2002 Grants Funded (CAN)

Studies of a serotonin transporter (5HTT) mutation associates with Asperger's syndrome

Matthew Beckman, University of Alabama at Birmingham (Young Investigator)
The search for genes associated with autism has been the focus of much attention by researchers in the past few years. Recently this search has led to the discovery of a mutation in the human serotonin transporter (SERT) coding sequence from patients with the following neuropsychiatric triad: Asperger's syndrome, obsessive-compulsive disorder, and anorexia. The function of a cloned serotonin transporter with this mutation (SERT-AS) has been characterized using [3H]5Ht uptake assays, binding assays employing a high affinity SERT ligand, ?-CIT, and immunocytochemical methods. When compared to the wild type SERT, mutant SERT-AS exhibits a roughly 1.5 fold increase in both maximum transporter velocity (Vmax) and binding sites (Bmax) and an increase in affinity (lower Kd) for ?-CIT, a high affinity ligand. Serotonin transporters have been shown to be highly regulated by various intracellular signal transduction pathways and direct phosphorylation of the transporter. Treatment of cells expressing the wild-type SERT with a nitric oxide donor S-Nitroso-N-acetylpenicillamine (SNAP) suggest a role for nitric oxide in the regulation of wild-type SERT, but the mutant SERT-AS is refractory to this stimulus. This work aims at characterizing the differences between the wild type SERT and the mutant SERT-AS using pharmacological, biochemical, and electrophysiological methods. Further, we hope to delineate the mechanism by which nitric oxide upregulates wild-type SERT with the goal of understanding why SERT-AS is constitutively upregulated. These studies offer the possibility for new insight into the pathogenesis and treatment of autism.

fMRI of perceptual gating and response selction of high-functioning autism and in autism siblings

Matthew Belmonte, Cambridge University (Young Investigator)
We propose to apply fMRI to ten-to-fourteen-year-old children with high-functioning autism or Asperger syndrome during performance of a visual attention task designed to factor out demands of perceptual selection from demands associated with response selection. Subjects will perform a two-stimulus colour-orientation discrimination in which different properties of each stimulus determine the status of the stimulus pair as a target or a non-target. In the stimulus in which orientation is the relevant property, colour will have to be suppressed, and in the stimulus in which colour is the relevant property, orientation will have to be suppressed. In a control task, both stimuli in each pair will be identical, and thus no suppression will be required. While both the experimental and the control tasks place demands on response selection, only the experimental task requires perceptual selection within each stimulus pair. We anticipate that perceptual selection will activate intraparietal sulcus, and that response selection will activate dorsolateral prefrontal cortex. We will compare results in autism to results in normal controls and also in autism siblings. The latter comparison is of interest since autism siblings share some of the cognitive characteristics of autism probands, yet they do not, in general, have autism. Siblings may thus offer a glimpse of the neurophysiological roots of the autistic syndrome, unobscured by later developmental sequelae.

Functional characterization of the SPCH1 gene

Joseph Buxbaum, Mount Sinai School of Medicine (Pilot Research)
A relationship has been postulated between certain language disorders and autism. The strongest evidence for this relationship is the identification of families with chromosomal abnormalities leading to language disorders in some family members and autism in others. Further support for this has been the genetic linkage of both language disorders and autism to a region of chromosome 7. Recently, a gene that underlies the language abnormality in a family with severe language disorder has been identified as foxp2. This gene is a transcription factor that has not been studied in detail. We and others are examining this gene as a candidate gene for autism, and we are also studying the expression of this gene during development. Given the importance of this gene, we here propose to disrupt this gene in mice ("knockout") and thereby study its role in development of the central nervous system and identify genes that are regulated by foxp2.

The Pre-school Social Communication Assessment Measure (PSCAM): Sensitivity to developmental change and early intervention efforts.

