The Puzzle of Autism

“For those of you that don’t know Ella, she doesn’t like to be touched, so every time you try to grab her she drops on the ground, fl aps her arms and starts to scream. On top of that, the more you try and get her to stay still, the more she twitches and the louder she becomes. She cannot stay still; she has to wander constantly.”

This mother’s introduction to her three-year-old, courtesy of the National Autistic Society, is not atypical among the half million families in the United States living with an autistic child.

One out of every 150 children born is diagnosed with the disorder; among boys, who are four times more likely to have autism, the number rises to one in 94. Currently, 1.5 million Americans are autistic.

Autism falls under a broader category known as Autism Spectrum Disorders (ASD), characterized by early onset developmental delay in communication and social interaction.

ASD manifests itself over a wide spectrum of severities, ranging from the less severe Asperger Syndrome, through Pervasive Developmental Disorder Not Otherwise Specified (PDD-NOS), all the way to full-scale autism. The disorder has a profound impact on a person’s capacity to understand social cues, express emotion, and form relationships.

Recently, rising numbers of diagnoses and increased awareness have shown the spotlight upon autism, with researchers and physicians across the nation working towards understanding and treating the complex disease.

The Autism Program at Yale

One of the leading clinical and research centers is the Autism Program at Yale, located in the Child Study Center. For over fifty years, the program has attracted premier clinicians and scientists across an interdisciplinary spectrum.

“The Autism Program at the Child Study Center focuses on bringing together the clinical world and world of science,” says the program’s director Dr. Ami Klin, Harris Associate Professor of Child Psychology and Psychiatry.

A National Institute of Health Autism Center of Excellence, the program encompasses clinical care, research, training, and advocacy. Teams of psychologists, psychiatrists, speech pathologists, geneticists, and others evaluate various aspects of a patient’s behavior, including communication, social interactions, and cognition.

The clinical program treats individuals of all ages, often retaining relationships with patients from infancy to adulthood. The center has also produced some of the foremost publications in the field, including the current leading textbook entitled Handbook of Autism and PDDs. Engaging in over 150 clinical training sessions annually, the training program caters to pre- and post-doctoral fellows, medical residents, graduate and medical students, and even undergraduates, with the popular seminar “Autism and Related Disorders” offering a practicum for students to work with autistic individuals in schools and private homes.

Moreover, the Autism Program at Yale advocates with legislative, educational, and scientific organizations around the world. “From genetic findings all the way to service and social policy, the Autism Program at Yale works to advance knowledge of causes and develop new treatments,” says Klin. “Multiple perspectives allow more comprehensive answers.”

A young boy with autism stacks cans. Repetitive and stereotyped behavior is one criteria of autism diagnosis.

Through the Eyes of the Autistic

Individuals with autism see the world differently – literally. In any environment, a person’s senses are constantly bombarded with a multitude of stimuli, and individuals ordinarily develop an intrinsic sense of what is important and what to ignore.

For example, in a social setting, people tend to focus on certain key aspects to gauge context and meaning and thus infer the appropriate reaction.

Autistic individuals, however, need explicit instructions on which cues are socially relevant to impose meaning on their contexts. The inability to naturally develop these traits poses significant challenges in everyday social situations.

What exactly are autistic individuals looking at when they see? To answer this question, Klin and colleague Warren Jones, a Research Associate at the Yale Child Study Center, developed a novel eye-tracking method. Over the past several years, they have conducted investigations in which autistic cases and non-autistic controls watch age-specific naturalistic social scenes.

During viewing, the eye-tracker quantifies, moment by moment, where the eyes of the watcher focus to create a “landscape of salience” – or comprehensive measurements of which environmental stimuli the participant considers important.

What Klin and Jones found was that these landscapes do not match up between the autistic cases and control subjects. When watching naturalistic social scenes, non-autistic individuals tend to focus on eyes, while autistic individuals are more apt to look at mouths, other body parts, or peripheral inanimate objects.

In a 2002 study, published in Archives of General Psychiatry, the Yale researchers found that compared to controls, individuals with autism focused twice as much on the mouth region, twice as much on the body region, and half as much on the eye region of dynamic faces.

