Characterizing and Improving Face-processing
Skills in Children with Autism

Prof Pawan Sinha
Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology
Cambridge, MA 02139
USA

Summary

Introduction: An integral component of a child's mental health is the ability to correctly interpret visual information about other people in the environment. Deficiencies in these skills can have devastating consequences. Indeed, one of the most marked correlates of autism is 'an impairment in the use of multiple nonverbal behaviors such as eye-to-eye gaze, facial expression, body posture, and gestures to regulate social interaction' (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition). In order to understand the causes and consequences of such visual impairments and to design rehabilitation procedures for mitigating them, it is imperative that we precisely characterize the visual skills of the affected children. With this motivation, we have undertaken a research project that seeks to improve the lives of children with autism.
Goal: We shall focus our studies on the face-perception skills of children suffering from autistic disorders. Preliminary experimental data suggest that such children may have impaired face processing abilities. This jeopardizes their interpretation of non-verbal cues, compromises effective social interactions and may lead to emotional disturbance and depression. Our goal is to experimentally characterize the nature and extent of face-perception impairments, to determine the processing deficiencies that may cause the observed impairments, and to refine and test VisTA (Visual Training and Assessment), a novel tool to help the children overcome their visual impairments.
Impact: This project will be the first to provide us with a comprehensive set of data regarding impairments in facial processing for autistic children. The data will be invaluable for understanding the precise nature of the deficits, evaluating the effectiveness of VisTA, and guiding the design of other treatment interventions.

Autism is a pervasive neuro-developmental disorder associated with marked deficits in a child's social and communicative abilities. Impairments in social skills are typically accompanied by emotional disturbances and severe problems in adjustment, making autism one of the most disruptive disabilities for a child's life. Its prevalence is estimated to be at least 1:1000 children (Centers for Disease Control and Prevention, 1997). More recent estimates have suggested an incidence as high as 1:500, making autism even more common than Down's syndrome. Given its prevalence and its devastating consequences on the lives of the children affected and, indeed, their families, our general lack of understanding of autism's causes and options for ameliorative interventions, is of grave concern.

We propose to focus on a hallmark deficit of autism - difficulties in social interactions. Specifically, we shall investigate whether this deficit is caused, at least in part, due to problems in high-level visual processing. Most person to person interactions are contingent on an exchange and interpretation of subtle facial cues. It is, therefore, conceivable that an inability to effectively process facial signals would manifest itself as impairments in social skills. Investigating this possibility is important since it has major implications for the design of treatment programs.

There is already some experimental evidence in support of this idea. Researchers such as Hobson (1986a, 1986b), Tantam et al. (1989), Gepner et al. (1994), Loveland et al. (1995), Celani et al. (1999), and Dawson et al. (2002), have reported that children with autism exhibit impairments relative to normal controls at tasks involving interpretation of facial emotions in images. While these findings have not gone uncontested (Ozonoff et al. (1990); Davies et al. (1994)), they serve as an excellent starting point for our investigations into the visual correlates of the social impairments in autism. Our work has three specific goals.
1. To identify core deficits in face processing by autistic children.
2. To determine the processing abnormalities that might cause the visual deficits.
3. To design training routines for improving visual skills.

We next describe the motivation for these goals and our planned approach for achieving them.

Goal 1: To identify core deficits in face processing

Establishing the basic face-processing deficits associated with autism is important from many perspectives; it is a pre-requisite for diagnostic purposes, for guiding the course of treatment and for constraining the search for causes. Past work, mentioned above, has made a beginning in this direction. However, the focus has been almost exclusively on the perception of basic facial expressions. Several other face perception tasks are largely unstudied. Four of the most important ones needed for normal social interactions are:
1. face localization in complex scenes (the first step in analyzing facial signals),
2. classification of complex expressions (for instance pride, shame and affection, common in everyday interactions),
3. assessment of attentional locus (a pre-requisite for shared attention tasks), and
4. facial identification across changes in viewpoint (as might occur during a normal interaction).

We shall test two groups of children on these four tasks. The first group will include autistic children and the second will be a control set comprising normal, age-matched children. Children will be diagnosed for autism at the collaborating hospitals using DSM-IV criteria and their syndrome severity will be assessed using the Childhood Autism Rating Scale (CARS) (Schopler et al., 1988). The children will participate individually in psychophysical experiments designed to probe performance on each face perception task. Psychophysical sessions will be augmented with ERP recordings whenever feasible to permit more comprehensive comparisons between experimental and control groups.

