Visual processing of biological motion (BM) produced by living organisms is of immense value for successful daily-life activities and, in particular, for adaptive social behavior and nonverbal communication. Investigation of BM perception in neurodevelopmental disorders related to autism, preterm birth, and genetic conditions substantially contributes to our understanding of the neural mechanisms underpinning the extraordinary tuning to BM. The most prominent research outcome is that patients with daily-life deficits in social cognition are also compromised on visual body motion processing. This raises the question of whether performance on body motion perception tasks may serve a hallmark of social cognition. Overall, the findings highlight the role of structural and functional brain connectivity for proper functioning of the neural circuitry involved in BM processing and visual social cognition that share topographically and dynamically overlapping neural networks.
Anecdotal evidence abounds on the ability of humans to extract accurate perceptual information based solely on the dynamic outline of a human body moving in the distance or in dim light. This is especially true when the walker is someone we know personally but occurs as well with acquaintances and strangers. For example, a century ago, the Russian psychiatrist Pyotr B. Gannushkin, a student of Sergei S. Korsakov, claimed that he was able to recognize mental conditions of patients simply by observing their changing outline as they moved about in a dimly lit room (Gannushkin 1933).
Visual processing of biological motion (BM) produced by living organisms is of immense value for successful daily-life activities and, in particular, for adaptive social behavior and nonverbal communication. Almost 4 decades ago, Gunnar Johansson, a prominent researcher at the Uppsala University in Sweden, introduced point-light displays depicting different types of human activities (Johansson 1973). Since that time, one of the most robust and replicable findings is that the visual system is exquisitely tuned to point-light BM, and in a number of studies, attempts have been made to clarify possible accounts for this extraordinary tuning (e.g., Proffitt et al. 1984; Cutting et al. 1988; Neri et al. 1998; Pavlova and Sokolov 2000; Jacobs et al. 2004; Bidet-Ildei et al. 2006; Thurman et al. 2010; Christensen et al. 2011). The main advantage of this methodology is that it helps to minimize availability of structural cues and thereby to separate information revealed by motion from other sources (e.g., shape, color). For these reasons, I focus here primarily on processing of point-light BM instead of full-light body motion.
Although in point-light displays only a few dots placed on the main joints of an otherwise invisible human body moving against a dark background are visible (Fig. 1), not only the mature visual system of adult humans is exquisitely sensitive to BM: human infants (Fox and McDaniel 1982), monkeys (Oram and Perrett 1994), newly hatched domestic chicks (Regolin et al. 2000; Vallortigara et al. 2005; Vallortigara and Regolin 2006), and bottlenose dolphins (Herman et al. 1990) perceive point-light displays. Mentally retarded adults have no difficulties in identifying such types of point-light locomotion as walking, running, and stair climbing (Sparrow et al. 1999).
The visual sensitivity to BM is not restricted to human forms. Adults and children identify animated point-light quadrupeds, such as elk, baboon, dog, cat, and camel as well as point-light birds (Mather and West 1993; Pavlova et al. 2001), discriminate between texture-defined displays of quadrupeds, and can endure substantial variations in reversed polarity (Bellefeuille and Faubert 1998). Comparative studies demonstrate that not just humans demonstrate high sensitivity to point-light displays. Cats discriminate a point-light cat from a similar display, in which local motion vectors are spatially scrambled (Blake 1993). Pigeons discriminate between displays with dots pasted on a pigeon and on a toy dog (Omori and Watanabe 1996) or between point-light walking and pecking birds (Dittrich et al. 1998). Recent data, however, indicate that the visual sensitivity to BM in non-humans may be restricted. For example, baboons (Papio papio) demonstrate very limited transfer to novel point-light stimuli (Parron et al. 2007). Rhesus monkeys require extensive training in order to discriminate forward versus backward point-light human locomotion (Vangeneugden et al. 2009), though single-cell recording shows that some neuronal populations in monkey are sensitive to forward and backward human locomotion without explicit discrimination training (Oram and Perrett 1996). In a water-maze visual discrimination task, rats distinguish between leftward and rightward motion of point-light walking humans, but they appear to be unable to generalize to a novel point-light display (MacKinnon et al. 2010). Female, but not male, common marmosets (Callithrix jacchus) exhibit curiosity to point-light BM of a hen as compared with inverted, spatially scrambled, static, and rotating display versions (Brown et al. 2010).
Latest behavioral findings suggest that in typically developing children, the perception of BM and social cognitive abilities are tightly linked, and therefore, during development, the perceptual system for analyzing BM might be functionally integrated with social abilities. In accordance with this assumption, human infants aged 12 months share attentional tuning expressed by an upright oriented, but not by inverted to 180° in the image plane point-light figure (Yoon and Johnson 2009). Furthermore, infants aged 9 months with an increased interest in BM and social causality possess higher scores on the developmental index than those who show a decreased interest (Kutsuki et al. 2009). The close connection between BM perception and social abilities is also confirmed by the following data. First, intranasally administered oxytocin, a neuropeptide that facilitates social abilities, selectively increases (relative to placebo) visual sensitivity to camouflaged BM (Kéri and Benedek 2009) and alters electroencephalographic (EEG) mu/alpha and beta oscillatory brain activity during BM processing (Perry et al. 2010). Second, patients with tumors in the left lateral cerebellum are compromised on a BM task (Sokolov, Gharabaghi, et al. 2010). The left lateral cerebellar substructures that are affected in these patients are also active during social cognition tasks (Kilts et al. 2003; Grèzes et al. 2004b, 2007; Gobbini et al. 2007). Third, healthy adults are proficient in inferring emotions, intentions, and dispositions of others represented by point-light BM (Runeson and Frykholm 1983; Dittrich et al. 1996; Brownlow et al. 1997; Pollick et al. 2001; Atkinson et al. 2004; Heberlein et al. 2004; Clarke et al. 2005; Rose and Clarke 2009; Sokolov et al. 2011). For example, perceivers can reliably judge emotional content of dance represented by a few moving dots placed on a dancer's body (Dittrich et al. 1996). Furthermore, as shown by Sverker Runeson, a student of Gunnar Johansson, observers can discriminate between deceptive and true intentions conveyed by body motion of point-light actors, and true information is precisely detected despite misleading endeavors (Runeson and Frykholm 1983; see also Grèzes et al. 2004a, 2004b for brain imaging data that elucidate the neural substrates supporting this ability). As stated by Runeson and Frykholm (1983), “‘hidden’ person properties might be conceptually linked with characteristics and states of a person's motor system whereby aspects of personality and emotion would be kinematically specified” (p. 585). The potential significance of this statement, the principle of kinematic specification of dynamics (KSD), for our understanding of social behavior and cognition is still underestimated. Finally, BM detection is modulated by the emotional context of gait (Chouchourelou et al. 2006), and the ability to recognize emotions in displays portraying human locomotion is related to the visual sensitivity to BM (Ikeda and Watanabe 2009). Understanding of social interaction (such as dancing or kicking) between agents in point-light displays enhances visual sensitivity to camouflaged BM (Neri et al. 2006; Manera et al. 2011; see Hirai and Kakigi 2009 for brain imaging data).