Tony Charman, Institute of Child Health (Psychometrics Award)
Significant progress has been made in the identification of autism spectrum disorder (ASD) in the pre-school period. Age on entry to a program and an emphasis on developing communication skills are important elements of successful early intervention programs. Although several screening (CHAT, M-CHAT, STAT) and diagnostic instruments (ADI-R, ADOS-G) have good clinical utility, they are unlikely to be sensitive measures of intervention effectiveness. In part this is because their primary aim is diagnostic. Further, they focus on measuring impairments in social and communicative capacities. As most intervention programs have increasing social communicative competencies as a key aim, novel instruments that measure competencies in these early social communication abilities are required. Non-verbal social communication abilities including joint attention, imitation, play and reciprocal interactions are "precursors" to later language and social-communicative competence in both children with ASD and typical development. As part of screening and early intervention research studies over the past 8 years, we have assessed 50 children with ASD under the age of 24 months. We developed an interactive assessment measure that focuses on early social and communicative capacities (Pre-school Social Communication Assessment Measure; PSCAM). Systematic analysis of the PSCAM videotapes from the two extant datasets will allow us to evaluate if the PSCAM is a sensitive index of developmental change in pre-school children with ASD and a sensitive index of outcome in pre-school children with ASD enrolled in an ongoing early intervention study. Funded In partnership with Repligen Corporation

Identification of genes involved in autism

Russell Ferland, Beth Israel Deaconess Medical Center (Young Investigator)
Autism is a neurological, developmental disorder that is characterized by abnormal communication and social interaction, and impaired cognitive development. Studies have demonstrated a clear genetic link with autism and have shown that it affects approximately 5:10000 people (Fombonne, 1999). Many groups have undertaken whole-genome screens in an effort to identify susceptibility loci in idiopathic autistic individuals living in the USA, with moderate success. Our laboratory has had great success in identify genes associated with neurological disorders through the use of families having consanguineous marriages from the Middle East. We are now applying this approach utilizing families with first-cousin marriages that have individuals affected with autism. We currently are examining families with first-cousin marriages in which approximately 25% of the children from these marriages have autism. We have conducted a genome-wide screen of these families and have demonstrated potential linkage to several loci in the genome. Our evidence suggests potential candidate regions on chromosomes 1, 3, 6, 8, 10, and 18 by multipoint analysis (multipoint LOD scores > 1). Currently, we are attempting to test these candidate intervals through homozygosity mapping, marker analysis, and candidate gene identification in these chromosomal regions. Furthermore, our laboratory has initiated a collaboration with a school for autistic children in Kuwait and will obtain DNA from other consanguineous families with multiple autistic children, allowing us to expand our study of recessive autism genes. Overall, these studies aim to identify genes associated with autism, which potentially will lead to improved diagnosis and treatment.

Studies of the relationship between autism and colonization of the intestinal tract by neurotoxigneic clostridial species

Sydney M. Finegold, M.D., University of California, Los Angeles, VA Medical Center (Bridge Grant)
This group has been investigating the possibility that neurotoxin-producing bacteria could cause or significantly contribute to autistic symptoms in a subset of children with autism. Previous qualitative and quantitative analysis of stools indicated that there were differences between the stool flora of children with autism compared to that of control children. They are closing examining the Clostridium species isolated from the stools of children with autism. Using polymerase chain reaction (PCR) they will specifically test these isolates for the production of tetanus neurotoxin.

Processing visual information in Autism. A pilot study

Dr. William Good, Smith-Kettlewell Eye Institute (Bridge Grant)
Studies on autism have shown abnormalities in a number of vision-based functions such as visual attention, face recognition and visual search (e.g. Plaisted et.al., 1998, Jolliffe T, Baron-Cohen S, 1997, O'Riordan et.al., 2001). The results of these studies have been attributed to different high-level cognitive impairments. However, there is no data showing whether the low level processing of visual information is working normally in autism.

Most visual information is transmitted through the striate pathway (the pathway that originates in the retina and courses to the primary visual cortex, or V1, passing through the lateral geniculate nuclei). At V1 in the visual cortex, visual information is distributed into two main streams, dorsal and ventral. It is important to understand whether defects in V1 processing cause abnormalities in higher based visual functions (attention, face recognition, visual search).

The present study is designed to investigate low level and middle-level visual processing in order to study how local information about individual features is combined into more complex visual patterns. A better understanding of all levels of visual functioning in the autistic population will promote further research and help to interpret previous studies.

