Biological Motion Perception, Brain Responses, and Schizotypal Personality Disorder

Importance  Exploration of the ability to process socially relevant events portrayed by biological motion and to identify underlying neuronal processes can provide clues for understanding the pathophysiology of psychosis. Individuals with schizotypal personality disorder (SPD) have pervasive interpersonal deficits and odd behaviors. An understanding of the neural mechanisms involved in the perception of biological motion and the relation of activity to clinical symptoms in those mechanisms is needed.

Objective  To investigate the specificity of brain regions responsive to biological motion perception in individuals with SPD compared with healthy control individuals.

Design, Setting, and Participants  Twenty-one patients diagnosed as having SPD and 38 age-, sex-, and IQ-matched controls underwent event-related functional magnetic resonance imaging. The SPD group completed the Scale for the Assessment of Positive Symptoms, the Scale for the Assessment of Negative Symptoms, and the Schizotypal Personality Questionnaire for assessment of symptom severity. During scanning, all participants were required to discriminate biological from scrambled sequences of point-light animations. Data were collected from September 21. 2011, to July 13, 2013, and analyzed from March to May 2015.

Main Outcomes and Measures  Blood oxygenation level–dependent signals during event-related scanning and symptom severity in the SPD group.

Results  The 21 individuals with SPD (16 men and 5 women) and 38 controls (29 men and 9 women) had a mean (SD) age, 22.8 (3.8) vs 22.2 (2.5) years and a mean (SD) IQ, 115.00 (12.55) vs 120.24 )7.68). Brain imaging revealed the presence of neuronal activation specific to biological motion within the posterior superior temporal sulcus. However, the individuals with SPD exhibited regions of neural responsiveness within brain regions forming the reward network, which consisted of the dorsal striatum and bilateral superior medial frontal cortex (all t ≥ 2.99, P of clusters <.002). The individuals with SPD also exhibited reduced activation in the anterior and middle cingulate cortices and the lingual and superior occipital gyri, which are brain areas responsive to biological motion perception and executive control of perception (all t ≥ 3.29, P of clusters <.001). In addition, significant correlations between the hyperdopaminergic clinical symptoms and enhanced neuronal activation in the caudate nucleus and frontal cortex were observed in the SPD group (all r ≥ 0.52, P < .02).

Conclusions and Relevance  Individuals with SPD display heightened activation in the neural circuitry involved in reward and decision making when viewing biological motion stimuli in addition to a positive correlation between increased blood oxygenation level–dependent signal responses related to biological motions and clinical symptoms. These findings suggest that enhanced responses arise within the reward network in individuals with SPD and are possibly related to the peculiar ways that individuals with SPD behave in social contexts.

The saying “actions speak louder than words” can be construed to mean that human body movements say more about a person’s intentions than facial expressions or words.¹ Posture and body movements convey varied, subtle messages about the person, and the perception of these social messages plays a large role in social communication.^2-4 This perceptual ability, termed perception of biological motion, is exemplified by performance when viewing a simple sequence of points of light depicting motion of the body and limbs of a human actor. The individual viewing these point-light (PL) animations may perceive the movements of the dots to be intentional and even emotional.⁵ Point-light animation sequences have been used to explore the visual processing of biological motion in healthy adults, children with normal development, and people with brain damage.⁴ Moreover, PL animations have been used with brain imaging techniques to identify cortical regions activated when people view biological motion animations. The posterior superior temporal sulcus (PSTS) is believed to integrate motion information from the dorsal system and object information from the ventral visual pathway system.^6,7 In addition, regions in the middle temporal gyrus that are sensitive to motion—the ventral lateral occipital cortex, lingual gyrus at the cuneus border, and cerebellum—are also thought to be involved in biological motion perception^8-11 and to work in tandem with the PSTS to form what has been termed a social cognition network.¹²

Given this perspective, people diagnosed as having schizophrenia show deficits in the perception of biological motion accompanied by abnormal activation levels in the PSTS. In addition to the core psychotic symptoms, such as hallucinations, delusions, and disorganized thoughts, schizophrenia is characterized by impoverished social skills.¹³ Some investigators believe that hallucinations and delusions experienced by individuals with schizophrenia may constitute fantasy social interactions that replace genuine social stimulation as a consequence of their self-imposed social isolation.¹⁴ This conjecture may be related to the intriguing finding that people with schizophrenia have unusually high false-alarm rates when undergoing testing with PL animations, some of which portray unstructured dot motions. Individuals with schizophrenia respond as if these noise sequences portrayed biological activity, and brain responses to those nonbiological animations are unusually large within the PSTS of those patients.¹⁵ These findings and their associated conjectures prompted our interest in studying individuals diagnosed as having schizotypal personality disorder (SPD).