Goal 2. To determine the processing abnormalities underlying deficits in face perception

By identifying the core deficits in face processing, as described above, we expect to improve our understanding of the causes of at least some of the social impairments exhibited by autistic children. To understand the genesis of the face processing deficits themselves, it is imperative that we explore another level of causation - what underlying visual processing deficits might lead to the observed problems in face perception? We shall address this question in a hypothesis driven fashion. Our hypothesis is derived from an observation that several high-functioning autistic individuals have made in describing their sensory experience (VanDalen, 1995; Williams, 1999). The world, to them, appears fragmented and lacks the 'built-in form of coherence' (Frith, 1989). We shall explore whether 'fragmented perception', or alternatively, the lack of configural processing, is a plausible causal factor for the observed deficits in face perception tasks. Our approach will involve psychophysical studies with autistic and normal children. The facial stimuli used in these experiments will be transformed so as to selectively influence piecemeal or configural processing (transformations will include Gaussian blurring, vertical inversion and image part permutations that preserve the features but not their configuration). By measuring the influence of these transformations on performance and face-specific ERP signals, we shall be able to infer the nature of visual processing deficits that might underlie face perception impairments in autism.

Goal 3. To design training routines for improving visual skills

Goals 1 and 2 seek to uncover the causes of social skill impairments in autistic children. Goal 3 is meant to apply this knowledge towards the design of methods that can alleviate the deficits. To the best of our knowledge, there are currently no interventions for improving visual processing by children with autism. We propose a novel approach, VisTA (Visual Training and Assessment) to address this need. VisTA has several of the characteristics considered desirable for autism related interventions (McConnell, 2002) such as the ability to involve both children with autism and their normally developing peers, portability for use in various settings throughout the day and easy monitoring of progress.

VisTA is based on a technique for image presentation we have recently developed called RISE (Random Image Structure Evolution) (Sadr and Sinha, 2001, 2003; Pollak and Sinha, 2002). RISE enables the presentation of images as time-series while carefully controlling potentially confounding influences from low-level image parameters. In a RISE sequence, images gradually evolve and become progressively more recognizable. An observer's point of perceptual object onset in such a sequence serves as a quantifiable marker for object perception proficiency. We have used RISE for assessing object perception skills of children with different developmental histories (Pollak and Sinha, 2002), for studying phenomena such as object priming and for identifying object-specific neural responses (Sadr and Sinha, 2001, 2003). VisTA extends the usage of RISE to the domain of visual training. VisTA presents time series showing face images evolving from randomness. The task of the observer is to try to determine what the evolving image depicts (say, a particular expression or a specific individual) as soon as possible after the beginning of the sequence. Since the image is very degraded near sequence onset, the observer is forced to use overall configural information rather than relying on piecemeal cues. Our pilot experiments with normal observers suggest that they find the task engaging, in the nature of a game, and show significant improvements in their ability to recognize objects using partial information after a few training sessions. We believe that the same procedure when conducted with children with autism, using appropriate facial stimuli, can mitigate their tendency to rely on fragmentary information and improve their ability to use overall facial configuration. This would translate into better performance at many of the socially relevant face-perception tasks such as recognition of subtle facial expressions.

Logistics: We plan to conduct our experimental studies at two institutions that draw a significant population of children with autism.

1. Children's Hospital in Boston
The Children's Hospital in Boston is a 300-bed comprehensive center for pediatric health care. The Hospital's Department of Psychiatry evaluates and treats children with a wide range of emotional and behavioral disorders, including autism. It typically records more than 15,000 outpatient visits and 400 inpatient admissions each year, making it one of the largest pediatric psychiatric services in New England. The Department develops interventions for various complex psychiatric problems using cognitive, behavioral, group, and biological methods.

2. The Open Door Center for Autistic Children in New Delhi, India
This center is operated by Action For Autism (AFA), a founding member of the World Autism Organization. AFA provides counseling, assessment and educational programs. Children diagnosed with autism are referred to AFA by prominent national hospitals such as the All India Institute of Medical Sciences and the National Institute of Mental Health and Neurosciences. AFA received 123 referrals in 2001. The OpenDoor Center is the first specialized school for children with autism in South Asia.

Our decision to work with AFA besides the Children's Hospital is driven by two considerations. First, it stands out in terms of the dedication of its members and their willingness to collaborate on such a project. The Director Merry Barua in particular has impressed me greatly with her deep passion for helping children with autism and the very welcoming attitude towards this undertaking. Second, it serves to bring much-needed attention to the problem of autism in India. India is estimated to have more than 2 million children with autism and yet public awareness of this disorder and governmental resources directed towards it are insignificant. We hope that reports of this work for the scientific and lay audiences will play an effective role in improving the state of knowledge in India regarding autism and possible interventions. This is necessary for improving the quality of life for the many children with autism in India who may otherwise be treated as social outcasts.

In summary, we propose to characterize face-processing deficits that may contribute to the social impairments observed in children with autism. We also seek to improve the children's face-perception skills through the use of a novel training technique, VisTA. The prospect of helping the many children who have to overcome challenges imposed on them by autism is a very compelling one and we look forward to getting the project underway by the beginning of year 2004.

 

References

Ceani, G., Battacchi, M. W., and Arcidiacono, L. (1999). The understanding of the emotional meaning of facial expressions in people with autism. Journal of Autism and Developmental Disorders, 29, 57-66.

Davies, S., Bishop, D., Manstead, A. S. R., and Tantam, D. (1994). Face perception in children with autism and asperger's syndrome. Journal of Child Psychology and Psychiatry, 35, 1033-1057.