To further develop several methods for reliable, quantitative, and reproducible measures of attention, and of implicit and explicit memory for individuals with autism

Barry Gordon, Johns Hopkins University School of Medicine (Psychometrics Award)
Attention and memory (learning) are often impaired in individuals with autism, particularly in lower-functioning individuals. Problems with generalized attention (alertness), directed attention, explicit learning, and implicit learning have all been identified. Normally, many different tasks are used to measure these functions. However, valid measures of these underlying functions can be obtained through performance on two relatively simple tasks that such individuals can perform: stimulus-response association learning (both visual-visual and auditory-visual) and the sequential reaction time task (which requires only touching a target or touching a target predictively). Determination of these underlying abilities is certain to be enhanced when standard response recording is augmented by observations of response speed, posture, head position, and gaze direction. Accordingly, these tasks will be used in the current proposal, tailored to each individual's level of functioning. Training/testing will be done on a specialized computer system (Foundations), which automatically assigns materials, runs the tasks, and collects all data in a relational database format. Video monitoring will be done concurrently with task performance. A total of 20 subjects with autism will be tested (10 low-functioning, 10 high-functioning). Testing will be repeated at least once, in different physical surroundings and with different examiners, to test reliability. Validity will be established by construct and by comparison within- and across-subjects and across methods (e.g., manual response vs. gaze direction). These measures and methods are expected to be useful in detailed assessments of the components of attention and memory in large-scale clinical trials. Funded In partnership with Repligen Corporation

Toward generation of an autism mouse model by making locus-specific duplications homologous to human chromosomes 15q11-q13 region

Yong-hui Jiang, Baylor College of Medicine (Young Investigator)
The numerous reports of autistic individuals with cytogenetic abnormalities of chromosome 15 have indicated that this region may harbor a susceptibility gene or genes for autistic disorder. The most common abnormalities are interstitial duplications and isodicentric chromosomes of exclusively maternal origin, suggesting that the gene or genes contributing to autism susceptibility may be subject to imprinting. Chromosome 15q11-q13 is an imprinted domain implicated in two neurological disorders, Prader-Willi syndrome (PWS) and Angelman syndrome (AS). Two maternally expressed genes, UBE3A and ATP10C, map within the region. UBE3A is the Angelman-causing gene and is expressed from the maternal allele in brain. We hypothesize that maternal duplication of 15q11-q13 may disrupt the normal expression of imprinted genes such as UBE3A and ATP10C and that abnormal expression of these two genes and/or others may contribute to autism susceptibility. We propose using the Cre-loxP system to prepare mutant mice with a duplication covering the entire 15q11-q13 region; a duplication covering the Ube3a gene and its possible regulatory elements; and a duplication covering the Atp10c gene. We will examine the expression pattern for Ube3a and do neuropathology, neurophysiology, and behavior testing on mutant mice to identify any impairment in morphology and social interaction. We believe there is a great likelihood that this will produce the first valid mouse model for autism. The mouse model will facilitate the genetic dissection of this trait and eventually lead to development of new and effective therapeutic interventions.

The development of perceptual integration in autism

Chantal Kemner, University Medical Center Utrecht (Pilot Research)
It is well known from clinical practice that subjects with autism are preoccupied with details, and they are usually found to excel in certain visuospatial tasks, which have in common that a tendency towards detail processing is advantageous. Extensive detail processing also seems to play a role in the social problems of autistics, such as face processing. Also, in autism structural as well as functional abnormalities are found in the occipital lobe. It is hypothesized that these findings indicate specific problems in early visual processing, viz. in visual perceptual integration. This is the ability to spatially integrate details of stimuli. An already funded study was started on perceptual integration in school age and adult autistic subjects. Two tasks are being used; one on texture segregation and one on face processing. During the tasks event-related brain potentials are measured, in order to get detailed information on the course of the perceptual integration process. However, there are indications that the abnormalities in autism are subjected to developmental influences. Therefore, it is warranted to study perceptual integration in autism as early as possible. In this project, the tasks will be presented to children with autism of about three years of age, who are recruited from a large screening study on early detection of autism in our department. The requested support will be used to accommodate the testing procedure for young children, to do the testing, and for analyzing the data.