Analysis of BM perception in neurodevelopmental disorders substantially contributes to further understanding of the mechanisms underpinning the extraordinary tuning to BM. Moreover, these studies help to clarify the interrelationship between BM perception and social cognitive abilities. At this juncture, I will discuss the findings available in typically developing children, in children who were born preterm, in autistic individuals, and in patients with genetic developmental conditions, such as Williams syndrome (WS), Down syndrome (DS), and Fragile X syndrome (FXS). I argue that performance on BM tasks may serve a hallmark of deficits in visual social cognition and competence. I will also highlight the findings uncovering the role of structural and functional brain connectivity for proper functioning of the neural circuitry involved in visual BM processing and social cognition that apparently share topographically and dynamically overlapping networks.
Development of Biological Motion Processing
The sensitivity to BM is thought to emerge early in perceptual development and is well preserved in the elderly. Aging does not affect perception of unmasked point-light displays (Norman et al. 2004) and moderately affects the perception of BM embedded into simultaneous noise (Billino et al. 2008; Pilz et al. 2010). Newly hatched domestic chicks and 2-day-old human babies are reported to exhibit spontaneous preference for a point-light display of a hen (and a cat) walking as if on a treadmill over random motion and inverted displays (Vallortigara et al. 2005; Vallortigara and Regolin 2006; Simion et al. 2008). Preference for BM is, therefore, postulated to be predisposed in the brain of vertebrates pointing to the existence of an inbuilt mechanism for BM detection. The findings in infants (Simion et al. 2008), however, contradict earlier data. First, by using a forced-choice preferential looking paradigm, it was shown that only infants older than 4 months exhibit preference for a point-light human figure running with no net translation over dynamic noise or the same display inverted 180° in the image plane (Fox and McDaniel 1982). For more complex and rare movement of point-light hands clasping an invisible glass, this preference was observed only for 6-month-olds. Second, 4-day-old neonates look longer at a single dot moving in conflict with the two-third power law, that is, nonbiologically, than at a dot moving in accordance with this law (Méary et al. 2007). Recently, 2-day-old human babies have been reported to exhibit spontaneous preference for a point-light hen over a non-BM display (a rotating static frame from a sequence depicting a point-light hen) but not over a spatially scrambled display that consists of the same amount of absolute motion but lacks an implicit body structure (Bardi et al. 2011). It appears therefore that preference of newborns to canonical BM is limited. In a series of systematic studies conducted in the 1980s, it was shown that by 3–5 months of age (but not earlier), infants discriminate a point-light walker from similar displays (Bertenthal et al. 1984, 1987) or from a walker without occlusion between dots during a gait cycle (Bertenthal et al. 1985). Infants aged 6 months are surprised if a point-light walker, but not a scrambled or inverted 180° walker, appears to pass through a table, that is, can interpret this event as implausible (Moore et al. 2007). Further research should be directed at clarification of whether BM processing is largely intrinsic rather than acquired through experience or, at least, whether it requires a postnatal period of growth.
Visual experience and perceptual learning beneficially affect performance on a variety of BM tasks (e.g., Grossman et al. 2004; Jacobs et al. 2004; Hiris et al. 2005; Jastorff et al. 2009). Typically developing young children steadily improve in their ability to spontaneously recognize point-light BM of humans and animals, with adult levels of performance achieved by the age of 5 years (Pavlova et al. 2001). Adults and 6-year-olds tolerate distortions caused by reverse transformation, that is, by showing the point-light film backwards, exhibiting the apparent facing effect (Pavlova et al. 2002).
Adding dynamic noise to a point-light BM display makes the task deliberately more demanding by reducing the sensory evidence available to observers. In their pioneering work, Cutting et al. (1988) report that the most effective mask to camouflage BM is composed of spatially scrambled dots on the body joints. Perception of BM embedded into a simultaneous mask requires a longer period of development and improves linearly with age (Freire et al. 2006; Annaz et al. 2010). In 14-year-old adolescents, learning for detecting a walker camouflaged by a scrambled walker mask proceeds more slowly than in young adults (Pavlova et al. 2000). Developmental trends in processing of camouflaged BM are in line with brain imaging data. Functional magnetic resonance imaging (fMRI) indicates stronger right posterior superior temporal sulcus (pSTS) activation in adults than in 5- to 7-year-old children, whereas in children, the right fusiform gyrus responds to BM stronger than in adults (Lichtensteiger et al. 2008). It is becoming apparent that the neural network for detection of camouflaged BM in 5- to 7-year-olds is still immature.
