarin face, or an object as a sample, and then a pair containing the previous stimulus and a novel stimulus was presented. The stimuli were either presented all in an upright orientation, or all in an inverted orientation. The novel stimulus in the pair was either an orientation change of the same face/object or a new example of the same type of face/object, and the stimuli were shown either in an upright orientation or in an inverted orientation. Preference to novelty scores revealed that humans attended most to novel individual human faces, and this effect decreased significantly if the stimuli were inverted. Tamarins showed preferential looking toward novel orientations of previously seen tamarin faces in the upright orientation, but not in an inverted orientation. Similarly, their preference to look longer at novel tamarin and human faces within the pair was reduced significantly with inverted stimuli. The results confirmed prior findings in humans that novel human faces generate more attention in the upright than in the inverted orientation. The monkeys also attended more to faces of conspecifics, but showed an inversion effect to orientation change in tamarin faces and to identity changes in tamarin and human faces. The results indicate configural processing restricted to particular kinds of primate faces by a New World monkey species, with configural processing influenced by life experience (human faces and tamarin faces) and specialized to process orientation changes specific to conspecific faces.     

The effect of inversion on the encoding of normal and “thatcherized” faces

In the “Thatcher illusion” a face, in which the eyes and mouth are inverted relative to the rest of the face, looks grotesque when shown upright but not when inverted. In four experiments we investigated the contribution of local and global processing to this illusion in normal observers. We examined inversion effects (i.e., better performance for upright than for inverted faces) in a task requiring discrimination of whether faces were or were not “thatcherized”. Observers made same/different judgements to isolated face parts (Experiments 1-2) and to whole faces (Experiments 3-4). Face pairs had the same or different identity, allowing for different process- ing strategies using feature-based or configural information, respectively. In Experiment 1, feature-based matching of same-person face parts yielded only a small inversion effect for normal face parts. However, when feature-based matching was prevented by using the face parts of different people on all trials (Experiment 2) an inversion effect occurred for normal but not for thatcherized parts. In Experiments 3 and 4, inversion effects occurred with normal but not with thatcherized whole faces, on both same- and different-person matching tasks. This suggests that a common configural strategy was used with whole (normal) faces. Face context facilitated attention to misoriented parts in same-person but not in different-person matching. The results indicate that (1) face inversion disrupts local configural processing, but not the processing of image features, and (2) thatcherization disrupts local configural processing in upright faces.     

The inversion effect in infancy: the role of internal and external features

The present work examined the changing role of inner and outer facial features in the recognition of upright and inverted faces in 5-, 7-, and 9-month-olds. Study 1 established that the “inversion effect” (impaired recognition of an inverted face) was present in infants as young as 5 months. In Study 2, internal and external features were inverted separately. Disrupting the internal configuration by inversion impaired recognition at all ages; disrupting the external configuration impaired recognition only at 5-months. In Study 3, an upright familiar face was paired with one having either novel internal or novel external features. The results confirmed that the 5-month-olds used only the external features to recognize faces, whereas older infants were as adept at using internal features as external ones. These findings suggest a shift, after 5 months, away from dependence on external features for face recognition and toward greater reliance on internal ones.     

Beyond U-Shaped Development in Infants' Processing of Faces: An Information-Processing Account

The development of the "inversion" effect in face processing was examined in infants 3 to 6 months of age by testing their integration of the internal and external features of upright and inverted faces using a variation of the "switch" visual habituation paradigm. When combined with previous findings showing that 7-month-olds use integrative processing of an upright face, but featural processing of an inverted face (Cohen & Cashon, 2001a), the present findings suggest that from 3 to 7 months, infants' ability to integrate facial features follows an N-shaped developmental pattern for upright faces and an inverted U-shaped pattern for inverted faces. We discuss these results in terms of a set of domain-general information-processing principles.     

Thatcher's children: development and the Thatcher illusion

Children do not show the same recognition disadvantage for inverted faces as adults do. It has been suggested that this is because the configural encoding (which is disrupted by inversion) becomes more useful or available as we get older. The distinction between configural processing and featural processing, however, is not always clear--it may be a dichotomy or a continuum. The perceived normal-to-grotesque switch in the Thatcher illusion was investigated, as the image was rotated, by people aged between 6 and 75 years. No effect of age was found, with young children showing the same effects as adults--the switch occurring at about 72 degrees. The development of face processing and the nature of facial configural encoding are discussed in the light of this result.     

An inversion effect modified by expertise in capuchin monkeys

The face inversion effect may be defined as the general impairment in recognition that occurs when faces are rotated 180°. This phenomenon seems particularly strong for faces as opposed to other objects and is often used as a marker of a specialized face-processing mechanism. Four brown capuchin monkeys (Cebus apella) were tested on their ability to discriminate several classes of facial and non-facial stimuli presented in both their upright and inverted orientations in an oddity task. Results revealed significantly better performance on upright than inverted presentations of capuchin and human face stimuli, but not on chimpanzee faces or automobiles. These data support previous studies in humans and other primates suggesting that the inversion effect occurs for stimuli for which subjects have developed an expertise.     

Thatcher illusion: dependence on angle of rotation

The expression of a face with its eyes and mouth inverted changes from 'pleasant' to 'grotesque' as the stimulus is rotated from 180 degrees to 0 degree (Thatcher illusion). We determined the angular orientation at which this change occurred for three manipulated faces. Mean thresholds for eighteen observers were found to lie between 94 degrees and 100 degrees relative to the vertical with an average overlap of about 15 degrees between an observer's ascending and descending thresholds. The sudden nature and relatively narrow zone of the changeover suggest a neuronal step-tuning of hypothetical face cells in the human brain, underlying the holistic ('grotesque') versus componential ('pleasant') processing of upright versus upside-down faces. Findings are discussed within the framework of cognitive, neuroimaging, and single-cell studies.     

The time course of processing external and internal features of unfamiliar faces

The time course of processing internal and external facial features was studied in a sequential face matching task, where first a target face was presented, followed by a test face. The exposure duration of the test face was varied systematically (90, 120, 150 ms, and self-paced). In three tasks, participants were instructed to match either the whole face, only external features, or only internal features of the target and test face. Taken together, the results in all the three tasks provide evidence for very fast matching processes. For upright faces, maximal performance was achieved at 90 ms exposure duration and longer exposure durations (120, 150 ms, self-paced) did not improve accuracy. For inverted whole faces, reduced exposure duration resulted in an increase of matching errors, suggesting that below 150 ms of exposure duration, inverted faces cannot be matched reliably. When matching selected facial features only, no such inversion effect was found. Our data challenges previous claims that external features are matched faster than internal: no difference of time course was found between external and internal features. However, external features were matched more accurately.     

The lady's not for turning: rotation of the Thatcher illusion

Inversion is especially detrimental to the processing of faces. This is clearly demonstrated by the Thatcher illusion. It has been suggested that this detriment is due to a loss of holistic or configural processing for inverted faces (Bartlett and Searcy, 1993 Cognitive Psychology 25 281-316). Stürzel and Spillmann (2000 Perception 29 937-942) suggest that this loss of configural processing occurs suddenly as a face is rotated slowly from upright to inverted. This hypothesis is tested in a study of the reaction times taken to indicate that a face has been Thatcherised at various angles of orientation. The results suggest that there is a gradual loss of configural information rather than a rapid switch from one type of processing to another.