Sensory experience, behavioral therapy, and neural plasticity: Implications for autism remediation

Michael Kilgard, Ph.D., University of Texas at Dallas (Pilot Research)
Autism is characterized by impairments in social interaction, disordered communication, repetitive behavior, and stereotyped interests. We propose that these symptoms lead to a form of isolation that disrupts the development of normal brain responses and in turn hinders the acquisition of critical language skills. Anatomical studies have established that even mild isolation during development can lead to significant weakening of cortical circuits. Although behavioral therapy and auditory training appear to improve language skills in autistic children. The neural basis of these improvements is not known. We have recently demonstrated that early sensory enrichment in rats leads to a profound increase in the sensitivity, frequency selectivity, response strength, and processing speed of neurons in primary auditory cortex. Similar brain plasticity may explain training-induced improvements of language function in autistic children. This proposal outlines experiments 1) to more completely characterize what specific aspects of an enriched environment strengthen cortical circuits, 2) to document how daily behavioral training improves cortical processing, 3) to establish the time course of developmental plasticity, and 4) to determine how dysfunction in the central cholinergic system affects auditory cortex maturation. Our preliminary findings indicate that during development cortical circuits are highly susceptible to degradation, and that targeted sensory enrichment can be extremely effective at restoring these circuits to normal performance. Results from the proposed experiments and others from the Autism Neural Plasticity Initiative will significantly influence the development of behavioral and pharmacological treatments for autism.

An investigation of somatosensory processing in autistic children using magneto-encephalography (MEG)

David McGonigle, University of California, San Francisco (Young Investigator)
The social and communicative handicaps present in autistic individuals have led a number of researchers to suggest that autistic individuals can be best described as suffering from a deficit in 'theory of mind' (e.g. Boucher, 1989). Others have focused on autistics' disorders of 'executive function' (Russel, 1998). An alternative hypothesis is that the higher-order cognitive deficits observed in autism are caused by disorders of basic sensory processing. Frith (1989) has proposed that the disorders of perception arising in autism may result from weak 'central coherence' (CC), or an inability to extract global features or context from stimuli. To examine sensory processing in autistics, we propose to investigate the neuromagnetic correlates of the processing of somatosensory stimuli in autistic children using magnetoencephalography (MEG), a non-invasive neuroimaging method. We predict that autistic individuals will have disrupted spatial patterning in their primary somatosensory cortex (SI), and will furthermore show differences in amplitude of their somatosensory magnetic mismatch response when compared to age-matched controls. By quantifying and localizing the neural substrates of this aspect of autism (disorders in sensory processing), we believe our work may suggest novel therapeutic approaches (e.g. Merzenich et al., 1996) to the problem of autism in children.

Using event-related potentials (ERP) to characterize auditory processing deficits in autism spectrum disorder and epilepsy

Elizabeth Pang, Ph.D., The Hospital for Sick Children Canada (Pilot Research)
Autism Spectrum Disorder (ASD) is a childhood neuropathology affecting language, social interaction, behavior and play. Landau-Kleffner Syndrome (LKS) is an acquired epileptic aphasia where normal children develop a devastating loss, or regression, of language. Two important similarities exist between ASD and LKS: 1) over one-third of the children with ASD also report a language regression, and 2) both ASD and LKS manifest auditory processing deficits. The question raised whether ASD and LKS are biologically distinct entities or whether LKS is a later manifestation of regression seen along the spectrum of ASD. Examination of these auditory processing deficits may shed light on the pathoaetiology of these disorders. Auditory processing in the cerebral cortex can be examined using event-related potentials (ERP), in particular, the N1. Our ERP Lab has recently published the developmental N1 norms. Our feasibility studies suggest that a specific and consistent N1 abnormality can be observed and appeared over one cerebral hemisphere in ASD and over both hemispheres in LKS. This raises the possibility that the N1 abnormality may be a biological marker for ASD and LKS and that the pattern of appearance of this abnormality may be an important diagnostic toll for differentiating these disorders. Objectives: (1) Verify whether the N1 abnormality is a biological marker for ASD and LKS. (2) Test whether the N1 abnormality can discriminate ASD and LKS. (3) Look for other consistent abnormalities in the N1 related to ASD. (4) Test whether more complex auditory stimulii are better markers for these disorders.