Schizotypal personality disorder shares genetic, clinical, and cognitive abnormalities with schizophrenia,^16-19 but SPD differs in that associated cognitive functions are relatively spared. Therefore, individuals with SPD have less severe clinical symptoms and social deterioration.²⁰ Individuals with SPD seem to immerse themselves in mental play by seeking eccentric and odd interests, sometimes focusing on perceptual aberrations and/or constructing referential ideas to which other individuals do not pay attention.²¹ Thus, current diagnosis of SPD tends to focus on cognitive-perceptual criteria, that is, ideas of reference, odd beliefs, and perceptual disturbance, and on oddness criteria, that is, odd behavior, odd speech and/or thought processes, and restricted affect.²² As a consequence, individuals with SPD have poor real-world social adjustment^23,24 despite having relatively intact mental and social cognition. These characteristics of SPD and its relation to patent schizophrenia motivated our study.

We used an event-related functional magnetic resonance imaging (fMRI) technique to measure patterns of brain activation during the performance of a biological motion task in individuals with SPD and healthy control individuals. Given the conflicting evidence concerning social perception in SPD^25,26 and the previous research on the atypical tendencies in mind perception, such as increased attributions,²⁷ we hypothesized that individuals with SPD would show atypical neural responses within brain areas activated during biological motion perception. Thus, during the fMRI scans, we required controls and individuals with SPD to perform an easy discrimination task that presented animations portraying biological motions and scrambled motions, with the aim of identifying the blood oxygenation level–dependent (BOLD) signals associated with the perception of biological motion. The task we used was not designed to uncover deficits in the perception of biological motion; the task did not strain the participant’s ability to perceive biological motion but ensured that the person’s attention remained focused on the PL animations during the extended period of brain imaging.

Box Section Ref ID

Twenty-one individuals with SPD and 38 controls underwent fMRI scanning. We found no significant differences in the demographic data between the SPD and control groups (P > .05; Table 1). Data were collected from September 21, 2011, to July 13, 2013. All participants provided written informed consent for the experimental protocol, which was approved by the institutional review board of Seoul National University. Additional information about our participants is available in the eMethods and eFigure 1 in the Supplement.

We used the Scale for the Assessment of Positive Symptoms (SAPS),²⁸ the Scale for the Assessment of Negative Symptoms (SANS),²⁹ and the Schizotypal Personality Questionnaire³⁰ to assess symptoms clinically and the Korean–Wechsler Adult Intelligence Scale to measure IQ.³¹ These measures are described in further detail in the eMethods in the Supplement.

We used 24 different PL animation biological motion sequences depicting various human actions, including running, jumping, kicking, and throwing, and scrambled versions of those actions⁹ (Figure 1). After each action presentation, the participant pressed 1 of 2 buttons to indicate whether the animation depicted human activity or a meaningless motion sequence; responses were made using an fMRI-compatible response pad. Additional information is given in the eMethods in the Supplement.

Data were analyzed from March through May 2015. Details regarding the imaging data acquisition and preprocessing are provided in the eMethods in the Supplement. First, we applied the 1-sample t test in Statistical Parametric Mapping software (version 8; MathWorks, Inc) to analyze data from each individual. To identify the main effects of biological motion vs the scrambled condition across both groups, contrast maps of each of the 2 conditions (biological vs scrambled motion) were created (P < .05, familywise error). Next, we applied a random-effects analysis in which the outputs from the individual-level analysis were entered into a general linear model to determine visual stimuli condition–related activation differences between the group-mean maps using 2-sample t tests. The significance level was set at the threshold P < .005 with a cluster of more than 20 voxels to achieve a reasonable compromise between type I and II errors.³²

We extracted β values from clusters with 6-mm spheres around the peak-activated voxels that showed significant stimuli-related differences (biological vs scrambled motion) from each participant’s data using MarsBaR toolbox (http://marsbar.sourceforge.net/). The particular sphere size was based on a precedent established in earlier studies.³³ Pearson correlation coefficients were then calculated using SPSS software (version 20.0; SPSS, Inc) to examine the association of those β values for the SAPS and SANS scores with the percentage of change in the activation of the BOLD signal.