Upside-down presentation prevents infants from discriminating a walker from similar displays (Bertenthal et al. 1984, 1985, 1987). Orientation specificity is one of the most intriguing phenomena in BM perception (Sumi 1984; Mitkin and Pavlova 1990; Pavlova and Sokolov 2000, 2003). Although an inverted display retains the same relational structure and amount of motion as an upright one, it is shown that inversion severely impedes integration of the local dots into a point-light figure (Pavlova and Sokolov 2000; Tadin et al. 2002). Sensitivity to BM drops by approximately 1/2 log-unit with inversion and is unaffected by amblyopia (Neri et al. 2007). The upside-down presentation has a number of advantages as a control for a BM task. The most important of them is that manipulation with display orientation allows one to keep the same amount of sensory information available. The only difference between upright and inverted displays is that in upright displays, the trajectories of moving dots correspond to human walking under natural conditions and in inverted displays, these trajectories contradict gravity forces. The human visual system is sensitive to the gravity laws (Pavlova and Sokolov 2000; Shipley 2003), and even newly hatched chicks appear to have a predisposition for gravity forces in point-light BM displays (Vallortigara and Regolin 2006). Visual impressions from inverted point-light displays remain impenetrable with respect to one's knowledge about display inversion (Pavlova and Sokolov 2003).
For clarification of the origins of orientation specificity in BM perception, brain imaging data might be of help. These data indicate a difference in brain activity during BM perception under upright and inverted display conditions. An upright walker elicits an increased blood oxygen level–dependent (BOLD) response over the right pSTS, whereas under inverted conditions, the BOLD signal over the pSTS is diminished by one half (Grossman and Blake 2001). An increased BOLD response to an upright as compared with an inverted walker is also reported over the frontal cortices (Grèzes et al. 2001). An easily recognizable upright walker robustly elicits consecutive peaks of oscillatory magnetoencephalographic (MEG) cortical activity over the left occipital, parietal, and right temporal lobes. In response to an inverted walker, however, the peak of oscillatory activity is topographically restricted to the left occipital cortex and does not reach the right temporal cortex (Pavlova et al. 2004). Negative component of the event-related potentials (ERPs) with a peak at 180 ms is higher for upright than for inverted BM over the left posterior cingulate and the right middle occipital gyrus (Jokisch, Daum, et al. 2005).
Neural Network Dedicated to Biological Motion Processing
Groundbreaking single-cell recording in awake macaque monkey by David I. Perrett and his coworkers (e.g., Perrett et al. 1985, 1990; Oram and Perrett 1994), brain imaging in humans (Grossman et al. 2000; Grèzes et al. 2001; Vaina et al. 2001; Grossman and Blake 2002; Pavlova et al. 2004; Pavlova, Birbaumer, et al. 2006; Peuskens et al. 2005; Peelen et al. 2006; Krakowski et al. 2011), and neuropsychological lesion studies (e.g., Battelli et al. 2003; Saygin 2007; Gilaie-Dotan et al. 2011) suggest that visual BM processing engages a specialized neural network that differs from processing of other moving stimuli. Positron emission tomography (PET) and fMRI indicate that the intact network dedicated to BM processing involves portions of the fusiform gyrus (Grossman et al. 2000; Vaina et al. 2001; Peelen et al. 2006), extrastriate body area (Grossman and Blake 2002; Peelen et al. 2006; Jastorff and Orban 2009), and portions of the parietal (Bonda et al. 1996; Grèzes et al. 2001) and frontal cortices (Saygin et al. 2004) primarily in the right hemisphere. The right pSTS is considered a key node for visual processing of expressive and meaningful BM (Grossman et al. 2000, 2010; Grossman and Blake 2001, 2002; Beauchamp et al. 2003; Puce and Perrett 2003; Pelphrey, Morris, Michelich, et al. 2005; Peuskens et al. 2005; Gobbini et al. 2007; Michels et al. 2009; see Fig. 2). However, fMRI provides only indirect measure of neural activity, assessing the changes in cerebral metabolism that are coupled in a complex way to changes in neural activity (Logothetis 2008). More important, most of the available fMRI findings are restricted to localization of brain regions involved in visual BM processing. Yet, for understanding proper functioning of neural circuits and its pathology, one has to consider the changes in brain activation unfolding over time. In healthy adults, visually perceived point-light BM robustly elicits consecutive peaks of oscillatory activity over the left occipital (100 ms), bilateral parietal (130 ms), and right temporal (170 ms) lobes (Fig. 3; Pavlova et al. 2004). This posterior portion of temporal cortex has been reported to be active during visual BM perception in fMRI and PET studies. Consistent with the MEG data on oscillatory brain activity, EEG reveals a topographically similar albeit slower response to point-light BM over the extrastriate and right temporal cortex (Hirai et al. 2003; Jokisch, Daum, et al. 2005; Krakowski et al. 2011). In contrast, the MEG response to an unattended walker is restricted to the left parieto-occipital cortex and does not reach the parietal and right temporal cortices (Pavlova, Birbaumer, et al. 2006). Modulation of the neural response to point-light BM by attention is also confirmed by EEG (Hirai et al. 2005) and fMRI (Safford et al. 2010) findings in healthy adults.
Brain imaging data appear to be in general agreement with lesional findings, which originate mostly from studies with a relatively small sample size and patients with heterogeneous lesion extent, etiology, and topography. Early case studies suggest that an ability to identify human actions in point-light displays is preserved in patients with bilateral damage to occipito-parietal cortex associated with low-level motion processing, if the temporal lobe is relatively spared (e.g., Vaina et al. 1990; McLeod et al. 1996). Deficits in visual perception of BM have been reported to result from unilateral right or left damage to the inferior parietal cortex (Battelli et al. 2003), lesions to the right temporal lobe in stroke patients (Vaina and Gross 2004), and damage to the superior parietal, lateral temporal, and medial frontal areas in patients with focal cortical lesions of diverse topography (Billino et al. 2009). The only study conducted in a larger cohort of 60 unilateral (right or left) stroke patients suggests involvement of premotor and parietotemporal regions, including the STS, with causal relationships to deficits in BM perception (Saygin 2007). The key role of the STS is further supported by the findings on reduced visual sensitivity to camouflaged BM following transcranial magnetic stimulation over the right pSTS (Grossman et al. 2005).