According to Klin, the understanding of a social situation is incredibly dependent upon the ability to infer meaning and context from the language of the eyes. Autistic individuals, however, do not distinguish verbal speech from the physical act of producing that speech; that is, since the sounds of language emerge the mouth, autistic individuals look primarily at that region.

Data showing visual focus for non-autistic viewers (white crosses) and autistic viewers (black crosses) during a movie clip, measured with eye-tracking technology. Typical individuals tend to focus on eyes, while autistic individuals show preferences for mouths, other body parts, and inanimate objects.

The result is difficulties in perceiving others’ thoughts or emotions and thereby impaired ability to predict behavior. This raises the question, says Klin, of “what experience is one having when he or she experiences another being?” That question may be taken one step further: “Could someone watch a face but not see a person?”

The tendency to make eye contact is innate in normal development. Infants as young as three months tend to stare at the eyes, rather than the mouth, of an approaching person. This preference is not exclusive to humans, but common to many primates. Scientists have observed mutual gaze between the eyes of infant and mother monkeys when the infant take its first steps.

Therefore, this likely emerged as an evolutionary mechanism. Infants are fragile, requiring much support by their caregivers. Natural selection rewards the animal able to attract the attention of the caregiver, and, as Klin explains, “eye attraction creates a loop between one organism and another.”

Thus, it is believed that eye contact became almost an extension of physical body contact. In humans, the strongest and most persistent eye contact is exhibited in two relationships: between a mother and an infant, and between lovers. “Eyes are not only the window to the soul, but the window into the social mind and social brain,” says Klin.

Eye-tracking investigations revealed another distinct inclination of autistic individuals – the tendency to view the inanimate over the social. For example, during a romantic movie scene, an adult with autism focused on a light-switch in the background rather than on the kissing couple. While watching a clip of Barney, a toddler with autism gazed upon the objects on the shelves behind Barney and the children in the center stage.

An additional trait of normal development is the ability to follow non-verbal gestures of communication, such as pointing. Klin and Jones’ eye-tracking studies showed that this ability, present in children as young as 12 to 14 months, is undeveloped among the autistic.

The naturalistic scene was a video clip in which an actor points to a painting on a wall, among a group of several paintings, and then asks “Who did the painting?” Viewers normally follow a similar pathway: first, they follow the pointing finger to look at the correct picture on the wall; then, after hearing the question, they look at the individual being asked the question and then back at the person who asked it.

A person with autism does not initially follow the pointing gesture, but rather waits to look at the wall until the question is verbally expressed; then, he scans the various pictures on the wall, unable to tell which is being referenced. Upon later questioning, the same autistic individuals were able to define and explain what pointing gestures mean. Simple explicit knowledge of the definition, however, does not translate into instinctive practice.

Klin, Jones, and others are now using eye-tracking in conjunction with measures of electrical activity in the brain. Functional magnetic resonance imaging (fMRI) is used to localize regions of brain activity, and electroencephalography (EEG) measures electric activity with temporal resolution. That is, the fMRI answers the question of “where,” while EEG provides the “when.”

By gaining a better sense of what is happening in the brain when autistic individuals react to certain stimuli, “the hope is that we can detect vulnerability to autism before the symptoms emerge,” explains Klin. The tracking methods, which were first studied with adults, are now being used to track the gazes of children, toddlers, and even infants. Early detection enables earlier intervention and treatment.

When a Square is More Than a Square: The Social Attribution Task

In order to gauge the inability of autistic individuals to naturally impose social context upon visual cues, Klin studied the responses of cases and controls to a 1944 cartoon animation showing geometric figures moving as if they were social beings.

When asked to provide a narrative describing the cartoon – a test known as the Social Attribution Task (SAT) – nonautistic viewers ascribed verbs and concepts relating to social behavior to the geometric figures, such as “being a friend,” “protecting,” “trapping,” and “being shy.”

Moreover, the individuals showed increased brain activity in regions associated with face perception; that is, the same regions active when thinking about actual people were also active when viewing objects that simply behaved in social ways. As Klin describes, “we tend to anthropomorphize our world.”