The control group consisted of 29 men and 9 women (mean [SD] age, 22.2 [2.5] years; mean [SD] estimated IQ, 120.24 [7.68]), and the SPD group included 16 men and 5 women (age, 22.8 [3.8] years; estimated IQ, 115.00 [12.55]). The control and SPD groups scored 96.8% and 95.7% accuracy, respectively, on the behavioral task (t = 0.70; P = .49). The absence of group differences was expected and, indeed, was the aim of the task design for this event-related procedure. In other studies of patients with schizophrenia, deficits in the discrimination of biological motion sequences were evident only when animations were rendered difficult to discern by the addition of masking noise or by the use of more subtle scrambling procedures.^15,34

Stimulus-selective clusters of voxels activated during PL biological motion were identified by contrasts between the biological and scrambled motion conditions using a 1-sample within-groups t test. This contrast detected extensive, intense activation differences in a number of voxel clusters localized within the temporal, occipital, and parietal cortices across combined participants. These clusters included voxels corresponding to the bilateral PSTS which, as in previous studies,^7-9 exhibited greater activation to biological motion than to scrambled motion (Figure 2) (Montreal Neurological Institute [MNI] coordinates for peak activations were −60, −58, and 12 for the left side and 56, −54, and 6 for the right side; t = 7.6 and t = 8.2, respectively). The brain regions exhibiting significant activations (biological contrasted with scrambled) are described in the eResults, eFigure 2, and eTable 1 in the Supplement.

Although the combined group data showed robust neural responses in the PSTS associated with viewing biological motion sequences, separate examination of areas associated with biological motion activation in the 2 groups disclosed different patterns of activation (Table 2). For example, activation in the anterior cingulate cortex was limited to a small region in the right hemisphere in the SPD group, whereas activation in the anterior and middle cingulate cortices was bilaterally present in the control group (number of voxels in each significant cluster [k], 68; P < .001). Moreover, the lingual gyrus and superior occipital gyrus, areas generally thought to be involved in the perception of biological motion, yielded lower activation in the SPD compared with the control group (k = 45 and k = 37, respectively; P < .001 for both). By contrast, the SPD group showed greater activation in the dorsal striatum, including the putamen and caudate nucleus, compared with the control group (k = 81; P < .001). In addition, the SPD group showed enhanced activation to biological motion in the bilateral superior medial frontal cortex and right middle temporal gyrus (k = 23 and k = 21; respectively; P = .002) (Table 2, Figure 3, and eFigure 3 in the Supplement).

To explore potential correlates of biological motion perception and SPD clinical characteristics, we performed correlation analyses between the β estimates of brain activation differing between the groups and clinical measurements within the SPD group only. A significant positive correlation was found between the activation in the caudate nucleus during the biological motion perception task and the SAPS Thought Disorder subscore (r = 0.61; P = .003) for individuals with SPD. Activation in the superior medial frontal cortex was positively correlated with the SAPS Hallucination subscore (r = 0.52; P = .02), SAPS Thought Disorder subscore (r = 0.65; P = .002), and total score (r = 0.55; P = .01). Activation in the anterior cingulate cortex was positively correlated with the SAPS Thought Disorder subscore (r = 0.50; P = .02) and total score (r = 0.54; P = .01) and with the SANS Flat Affect (r = 0.46; P = .03) and Attentional Disturbance (r = 0.47; P = .03) subscores. Finally, we found a positive correlation between activation in the putamen and the SANS Attentional Disturbance subscore (r = 0.48; P = .03) (eTable 2 in the Supplement).