The right posterior temporal cortex appears to be a cornerstone of the social brain: This cortical area is heavily involved in processing of visual information about animacy, agency, emotions, dispositions, and intentions of others revealed by actions and movement (Pelphrey et al. 2004; Grèzes et al. 2004a, 2007; Morris et al. 2005; Kret et al. 2010), for example, in Heider-and-Simmel-like (Heider and Simmel 1944) animations (e.g., Castelli et al. 2000, 2002; Schultz et al. 2005; Gobbini et al. 2007; Tavares et al. 2008; Pavlova, Guerreschi, Lutzenberger, Krägeloh-Mann, 2010). Heider-and-Simmel-like displays have been an indispensable tool for the investigation of social interaction revealed by motion. When simple geometric shapes (disks or triangles) irregularly move, observers systematically tend to perceive these events as social interaction (one shape pushes, entrains, chases, or launches the other one), and their impressions depend on different spatiotemporal characteristics (velocity, acceleration). Moreover, particular personal traits, needs, dispositions, and emotions (aggression, fear, or escape of safety) are often attributed to these shapes. In the same participants, fMRI activation during point-light BM processing overlaps topographically, especially, in the right pSTS, with the network engaged in visual perception of agency in Heider-and-Simmel animations (Gobbini et al. 2007). It seems, however, that the right STS is employed by different brain networks subserving social cognition tasks. Comparison of temporal characteristics of the oscillatory MEG response to point-light BM and Heider-and-Simmel-like animations provides support for this view. In response to point-light BM, the peak in oscillatory activity occurs at 170 ms from the stimulus onset (Pavlova et al. 2004). Heider-and-Simmel-like animations elicit the first gamma boosts over the posterior temporal cortex at nearly the same latency after the culmination point of the event (Pavlova, Guerreschi, Lutzenberger, and Krägeloh-Mann 2010). In response to Heider-and-Simmel animations, however, a subsequent boost of gamma activity occurs at about 300 ms after the culmination point that may reflect more complex processing of these animations. It appears, therefore, that the right temporal cortex is engaged in processing of BM and Heider-and-Simmel animations with different temporal characteristics of brain activity. This view is in line with the outcome of the meta-analysis of fMRI findings in the right STS that emphasizes the role of network connectivity (Hein and Knight 2008).
To the present date, neural mechanisms underlying the development of the ability to perceive BM have been poorly understood. There are a few EEG studies trying to shed light on this issue. In healthy 8-month-old infants, the averaged negative amplitude of the ERPs in the right hemisphere is greater in response to canonical than to scrambled point-light BM (Hirai and Hiraki 2005). While viewing upright as compared with inverted point-light BM (walking and kicking), infants of this age exhibit larger positive ERP amplitude over the right parietal cortex at a latency of 200–300 ms (Reid et al. 2006). These data are in agreement with neuroimaging studies in healthy adults inasmuch as the network dedicated to BM processing is more often localized in the right hemisphere. In 8-month-old infants, viewing of biomechanically impossible motion results in increased ERP positive amplitude over the parietal channels (Reid et al. 2008). This indicates that the brain of 8-month-old infants processes biomechanically impossible motion in a different way than doable actions. In 7- to 14-year-olds, both the P1 amplitude (intensity) and N1 latency (timing) on the bilateral occipitotemporal electrodes linearly decrease with age in response to body motion (Hirai, Watanabe, et al. 2009). Functional near-infrared spectroscopy in 5-month-old infants during perception of full-light motion of eyes, mouth, and hands indicates dissociable patterns of responses to different types of BM (Lloyd-Fox et al. 2010) that topographically resemble fMRI activation in the adult brain (Pelphrey et al. 2005). Comparison of passive viewing of BM (a full-light walker and a robot) versus non body motion (a grandfather clock and a disjointed mechanical figure) in 7- to 10-year-olds shows increasing specificity for BM with age in the right STS (Carter and Pelphrey 2006).
It appears that we are still far apart from understanding the neural network underpinning BM processing. Veridical perception of BM requires intact communication within the distributed brain network. Information flow, however, is primarily considered to be limited to cortico-cortical connections, and engagement of brain structures beyond the cerebral cortex, such as the cerebellum and amygdala, has not yet been properly identified, though brain activation in these regions is incidentally reported (Bonda et al. 1996; Grossman et al. 2000; Vaina et al. 2001; Ptito et al. 2003). Findings in patients with cerebellar lesions of similar extent and topography point to left lateral cerebellar engagement in the network dedicated to BM processing (Fig. 4; Sokolov, Gharabaghi, et al. 2010, but cf. Jokisch, Troje, et al. 2005 in patients with heterogeneous lesion topography). Most recent fMRI data with subsequent seed-voxel regression functional connectivity analysis along with dynamic causal modeling indicate that the left lateral cerebellum bidirectionally communicates with the right pSTS (Sokolov, Erb, et al. 2010). For the first time, this work indicates that the left lateral cerebellum not only receives input from the right STS but also modulates functional activity in the right STS. The evidence for 2-way communication between the right STS and the left lateral cerebellum agrees with the previous findings on closed loops between the cerebellum and the parietal, motor, and frontal cortices (Strick et al. 2009). Findings in children with altered brain connectivity further shed light on the role of intact communication within the network specialized for BM.
Biological Motion Processing in Children Born Preterm
In a series of studies, point-light BM processing was explored in children with periventricular leucomalacia (PVL), the dominant form of brain injury in survivors of premature birth. The prevalence of children who were born preterm (below 37 weeks of gestation) is about 11% in Western European countries reaching 12–13% in the United States of America (Goldenberg et al. 2008). Among those children, 30–50% may suffer a specific early damage to peritrigonal brain regions, PVL, which is characterized by gliosis in the white matter and tissue loss with secondary ventricular dilatation (Krägeloh-Mann et al. 1999) thereby affecting brain connectivity (Thomas et al. 2005; Skranes et al. 2007; Smyser et al. 2010).