A pilot cross-cultural study of autism phenotypes in Puerto Rico

Jeremy Silverman, Mount Sinai School of Medicine (Pilot Research)
The Mount Sinai Autism Family Studies Research Center recently studied a very large number of primarily non-Hispanic affected sibling pair families for genetic and phenotypic studies. In one study we identified several autism related clinical features that show evidence of familiality (i.e. the level of clinical expression among affected members in the same family are more likely to be similar than between unrelated affected individuals). Indeed, to date, we have looked at one of these features (phrase speech delay) and found substantially increased evidence for genetic linkage on chromosome 2 among affected sibships who share this characteristic. In the present pilot project we will recruit affected sibling pair families from Puerto Rico, a distinct population from the families previously studied, to begin to examine whether similar patterns of familiality are evident in affected sibling pair families in this culture as well. If so, the familiality of such features are more likely to be explained by similar underlying genetic factors as opposed to cultural/environmental ones. Blood samples too will be collected for our genetic studies and shared with AGRE, under the terms of our ongoing collaboration, and to examine lead levels in Puerto Ricans with autism. Puerto Rico is especially valuable as a separate recruitment site because the genetics of Puerto Rico is largely distinct from the U.S. mainland population and relatively homogeneous. In addition, families are frequently very large and highly cooperative, yet understudied despite the fact that Puerto Rican people constitute a major minority group in the United States.

Provacative urine excretion of heavy metals using meso-2,3-dimercaptosuccinic acid (DMSA) in children with autism

Sarah E. Soden and Jennifer A. Lowry, Children's Mercy Hospitals & Clinics (Bridge Grant)
The investigators propose measurement of urinary mercury, lead, cadmium, arsenic, and aluminum prior to and during a 24 hour provocative urine excretion study; comparing children with autism to non-autistic age matched controls. For the provoked excretion, we will use a heavy metal chelator meso-2,3-dimercaptosuccinic acid (DMSA). Our goals are to document difference that may exist between the two groups, to direct follow-up study, and to collect normative data for provocative urinary excretion of these environmental toxins.

Cellular mechanisms controlling central oxytocinergic activity: Role in the pathophysiology of autism

Javier Stern, Wright State University (Pilot Research)
Autism, a behavioral syndrome characterized by impairments in socialization, communication and stereotyped behaviors, has been the subject of substantial investigation. While major advances have been made, no pharmacological treatment has yet been found to consistently improve the symptoms or course of the disease. Developing a more effective treatment for autism requires research on the underlying pathophysiology. Even though autism is recognized as a neurodevelopmental syndrome, its neurobiological basis is poorly understood. Growing evidence indicates that the neuropeptide oxytocin plays an important role in social behavior, cognition and motor stereotypes. Furthermore, an abnormal function of oxytocinergic systems has been proposed to be involved in the pathogenesis of autism. Little is known about the cellular and functional properties of central oxytocin neuronal circuits, and how abnormal changes in these properties might be involved in the pathophysiology of autism. Using the oxytocin knock out mouse as an animal model for autism, we aim to determine a) the cellular mechanisms controlling neuronal excitability in oxytocin neurons that innervate limbic
structures involved in behavioral and social functions, b) the cellular mechanisms by which oxytocin modulates neuronal activity in these areas, and c) the mechanisms underlying altered neuronal function in limbic areas in the oxytocin knock out mouse. This work will provide fundamental information on the cellular mechanisms regulating central oxytocin activity. Furthermore, it will lead to a better understanding of the cellular mechanisms that contribute to the pathophysiology of autism, providing a basis for the development of therapeutic strategies for the treatment of this disease.

Mechanisms underlying receptors delivery to synapses in early postnatal development

Takuya Takahashi, Cold Spring Harbor Laboratory (Young Investigator)
Large deficits in information processing such as learning and memory have been found in autistic patients. It is suggested that disruption of the glutamatergic neurons is responsible for these deficits. Long-term potentiation (LTP), the long lasting enhancement of synaptic strength induced by repetitive activation of glutamatergic excitatroy synapse, is believed to have important roles in learning and memory. One possible mechanism of LTP expression is the rapid delivery of the functional AMPA-type glutamate receptors to synapses from non-synaptic sites. Indeed, GFP-tagged AMPA receptor 1 (GluR1) is delivered to synapses after LTP-inducing stimuli. In contrast, GluR2 is continuously delivered to synapses and replaces existing synaptic receptors. Signaling mechanism of synaptic delivery of AMPA receptos is porrly understood. Interestingly, GluR4 is expressed exclusively in the first postnatal week and delivered to synapses by spontaneous activity. This suggests that GluR4 has a crucial role in information processing during early postnatal development. Considering the early onset of autism, it is important to understand the cellular and molecular mechanism of infomration processing during early postnatal development. We have characterized the important region for delivery in the carboxyl terminus of GluR4 and isolated Hsc70 as a binding protein of this region. We will analyze the functional role of Hsc70 on the delivery of GluR4.