Perception of the activities and thus the intentions of other people constitute important social skills. A useful way to study those skills uses PL animations that isolate the kinematic information portraying human activity.¹² Previous work using PL animations documents the existence of brain areas that constitute a network of processing nodes putatively involved in the perception of biological motion.^7,9 Moreover, behavioral and brain imaging studies in patients with schizophrenia reveal reduced activation in some of those nodes, most notably in the PSTS.¹⁵ Therefore, we investigated whether this biological motion network might differ in individuals with SPD. Those individuals constitute 1 region of the schizophrenia spectrum and share some but not all of the behavioral symptoms characteristic of full-blown schizophrenia. Our inquiry found some surprising results that, in hindsight, may be related to the neural bases of some of the characteristics that define SPD.

As a reminder, when we compared individuals classified as having SPD with controls while they viewed PL animations that visually portrayed biological motion, we observed the following. First, no significant activation differences were found between groups within the PSTS, an area thought to be a linchpin within the brain network involved in the perception of biological motion.^7,35 Second, heightened activation was detected in the striatum, including the putamen and caudate nucleus, bilateral superior medial frontal cortex, and right middle temporal gyrus. Third, reduced activation was detected in the lingual gyrus and superior occipital gyrus, areas that are normally responsive during biological motion perception. Finally, we found a positive correlation between increased BOLD responses evoked by biological animations and clinical symptoms. What do these findings tell us about the possible behavioral, emotional, and cognitive concomitants of those patterns of activation in individuals with SPD?

The array of brain areas exhibiting enhanced biological motion activation in people diagnosed as having SPD has been identified in other studies as consisting of neural circuitry involved in reward and decision making.^36-40 Therefore, might the unusual BOLD responses in the dorsal striatum arise from overactive attentional modulation, a schizotypal-specific cognitive feature? In an imaging study,⁴¹ participants with SPD showed increased BOLD signals within the frontal-striatal-thalamic circuitry, which includes the dorsolateral prefrontal cortex, striatum (caudate and putamen), and thalamic nucleus, when the region failed to ignore unimportant stimuli. The authors of that study proposed that the altered neural activations among individuals with SPD might reflect their inappropriate attention to (and thus failure to ignore) unimportant stimuli, perhaps contributing to their thought disorder. People with SPD tend to exhibit behaviors that imply odd ideas or magical beliefs.⁴² For instance, individuals with high levels of schizotypal features are more likely to assign agency and mentation to dead people, robots, animals, or trees, attributions never made by healthy people.²⁷ Perhaps hyperactivation to biological PL animations represents another expression of this bizarre mode of thinking. However, why would hyperactivation in individuals with SPD not have an equal effect on scrambled PL animations, the contrasting condition for identifying selective biological motion activation? We can think of several possible reasons. First, the social stimulus may give rise to significant neural responses across the dopaminergic regions.^43,44 A recent study⁴⁵ using a procedural learning task found increased BOLD signals across the dopaminergic regions, including the striatum, insula, and frontal regions, in response to patterned relative to random movements. In that study, these relatively strong responses were positively associated with the likelihood of a psychosis-prone personality. In other words, meaningful stimuli appear to trigger higher-than-normal levels of activation in the dorsal striatum in people with subthreshold proneness to psychosis.

How might this abnormally high level of activity within the dorsal striatum express itself in thought or behavior? Our study does not include measures that answer that question directly; however, based on other studies that implicate the prefrontal cortex and dorsal striatum in social action^43,44,46 and the reward of social action,⁴⁷ we can offer a conjecture about the possible consequences of hyperactivity specific to biological activity. Perhaps when observing biological motion, individuals with SPD recruit brain regions involved in the reward circuitry in addition to the typical regions activated in healthy individuals. Striatal neurons are activated during rewarded activity but not during unrewarded activity.^48,49 Thus, for individuals with SPD, whose symptoms include peculiar semantic processing and a limited capacity to form social interactions,^21,50 these PL stimuli may provoke an abnormal sense of animacy and mentation and consequently may imbue those simple stimuli with reward value. Our present results do not necessitate that conclusion, but the conjecture is certainly testable in future work by actual measurement of experiences of liking or desiring with BOLD activation levels during viewing of biological motion sequences.