Adolescents who were born prematurely with signs of PVL on an MRI scan (without additional cortical or subcortical lesions, aged 13–17 years, verbal IQ in the normal range) exhibit compromised ability for visual processing of BM: their visual sensitivity is lower than in term-born and preterm-born peers with normal MRI scans (Fig. 5; Pavlova et al. 2003, 2005). The severity of this impairment is specifically related to topography and volumetric lesion extent: neither frontal nor temporal PVL relates to deficits in BM processing, whereas the sensitivity decreases with an increase in the volumetric parieto-occipital PVL extent in both hemispheres, most strongly in the right one (Pavlova, Sokolov, et al. 2006).
Different vulnerability of BM and global motion perception in premature children aged 5–9 years suggests that deficiencies on these tasks are connected with different neural networks (Taylor et al. 2009). It remains unclear, however, whether deficits in biological and global motion perception are related to the occurrence of lesions because the data had been collapsed across preterms with and without brain injury. In low birth weight preterm children aged 5–6 years, deficits in motion-defined form recognition relate to the presence of brain injury and retinopathy of prematurity, that is, to preterm birth complications, rather than to a history of prematurity per se (Jakobson et al. 2006). In accordance with this, on a motion-defined form (segmented motion) recognition task, the sensitivity of 10-year-old full-terms and preterms without PVL is equal, and it is higher than in preterms with lesions (Guzzetta et al. 2009).
Combination of MRI with psychophysics reveals that the factor of premature birth per se is irrelevant for visual BM processing (Pavlova et al. 2003, 2005; Pavlova, Sokolov, et al. 2006). Instead, deficiencies are associated with PVL that, by affecting brain connectivity, leads to disintegration of the underlying neural network. This assumption is supported by the MEG evidence. The early (140–170 ms) evoked root mean square (RMS) response to BM over the right parietal cortex is weaker in PVL patients (Pavlova, Marconato, et al. 2006). In contrast, for a control configuration (a spatially scrambled walker), the MEG response does not differ between PVL patients and healthy controls. The modulations of the MEG response are, therefore, stimulus specific and topographically distinct.
Similar to healthy adults, in typically developing adolescents, the oscillatory MEG response to BM peaks at a latency of 170 ms over the right temporoparietal cortex (Fig. 6; Pavlova et al. 2007). This increase is absent in PVL patients. Instead, peaks in the oscillatory response of lower frequency occur in PVL patients later, at a latency of 290 ms over the left temporal region. The posterior thalamocortical fibers are longer, thinner, and less numerous in PVL patients, primarily because of ventricular enlargement along with gliosis in the white matter (Thomas et al. 2005; Smyser et al. 2010). This might account for the delayed latency of oscillatory MEG response. At first glance, the left-lateralized topography of MEG peak appears surprising since the damage to periventricular regions is bilateral and symmetrical, almost equally affecting both hemispheres. One possible account for the left-hemispheric shift in PVL patients is that in healthy participants, the posterior thalamocortical fibers are usually more numerous and widely distributed on the left side (Thomas et al. 2005), and therefore, the brain compensatory potential might be greater in the left hemisphere.
Despite everyday visual experience with human locomotion that improves neural processing of BM (Grossman et al. 2004), even adolescents with mild PVL exhibit compromised processing of point-light displays. In particular, alterations in brain connectivity with the right temporal and frontal cortices may lead to disintegration of neural networks underpinning visual BM processing (Pavlova et al. 2009). This is in a good agreement with the data of unilateral stroke patients (Saygin 2007) and fMRI findings (Saygin et al. 2004). In contrast, long-term visual deprivation does not affect BM processing. After 40 years of early acquired blindness, the patient MM is reported to show intact sensitivity to BM (Fine et al. 2003).
Some experimental data and theoretical reasoning suggest that production and perception of BM may be closely linked and share a common representational neural network (e.g., Rizzolatti and Sinigaglia 2010). PVL patients represent a proper model for exploring this issue because some individuals exhibit signs of motor disorders in a form of bilateral leg-dominated spastic cerebral palsy ranging from mild impairments in walking pattern to a complete walking disability. Motor disorders in this population occur from the very beginning in life preventing normal experience in human locomotion that extends beyond their former or actual capabilities. The findings indicate that the visual sensitivity to point-light human gait is not substantially affected by observers' early restrictions in locomotion, and therefore, walking experience does not appear to be a necessary prerequisite for the visual analysis of BM (Pavlova et al. 2003). These findings challenge a tight link between the ability for BM production and perception, though there are also some data indicating that motor training can improve visual BM processing (e.g., Casile and Giese 2006; Bidet-Ildei et al. 2010; Christensen et al. 2011). It is also reported that hemiplegic patients are compromised in recognition of point-light gestures portrayed by the affected contralesional arm (Serino et al. 2009). However, these data appear to be difficult to interpret in favor of the common code between BM perception and production because they leave room for several alternative possibilities: 1) motor impairment affects action recognition, 2) brain lesions affect understanding of actions of the contralesional arm, and 3) both brain lesions and motor impairment jointly affect action comprehension.
Adolescents with PVL exhibit difficulties in perception and understanding of others' actions depicted in a series of snapshots in a comic-strip fashion (event arrangement [EA] task), and the extent of periventricular lesions over the right temporal region serves as the best predictor of the severity of this impairment (Pavlova et al. 2008). Bearing in mind that a wealth of brain imaging and neuropsychological data point to the right temporal lobe as an important contributor to both visual BM processing and social cognition (e.g., Allison et al. 2000; Adolphs 2003; Puce and Perrett 2003; see also section Neural Network Dedicated to Biological Motion Processing above), one can assume that temporoparietal PVL disrupts subcortico-cortical and corticocortical connections with the right temporal cortex and lead to difficulties in perception and understanding of social properties such as intentions, desires, and dispositions of others. In PVL patients, deficits in the visual sensitivity to BM and performance on EA task are closely related. This indicates that the neural system for analyzing BM might be functionally integrated with the development of social cognition.