Dawson, G., Carver, L., Meltzoff, A. N., Panagiotides, H., McPartland, J., and Webb, S. J. (2002). Neural correlates of face and object recognition in young children with autism spectrum disorder, developmental delay and typical development. Child Development, 73(3), pg. 700-717.

Frith, U. (1989). Autism: Explaining the enigma. Oxford: Basil Blackwell.

Gepner, B., de Schonen, S., and Buttin, C. (1994). Face processing in young autistic children. Infant Behavior and Development, 17, 661.

Hobson, R. P. (1986a). The autistic child's appraisal of expressions of emotion. Journal of Child Psychology and Psychiatry, 27, 321-342.

Hobson, R. P. (1986b). The autistic child's appraisal of expressions of emotion: A further study. Journal of Child Psychology and Psychiatry, 27, 671-680.

Loveland, K. A., Tunali-Kotoski, B., Chen, R., Brelsford, K. A., Ortegon, J., and Pearson, D. A. (1995). Intermodal perception of affect in persons with autism or Down syndrome. Development and Psychopathology, 9, 579-593.

McConnell, S. R. (2002). Interventions to facilitate social interaction for young children with autism: Review of available research and recommendations for educational intervention and future research. Journal of Autism and Developmental Disorders, 32(5), 351-372.

Ozonoff, S., Pennington, B. F., and Rogers, S. J. (1990). Are there emotion perception deficits in young autistic children? Journal of Child Psychology and Psychiatry, 31, 343-361.

Pollak, S. and Sinha, P. (2002). Effects of early experience on children's recognition of facial displays of emotion. Developmental Psychology, 38, 784-791.

Sadr, J. and Sinha, P. (2001). Exploring object perception with Random Image Structure Evolution. MIT Artificial Intelligence Laboratory memo.

Sadr, J. and Sinha, P. (2003). Exploring object perception with Random Image Structure Evolution. Cognitive Science (in press).

Schopler, E., Reichler, R. J., and Renner, B. R. (1988). The childhood autism rating scale. Los Angeles, CA: Western Psychological Services.

Tantam, D., Monaghan, L., Nicholson, H., and Stirling, J. (1989). Autistic children's ability to interpret faces: a research note. Journal of Child Psychology and Psychiatry, 30, 623-630.

VanDalen, J. G. T. (1995). Autism from within: Looking through the eyes of a mildly afflicted autistic person. Link, 17, 11-16.

Williams, D. (1999). Like colour to the blind. Jessica Kingsley Publishers.

Researcher goes face to face with autism
June 2, 2003

Pawan Sinha, assistant professor of visual neuroscience at MIT, wonders if autistic individuals simply don't see what others see, focusing on details and losing sight of the whole.

Sinha is one of three promising young scientists in the United States who recently received a four-year fellowship from the John Merck Scholars Program in the biology of developmental disabilities in children. The Merck Scholars receive $300,000 apiece over four years to support their research. Sinha plans to use the Merck fellowship for his work on characterizing and improving face processing skills in children with autism and those who have suffered extended periods of visual deprivation.

Autism is a complex developmental disability that typically appears during the first three years of life. The result of a neurological disorder, autism is four times more prevalent in boys than girls and knows no racial, ethnic or social boundaries. As many as 1.5 million Americans are believed to have some form of autism, according to the Autism Society of America. That number is growing at a rate of 10-17 percent per year.

Autistic children typically have trouble maintaining eye contact and reading social cues such as facial expression, body posture and gestures. Sinha is exploring whether autistic children have impaired face perception skills and is developing a methodology called VisTA (Visual Training and Assessment) to help them refine those skills.

Through collaborations with Children's Hospital in Boston and the OpenDoor Center for Autistic Children in New Delhi, India, the first specialized school for children with autism in South Asia, Sinha plans to administer tests to several autistic individuals over the coming two years. The tests will simultaneously determine whether the children have trouble seeing faces and help them bring disparate visual elements together into a comprehensible whole.

"We are focusing on a hallmark deficit of autism--difficulties in social interactions," Sinha said. "Most person-to-person interactions require an exchange and interpretation of subtle facial cues. It is conceivable that an inability to effectively process facial signals would manifest itself as an impairment in social skills. We want to find out what visual processing deficits might lead to problems in face perception."

If visual processing by the brain is the problem, this would have significant implications for the design of treatment programs. VisTA is intended to be one such program. "We believe that this procedure, when conducted with children with autism, might mitigate their tendency to rely on fragmentary information and improve their ability to use overall facial configuration," Sinha said.

Sinha believes that conducting some of these studies in India as well as in the United States can accomplish important objectives. "We hope to bring some much-needed attention to the problem of autism in India, where the number of affected children is estimated to be more than 2 million. Public awareness and government resources directed toward it are almost nonexistent," he said. "We hope to improve the quality of life for many children with autism in India who may otherwise be treated as social outcasts. And we hope that these studies, whether in the United States or India, will eventually help mitigate the toll that this disorder exacts on the lives of many children and their families all over the world."