We should consider those brain areas that show reduced activation to PL animations in individuals with SPD, including the anterior and middle cingulate cortices, lingual gyrus, and superior occipital gyrus. Previous studies^8,51,52 have shown that the lingual gyrus is more specific to higher-level motion processing, such as speed or direction of motion, which differs from simple motion processing in the middle temporal gyrus and social meaning processing in the superior temporal sulcus region. In addition, the superior occipital gyrus has been suggested to be involved in the observation of meaningful movements,⁵³ and the anterior midcingulate cortex has been postulated to be a crucial area for the executive control of perception, especially for error detection and monitoring.⁵⁴ Perhaps the individuals with SPD used minimal sources to perceive biological motions. Although the possibility of deficits in top-down processing is intriguing, a future study should focus on these processes in the biological motion perception of individuals with SPD.

Finally, we should consider the strong associations between brain activation and clinical measurements in individuals with SPD. Enhanced activation in the dopaminergic regions and SAPS scores are closely linked. These findings agree with those of reports that demonstrated that the SAPS total score and the SAPS Thought Disorder subscore can be used to evaluate the subcortical hyperdopaminergic symptoms of psychosis.⁵⁵ A prior study⁵⁶ described the effects of prefrontal dysfunction or dysregulation of subcortical dopaminergic function and provided feasible neuronal evidence for the positive (eg, paranoia) and negative (eg, anhedonia) symptoms of schizophrenia spectrums, especially in social situations. Therefore, enhanced neural activity in the dopaminergic region may reflect these spontaneous bizarre concerns about others or external stimuli in individuals with psychoticlike experiences.⁵⁷ The correlations between brain activation and symptoms of SPD in the present study allow us to speculate possible neural mechanisms for the schizotypy-specific eccentric ideation.^23,28 In addition, we are reminded of the social deafferentation hypothesis, which states that social withdrawal may produce abnormal cognitive attributions in individuals with an unfulfilled desire for social contact. Perhaps abnormal neural activations seen in people within the spectrum of schizophrenia disorder promote misinterpretation of cues guiding social interactions.

We have obtained suggestive evidence that the brain in patients diagnosed as having SPD exhibits unusual patterns of neural activity in response to viewing biological motion sequences. One unusual pattern is the higher-than-normal activity within the brain network commonly believed to form a reward circuit. Another unusual pattern is the weaker-than-normal activity within regions ordinarily involved in executive function, such as error monitoring. In addition, enhanced activations in dopaminergic regions are implicated in clinical symptoms of individuals with SPD. These findings may help to characterize the interaction between disturbed reward circuitry and the bizarre ways of perceiving and experiencing social stimuli observed in individuals with SPD, whose condition tends to isolate them from everyday social interaction.

Corresponding Author: Jun Soo Kwon, MD, PhD, Department of Psychiatry, Seoul National University College of Medicine, 101 Daehak-no, Chongno-gu, Seoul 03080, Republic of Korea (kwonjs@snu.ac.kr).

Submitted for Publication: August 4, 2015; final revision received October 18, 2015; accepted November 19, 2015.

Published Online: January 20, 2016. doi:10.1001/jamapsychiatry.2015.2985.

Author Contributions: Drs Hur and Kwon had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Hur, Blake, J. Kim, Kwon.

Acquisition, analysis, or interpretation of data: Hur, Cho, J. Kim, S.-Y. Kim, Choi, Kang, Kwon.

Drafting of the manuscript: Hur, Blake, Cho, Kang, Kwon.

Critical revision of the manuscript for important intellectual content: Hur, J. Kim, S.-Y. Kim, Choi, Kwon.

Statistical analysis: Hur, S.-Y. Kim.

Obtained funding: Kwon.

Administrative, technical, or material support: J. Kim, S.-Y. Kim, Choi, Kwon.

Study supervision: Blake, Kang, Kwon.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Basic Science Research Program through the National Research Foundation of Korea, which is funded by grant 2013R1A2A1A03071089 from the Ministry of Science, ICT, and Future Planning, and by grant R31-10089 from the World Class University program sponsored by the Korea Science and Engineering Foundation with funds from the Ministry of Education, Science and Technology.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Baxter P. Rogers, PhD, Vanderbilt University School of Medicine, assisted in the analysis of the original study. He received no compensation for this role. We thank the participants involved in this study.