In sum, the data of PVL patients shed light on the role of brain connectivity for visual processing of BM. Disturbances in structural brain connectivity caused by periventricular brain damage lead not only to reduced visual sensitivity to BM but also to alterations in functional brain activity. Furthermore, the findings suggest that PVL patients are impaired not only on BM processing but also on some aspects of social cognition. Future research should clarify whether PVL patients have deficiencies in body language reading, that is, in visual perception of social properties (emotions, intentions, and expectations) conveyed through BM.
Visual Processing of Biological Motion in Autism
Individuals with autistic spectrum disorders (ASD; prevalence is close to 6 per 1000; Newschaffer et al. 2007) is an essential population for understanding the relationship between visual BM processing and deficits in social cognition. Autism is a life-long, presumably, genetic disorder characterized by impairments in several aspects of social cognitive behavior, among which are (1) avoiding direct eye contact and specific pattern of eye movements during social interaction and (2) deficiencies in reciprocal social interactions and in social perception, in particular, at nonverbal communication level. It has been proposed that ASD individuals experience difficulties in social cognition, and this impairment is associated with deficient BM processing (Dakin and Frith 2005; Pavlova 2005).
Up to the present date, however, the findings are controversial. BM perception (discrimination between unmasked point-light activities such as throwing, jumping, or kicking and phase-scrambled configurations) is compromised in 8- to 10-year-old autistic children, and the visual sensitivity to BM correlates with the severity of disorder (Blake et al. 2003). In accordance with this, 5- to 12-year-old autistic children are reported to have impairments both in discrimination of unmasked point-light BM from scrambled displays and in detection of camouflaged BM, though visual sensitivity to BM is unrelated to performance on coherent motion and form-from-motion tasks (Annaz et al. 2010). While typically developing children aged 7–11 years exhibit enhanced visual sensitivity to slightly camouflaged and noncamouflaged point-light humans as compared with point-light inanimate objects (tractors), adults with ASD do not (Kaiser, Delmolino, et al. 2010). Even 2-year-olds with autism are impaired on a preferential looking task with point-light BM (Klin and Jones 2008; Klin et al. 2009). Autistic toddlers do not exhibit any preference for upright-oriented point-light human figures performing “peek-a-boo” or “pat-a-cat” games over the same figures inverted in the image plane, whereas their typically developing peers and developmentally delayed children without autistic traits prefer upright-oriented BM (Klin et al. 2009). Visual information about BM in this study is accompanied by a soundtrack, which is played backward during the presentation of an inverted human figure. Although combination of BM with sound increases ecological validity of the stimuli, visual preference for upright-oriented BM may be potentially confounded with influence of auditory information. This work is of particular importance for our understanding of autism because for the first time it shows impairments in preference to BM in autism early in life.
Autistic adolescents with impaired high-level symbolic processing can reliably differentiate point-light human actions from similar moving configurations of inanimate objects (Moore et al. 1997). Autistic adolescents, however, are impaired on detection of the direction of a point-light walker embedded in a coherent motion mask (Koldewyn et al. 2010), although they perform similar to typically developing controls on a static coherent form task and on a coherent motion task. Most intriguing, performance on a BM task correlates to autism traits as measured by the autistic diagnostic observation schedule (ADOS), whereas coherent motion deficits are unrelated to the ADOS scores. This study has far reaching implications for our understanding of autistic impairments in social cognition because it provides evidence in favor of the hypothesis that poor social cognition is associated with deficient BM processing.
Adults with ASD are reported to be equally able to detect the direction of a camouflaged point-light walker as healthy controls (Murphy et al. 2009). This report should be taken with caution because translational component of motion (which is decisive for the facing/direction determination of a canonical walker) might have resulted in the lack of difference in sensitivity. However, adults with ASD also have no difficulties in detecting the direction of a camouflaged walker moving as if on a treadmill (Saygin et al. 2010). During performance of a direction discrimination task, young autistic male adults do not exhibit behavioral deficits, but they have lower fMRI activation over the temporal cortex as compared with healthy controls (Herrington et al. 2007). On the other hand, adolescents and young adults with ASD need longer to respond to unmasked human locomotion despite previous experience with such stimuli in an fMRI scan (Freitag et al. 2008). Autistic adults are not only less accurate in emotion identification in point-light and full-light BM, but this deficit is related to elevated thresholds on a coherent motion task (Atkinson 2009). It remains unclear whether autistic individuals exhibit deficient processing of BM itself or they have problems with segmentation in dynamic noise that requires additional attentional resources.
Although young children, adolescents, and adults with ASD appear to identify unmasked human actions in point-light BM displays, they apparently have deficits in revealing information about emotions from point-light actions (Moore et al. 1997; Hubert et al. 2007; Parron et al. 2008; Atkinson 2009). Because of methodological issues such as differently created point-light displays, unmasked or camouflaged targets, different types of masks, stimulus duration, tasks (recognition of point-light actions, detection of a walker, facing determination), IQ level of autistic and control participants, and severity of ASD, it is still unclear whether visual BM processing, processing of social properties (emotions, intentions, and dispositions) solely, or both are deficient in autism.
Functional MRI activation in autistic individuals is decreased in the right STS in response to unexpected gaze directions of others (Pelphrey, Morris, and McCarthy, 2005). Functional MRI in young autistic adults also points to decreased activation in the parietotemporal network including the right temporal cortex during point-light BM processing, though this activation is also lower in response to a control scrambled display (Freitag et al. 2008). Pioneering fMRI work in autistic children aged 4–17 years shows that, as compared with typically developing children and unaffected siblings, autistic individuals exhibit reduced fMRI response to point-light BM in several brain areas including the right pSTS and bilateral fusiform gyri (Kaiser, Hudac, et al. 2010).