To assess a total of 35 SNPs (single-nucleotide polymorphisms) in 6 to 10 positional candidate genes using a complete set of DNA from well-characterized families with autism spectrum disorders from the AGRE family collection of biomaterials.

Rudolph Tanzi, Massachusetts General Hospital (Contracted Research)
Autism and autism spectrum disorders are genetically complex neuropsychiatric disorders of early life. To date, several whole genome scans on independent samples were performed indicating the existence of numerous putative autism loci across different chromosomes. One of the most recent studies was performed by the Autism Genetic Resource Exchange (AGRE) Consortium and yielded suggestive linkage evidence on at least four chromosomes (i.e. chromosomes 5, 8, 19 and X; Lui et al., 2001). In addition, using different analytic methods and additional markers other investigators were able to show linkage to other loci in the same sample (e.g. Buxbaum et al., 2001). Our study aims at assessing a total of 35 SNPs (single-nucleotide polymorphisms) in 6 to 10 positional candidate genes in these regions in the AGRE dataset. Specifically, we will use high-throughput fluorescence-polarisation single-base extension (FP-SBE) detection for genotyping of 3 to 5 SNPs per gene, which will then be analyzed either individually using the family-based association test program (FBAT), or as multi-locus haplotypes using TRANSMIT. Both programs allow to estimate valid significance values even in the presence of known linkage - as is the case here, with positive results from the prior genome scans - while making full use of all the available genotype information. Our laboratory has extensive experience will all aspects of high-throughput SNP genotyping and analysis, which can be directly applied to the AGRE sample. These efforts to identify new genetic risk factors will, together with the contributions of the other AGRE Consortium members, help to unravel the genetics of this appalling disorder.

Analysis of candidate genes in autism

Enrique Villacres, University of Washington (Pilot Research)
Current evidence suggests that autism is a polygenic disorder with at least five or more genes causing the disease. We hypothesize that chromosomal translocations found in some autistic patients cause or contribute to autism. The genes disrupted by these translocations are candidate genes for autism. These genes may be major effect loci coincident with chromosomal regions identified by linkage analysis or minor effect loci not detectable by standard family-based methods. Point mutations or polymorphisms in the genes identified may contribute to autism in subjects with normal karyotype. We have performed molecular analysis of a balanced chromosomal translocation t(7;20) (q11.2; p11.2) in a pair of twins with autism. Specifically, we have identified a novel gene (AUTS2) at 7q11.2, that spans this breakpoint. Two recent searches for susceptibility loci in autism using genetic linkage methods yielded positive evidence for an autism locus on chromosome 7q (Barret et al., 1999; IMGSAC, 2001). AUTS2 is expressed in fetal brain, suggesting that alterations in this gene could result in abnormal development leading to autism. Preliminary evidence does not support allelic association with autism with only two SNPs studied. The goals of this proposal are to 1) identify the chromosome 20 gene/genes disrupted by t(7;20) in the monozygotic autistic twins; 2) identify genes disrupted by all the breakpoints in an autistic boy with a complex rearrangement t(1;7;21); and 3) screen translocation breakpoint genes for mutations in non-breakpoint subjects. Finding one or more genetic cause of autism will provide clues to the treatment of this disorder.

Pilot pediatric diagnostic study for autism and other developmental problems

Stuart Shanker, Ph.D. and Stanley Greenspan, M.D., York University (Contracted Research)
There is great interest in identifying children at risk of developing Autism Spectrum Disorder (ASD) at earlier ages in order to enhance the opportunity for early intervention as well as to understand the early stages of this disorder in the context of the search for neurobiological correlates and preventative strategies. Recent research suggest that an early sign of the developmental trajectory leading to ASD is a deficit that appears in infants around the age of 9 months in their capacity for reciprocal exchanges of emotional expressions. In these exchanges of affect, and impairment can also be observed in the detail and complexity of the facial expressions of emotion, as well as a discord between the infant's movements in response to her caregiver's vocalizations, facial expressions, and gestures. This study proposes to develop a systematic procedure for looking at the dynamic process of co-regulated facial expressions of affect and the rhythmicity of the infant's movements as a tool for identifying 9 month-old infants that are at risk of developing ASD.