The autistic viewers demonstrated significant shortcomings in perceiving these social relationships. The 2000 study, published in the Journal of Child Psychology and Psychiatry, revealed that, on average, autistic viewers recognized only one quarter of the social behaviors considered necessary to deducing the plot.

The study also showed that “a third of their attributions were irrelevant to the social plot,” and that they were “unable to derive psychologically based personality features from the shapes’ movements.”

In this 1944 cartoon animation, geometric blocks behave as social beings to relate a narrative. In the Social Attribution Task, autistic individuals show impaired ability to recognize the behaviors of the shapes as social.

Autistic viewers did not have difficulties, however, explaining the literal motions of the geometric blocks. Further demonstrating the specific social nature of this deficiency was a 2006 study by Klin and Jones that used a novel Physical Attribution Task (PAT) in addition to the SAT.

They found that autistic individuals were successful at constructing narratives relating to physical behaviors represented by geometric animations, such as when shapes moved in a motion resembling the launching of a rocket. In addition, the deficits in imposing social meaning did not correlate with deficits in either verbal IQ or level of metalinguistic skills.

These findings challenge the theory that autistic individuals have overall tendencies to process their environments in a “fragmented fashion,” focusing on the parts rather than the whole to provide meaning. Rather, this claim seems valid only with regard to specific domains – holding true, for instance, more often in the social world than in the physical realm.

No “One Gene” Cause

Siblings of autistic children are enrolled in the Autism Program at Yale before they are even born. The reason for this is the strong genetic component of autism; in fact, of all neuropsychiatric disorders, autism is believed to be the most highly heritable.

Studies of concordance rates between siblings and twins (which measure the proportion of pairs that share a trait) have provided evidence for this theory. In monozygotic (identical) twins, the concordance rate is 60 percent. This number shoots up to 90 percent for Autism Spectrum Disorder. The concordance rates for dizygotic (fraternal) twins and siblings is 5 to 10 percent. In contrast, an individual randomly selected from the population has a less than one percent chance of having the disorder.

Despite the high heritability, the etiology of autism has remained more of a mystery. In particular, attempts to locate a single gene or mutation implicated in autism have been unsuccessful. “The misconception used to be that autism has one clear, biological basis,” says Dr. Matthew State, the Harris Associate Professor of Child Psychiatry and Genetics.

However, lower concordance rates for fraternal twins and siblings compared to identical twins, patterns of inheritance revealing multiple genetic abnormalities in autistic patients, and a wide spectrum of phenotypes among those affected support a theory of a multi-genic basis of the disorder.

Current estimates state that at least two and as many as one hundred genetic mutations may contribute to the development of autism. Each genetic abnormality likely contributes a “small increment of risk” that, when combined, increase susceptibility to the disorder.

Why does each common mutation contribute only a small risk? The answer lies in natural selection. The chronic, early-onset nature of the disorder impairs the reproductive fitness of an allele, or form of a gene, carrying any appreciable risk. That is, natural selection tends to weed out high-risk alleles, so that any alleles common in early onset contribute only small effects.

“For disorders like diabetes, we likely overestimate how much risk those common alleles carry,” State explains. “With autism, it is likely that common alleles carry even smaller risks — meaning we need even larger samples to identify them.”

As a result, “We focused on rare genetics variants contributing to autism from the start,” said State, rather than on the common small-effect variants. Although the search for rare variants is more common today, the technique was relatively novel when the State lab began in 2000.

Searching for Susceptibility Genes

In the search for rare genetic variants, the State lab employs two main approaches: the study of inbred families that pass along a phenotype of Autism Spectrum Disorder, and the study of chromosomal abnormalities in affected individuals. Over the past several years, enhanced technologies and greater samples sizes have rapidly augmented the search for the susceptibility genes linked to the disorder.

In 2004, a study published in the American Journal of Human Genetics identified in a child with PDD-NOS an abnormality on chromosome 3 that disrupts a gene known as contactin 4 (CNTN4). This gene encodes for a neuronal cell adhesion molecule involved the growth and development of axons, which are extensions of nerve cells that transmit impulses.