High-functioning autistic patients have difficulties not only on some BM tasks but also in interpreting Heider-and-Simmel-like animations (e.g., Abell et al. 2000; Klin and Jones 2006; Boraston et al. 2007), and in this population, the extrastriate regions show reduced functional connectivity to the STS (Castelli et al. 2002). High-functioning autistic adults are impaired in recognition of particular emotions (sadness) in Heider-and-Simmel animations, and this deficit is linked to severity of impairment in reciprocal social interaction as assessed by the ADOS (Boraston et al. 2007). Most recent MEG findings reveal dynamic properties of the cortical network engaged in perception of social interaction in Heider-and-Simmel animations that might be impaired in autistic individuals (Pavlova, Guerreschi, Lutzenberger, and Krägeloh-Mann 2010; Pavlova, Guerreschi, Lutzenberger, Sokolov, et al. 2010).
Recent progress in brain imaging indicates that ASD may be largely associated with structural and functional brain abnormalities (McAlonan et al. 2005; Waiter et al. 2005; Courchesne et al. 2007; Assaf et al. 2010; Cheng et al. 2010). However, there is a lack of consensus in respect to specific structural brain abnormalities accompanying social cognition deficits in ASD. Reduction in gray matter is repeatedly reported in the temporal cortex, the region that is important for social cognition (Zilbovicius et al. 2006). It appears extremely unlikely that deficiencies in social cognition can be explained by abnormalities in a single brain region. Diffusion tensor imaging (DTI) in autism indicates disrupted white matter connections between regions implicated in social cognition including the right temporal cortex (Barnea-Goraly et al. 2004; Noriuchi et al. 2010; Jou et al. 2011). It becomes apparent that deficits in visual social perception in autism are associated with the right STS, and abnormalities are characterized by 1) decreased gray matter volume, 2) reduced structural and functional connectivity, and 3) abnormal functional activation during performance of social cognition tasks. Bridging the gap between appropriate psychophysical methodologies and analysis of functional brain activity will lead to better understanding the nature of deficits in BM processing and social impairments in autism.
Biological Motion Processing in WS, DS and FXS
Williams syndrome (WS), Down syndrome (DS) and Fragile X syndrome (FXS) are genetic conditions with distinct profiles of social cognitive impairments. Because of my overarching hypothesis that intact BM processing may be considered a fundamental basis for preserved social cognition, it is of particular interest to analyze whether BM processing is intact in these disorders.
WS is a relatively rare genetic condition (estimated to occur in 1 per 20 000 live births with an assumed equal sex ratio; Bellugi et al. 1999) resulting from a homozygous submicroscopic deletion on chromosome 7q11.23. The syndrome is associated with mild to moderate mental retardation and uneven cognitive profile with disproportionately severe visual–spatial deficits and spared language abilities. WS individuals show reduced thalamic and occipital lobe gray matter volumes and reduced gray matter density in regions involved in the visual–spatial system, whereas gray matter volume and density in several areas implicated in emotion and face processing (the amygdala, orbital and medial prefrontal cortices, cerebellum, anterior cingulate, fusiform gyrus, and superior temporal gyrus) are preserved or even enlarged (Reiss et al. 2004; Chiang et al. 2007; Golarai et al. 2010; see also Meyer-Lindenberg et al. 2005). Notably, DTI provides evidence for alterations in white matter tracts formations and brain connectivity in WS (Marenco et al. 2007). Individuals with WS exhibit specific features such as “elfin” facial appearance, connective tissue malformation, cardiovascular problems, and reduced overall brain volume. They commonly possess a hypersocial personality profile, with enhanced emotionality and face processing (Tager-Flusberg et al. 2003).
WS children aged 9–18 years and adults are unimpaired on visual BM processing. They can recognize point-light actions (jumping, slipping on a banana) and possess intact ability for direction detection (facing left or right) of a point-light walker moving as if on a treadmill and embedded in a static or dynamic mask (Jordan et al. 2002; Reiss et al. 2005). At the same time, children and adults with WS are severely impaired on a 2D form-from-motion task. To this day, there is a lack of consensus in regard to the ability of WS individuals to detect coherent motion. There are contradictory data indicating that this ability is either impaired (Atkinson et al. 2006) or preserved (Nakamura et al. 2002; Reiss et al. 2005). This is an important issue because in the light of deficient ability for other kinds of motion, the selective sparing of BM processing would point to its special status. The possibility for selective sparing is consistent with the previous neuropsychological and brain imaging findings indicating that BM processing engages a specialized neural network that differs from processing of other moving stimuli (Pavlova, Sokolov, et al. 2006). An MEG case study in a single WS patient aged 20 years indicates that the peak amplitude and latency of the evoked neuromagnetic response to point-light BM over the right occipitotemporal cortex do not substantially differ (are within 2 standard deviations) from healthy controls (Hirai, Nakamura, et al. 2009).
Intact BM and face processing in WS individuals is of interest in the light of evidence for deficient BM processing in patients with congenital prosopagnosia, a selective impairment in face recognition (Lange et al. 2009). Because both BM and face processing are essential for daily-life social cognition, this dissociation appears remarkable.
DS is a set of cognitive and physical conditions that result from having an extra copy of chromosome 21 (prevalence about 1 per 1000; Bishop et al., 1997). Cognitive and motor development is delayed in individuals with DS. In accordance with popular wisdom that DS individuals are socially approachable, recent data show that adults with DS exhibit a tendency to judge emotional face expressions more positively than controls (Hippolyte et al. 2008). Data on BM perception in DS are sparse. While DS children aged 8–15 years can reliably differentiate between point-light BM and moving objects, they are significantly less accurate in their judgments than typically developing controls aged 4–8 years (Virji-Babul et al. 2006). DS individuals are able to reliably (above chance) recognize emotions (such as happiness, anger, and fear) from point-light dance but fail in recognition of sadness. Again performance of DS children is lower than performance of controls. Interestingly, on an emotion attribution task with faces, adults with DS also exhibit difficulties with sad face expressions (Hippolyte et al. 2009). Finally, DS children are unable to differentiate between typical and atypical (represented by children with cerebral palsy and DS) point-light gaits (Virji-Babul et al. 2006). In general, DS individuals can reveal social information from BM, but this ability is compromised and modulated by the emotional context of actions. It remains unclear whether and, if so, how brain networks underpinning BM and social properties revealed through body motion are impaired in DS individuals.
FXS or Martin-Bell syndrome is a rare genetic condition (1:4000 in males and 1:8000 in females, Crawford et al. 2001) that is associated with the expansion of a single trinucleotide gene sequence (CGG) on the X chromosome and yields a failure to express the FMR1 protein, which is required for normal neural development. This results in a spectrum of specific physical, intellectual, emotional, and behavioral symptoms, which range from severe to mild in their manifestation. FXS is characterized by social anxiety, including gaze aversion, prolonged time to commence social interaction, and difficulties in forming peer relationships.
Social anxiety in individuals with FXS is related to deficits in face encoding (Holsen et al. 2008). Females with FXS are compromised on a point-light BM task requiring detection of a point-light walker embedded into a simultaneous mask (Kéri and Benedek 2010). This impairment is related to higher depression, but not anxiety, scores. However, this outcome is questioned by the finding that also visual perception of point-light mechanical motion (a rotating shape) is impaired in this population.
Overall, visual perception of BM appears to be compromised in genetic conditions but only if manifestation of disorder is characterized by daily-life difficulties in social cognition, establishing social contacts, and low social competence. WS individuals exhibit intact or even enhanced social skills, and they are reported to be unimpaired on BM tasks. In contrast, FXS patients with difficulties in social interaction appear to be compromised on BM tasks. Individuals with DS who might experience moderate problems in social cognition exhibit also moderate difficulties in BM perception. Future research should shed light on functional brain mechanisms associated with impairments in social perception through BM and body language reading. Such research should also help to better understand proper functioning of the social brain circuitry in typical development.
Sex Differences in Biological Motion Processing and Body Language Reading
Is the social brain sex specific? This is an open question. According to widespread popular beliefs about female superiority in social cognition, there are some indications for sex impact on BM processing in common marmosets (C. jacchus): Females but not males exhibit curiosity to point-light BM (Brown et al. 2010). Gender affects visual priming of camouflaged point-light BM (Bidet-Ildei et al. 2010). Early MEG response to point-light BM over the right (but not left) temporal cortex is greater in females (Pavlova et al. 2011). Gender effects, however, are not evident in the neural circuitry underpinning visual processing of social interaction represented by Heider-and-Simmel-like animations, but rather in the regions engaged in perceptual decision making: The MEG gamma response over the left prefrontal cortex peaks earlier in females (Pavlova, Guerreschi, Lutzenberger, Sokolov, et al. 2010). For expressive full-light (neck to knees or ankles) BM video clips, superiority of females in body language reading (Blanck et al. 1981) and greater fMRI brain activations in males for male threatening versus neutral body displays (Kret et al. 2011) are reported. The first study on body language reading that makes use of point-light BM displays representing knocking at a door with different emotional expressions indicates that gender effects are modulated by emotional content of actions: Males surpass in recognition accuracy of happy actions, whereas females tend to excel in recognition of hostile angry knocking (Sokolov et al. 2011).
There is emerging evidence of early sexual dimorphism of the brain (Cahill 2006; Vasileiadis et al. 2009). Investigation of sex differences would shed light on neurodevelopmental disorders characterized by deficient social cognition. It is known that males are more commonly affected by ASD than females, with a ratio of about 4:1 (Newschaffer et al. 2007). Females, however, are affected much more severely, and therefore, in high-functioning autistic individuals, this ratio is even much higher. The lack of studies in females with ASD calls for a thorough investigation of their profile. The other important issue for future research is sex differences in visual social cognition in survivors of premature birth: Males are at a 14–20% higher risk of premature birth (Melamed et al. 2010) and of its complications in the brain and cognition. Among live births, males have higher DS rates than females with an overall sex ratio of 1.28 (Bishop et al. 1997) and higher FXS rates with a ratio of 2 (Crawford et al. 2001). Gender effects in BM processing and body language reading in these genetic conditions are largely unknown. Clarification of gender impact on BM processing, body language reading, and underlying brain networks would provide novel insights into understanding of gender vulnerability to neuropsychiatric and neurodevelopmental impairments in social cognition.
One of the most prominent conclusions of this analysis is that individuals who exhibit deficits in visual BM processing are also compromised on daily-life social perception. This raises a question of whether performance on BM tasks may serve a hallmark of deficits in visual social cognition. Overall, the findings highlight the role of structural and functional brain connectivity for proper functioning of the neural circuitry involved in BM processing and visual social cognition that share topographically and dynamically overlapping neural networks. Future research should be directed at uncovering sex differences in BM processing, social cognition, and body language reading in typical and atypical development. Although sex differences represent a rather delicate topic, underestimation or exaggeration of possible effects can retard progress in the field.
Else Kröner-Fresenius-Foundation (Grants P63/2008 and P2010_92), the Werner Reichardt Center for Integrative Neuroscience funded by the German Research Foundation (pool project 2009–24), and the Reinhold-Beitlich Foundation to M.P.
This work is devoted to the memory of my mentor and friend Prof. John C. Baird, who passed away on June 8, 2011. I thankfully acknowledge my collaborators, in particular, Inge Krägeloh-Mann, Christoph Braun, Alexander N. Sokolov, Martin Staudt, and Werner Lutzenberger, and my former and present students Christel Bidet-Ildei, Michelle Guerreschi, Gunnar Erz, Samuel Krüger, Fabio Marconato, and Arseny A. Sokolov for contributions to work analyzed and described here. I appreciate the patients, family members, and care-providers for their kind cooperation. I am grateful to my husband and my son for their encouraging support. Conflict of Interest : None declared.