Then, in the past year, the lab identified rare chromosomal abnormalities that disrupted three genes in the same family of molecules: one was CNTN4, and a second gene of interest was Contactin-associated protein-like 2 (CASPR2), located on chromosome 7. The State lab published its discovery in January 2008 in the American Journal of Human Genetics.

Remarkably, in the same journal, two other labs – one at Johns Hopkins School of Medicine and the other at UCLA – also reported association of CASPR2 with autism, using different methodologies. Moreover, all three papers overlapped with a study in the New England Journal of Medicine that linked the same CASPR2 gene to symptomatic focal epilepsy. CASPR2 is involved in the language development in the brain.

“Looking across all the findings, there is more agreement on CASPR2,” says State. “Still,” he continues, “we are a long way from understanding what exactly the contribution is and how it works.”

The lab is currently investigating the relevance of mutations in contactin and associated molecules to the development of autism, collaborating with researchers in neurobiology, neurosurgery, and molecular, cellular, and developmental biology.

First in the Family: The Role of Spontaneous Mutations

Although autism is highly heritable, the majority of cases show no family history of the disorder. How can this be explained? The hypothesis is copy number variations (CNVs), which are large-scale (100 kilobases and over) chromosomal structural variations, such as duplications and deletions. These variations are formed de novo, or spontaneously, in the parental germ line, meaning that the disorder will express itself in children even when neither parent is affected.

Studies have already linked de novo CNVs to various disorders, including Prader-Willi syndrome, Angelman syndrome, and spinal muscular atrophy. Single nucleotide polymorphism (SNP) chips and competitive genomic hybridization arrays (aCGH) are used for genotyping, producing data that enables the detection of duplications and deletions.

Funded by the Simons Foundation in New York and with grants from the National Institute of Child Health and Development and from the Autism Center of Excellence, the State lab will be leading a national genome-wide search for copy number variations in a large sample of simplex autism families.

In a simplex family, a single child is affected with the disorder, and both parents are unaffected. Yale is one of a dozen sites contributing patients to the investigation. This investigation will continue over the next several months, “by far the largest simplex collection ever evaluated for autism,” says State. A total of 2000 affected children and 6000 family members will be evaluated in a search for de novo common and rare variants associated with autism.

Collaboration is the Crux

As evidence mounts for the complex and multi-genic nature of the etiology of autism, clinical collaboration continues to be increasingly essential. “It is very unlikely that genes map one-to-one onto the clinical constructs,” says State.

“The genetics perspective is valuable in that mutations could increase susceptibility to developmental delays that are cognitive, social, or a combination – but all it does is say that genes set the stage…. They are not fate.”

The motif underlying all the recent research and advances indeed seems to be interdisciplinary efforts. “We go from the bedside, starting with autism research group at the Child Study Center, all the way through to biology to disentangle a complex and devastating disorder,” says State.

As our understanding of autism expands in scope and complexity, it is not surprising that we must draw upon knowledge and resources harnessed from a wide array of sources.

Research may be ongoing, but every finding brings us one step closer to providing answers and treatment for the 1.5 million and their families suffering every day.

ILANA YURKIEWICZ is a a junior in Calhoun College majoring in Molecular, Cellular,
and Developmental Biology.

The author thanks Dr. Ami Klin and Dr. Matthew State both for discussing their
own research and for providing general information about the Autism Program at

Yale Child Study Center:
Gupta, A.R., and State, M.W. (2007). Recent advances in the genetics of autism. Biol. Psychiatry 61, 429–437.
Klin, A. (2000). Attributing social meaning to ambiguous visual stimuli in higher functioning autism and Asperger syndrome: The Social Attribution Task. Journal of Child Psychology and Psychiatry, 41(7), 831-846.
Klin, A., Jones, W., Schultz, R.T., & Volkmar, F.R. (2003). The Enactive Mind – from actions to cognition: Lessons from autism. Philosophical Transactions of the Royal Society, Biological Sciences, 358, 345-360.