Show Summary Details
Page of

Environmental Exposures That Increase the Risk of Autism Spectrum Disorders 

Environmental Exposures That Increase the Risk of Autism Spectrum Disorders

Environmental Exposures That Increase the Risk of Autism Spectrum Disorders

Patricia M. Rodier

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2015. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy).

Subscriber: null; date: 16 August 2017

Points of Interest

  • More is known about environmental risk factors for autism than for other developmental disabilities.

  • The risk factors that have been established are all potent teratogens, which cause many different birth defects.

  • The relative risk for each of the exposures is extremely high.

  • The critical period for exposure for each environmental risk factor is during the first trimester.

  • Phenotypic features that accompany the autism outcome after exposure are shared by several exposed groups and are seen in idiopathic cases and genetic syndromes, as well.

Autism spectrum disorders are unique among the conditions described in the Diagnostic and Statistical Manual of the American Psychiatric Association (DSM), in that several environmental exposures have been determined to increase the risk of ASDs. Although the etiology of autism is known to have a strong genetic component (e.g., Bailey et al., 1995), these environmental factors offer other avenues by which the nature of ASDs can be investigated. For example, they offer scientists the opportunity to create animal models based on exposures to risk factors, just as genetic discoveries offer the opportunity to create models by manipulating animal genotypes.

The environmental factors identified thus far account for a very small proportion of autism cases, but that does not diminish the importance of their contribution to our understanding of ASDs. The purpose of this chapter is to summarize our present knowledge of the environmental factors that are known to increase autism risk.


Evidence and Relative Risk

The first environmental factor associated with autism was rubella infection (Chess et al., 1978; Chess & Fernandez, 1980). Because rubella exposure of the embryo has long been known to be a cause of birth defects including brain damage (Gregg, 1941; Ariens Kappers, 1957), it is hardly surprising that it might produce injuries that lead to autism. What is more remarkable is that a significantly elevated rate of autism was found using the very narrow criteria of Kanner (1943). It seems likely that the same population of rubella-exposed children, if examined with today’s much broader definition of ASDs, would have shown an effect that was even more striking. However, comparing the 18 autism cases out of 243 rubella-exposed cases (Chess et al., 1978) to modern rates for autistic disorder (16.8/10,000, from Chakrabarti & Fombonne, 2001) gives a relative risk greater than 40. This is an enormous effect.

As we consider each of the known risk factors for autism, it will become apparent that each has a relative risk that is much higher than those familiar from toxicology studies. For example, the National Institutes of Health (NIH) recognizes a relative risk of 2 to 4 for the effect of smoking in increasing the risk of heart disease, and this is considered a large effect. It is worthwhile to pause here and consider what makes the risk factors for ASDs’ relative risks dramatically higher. The effects of teratogens on development are different from the effects of toxic agents to which we are exposed over long periods. For example, in the case of rubella, postnatal exposure or exposure late in pregnancy cannot cause any of the permanent defects we associate with this virus. Instead, there is a brief window during development when exposure leads to injury. This is called the critical period, and it differs for different teratogens and different defects. Notice that Chess ascertained cases exposed to rubella in an institutionalized population. That is, all the cases were severely injured. Therefore, only cases exposed during the critical period were represented in the 243 cases examined. The relative risk of 40 does not represent the risk associated with exposure to rubella at any time, or even with exposure at any time during pregnancy; it is the risk of autism among cases exposed during the critical period for severe injury. Throughout this chapter, the reader will find more examples of cases ascertained from selected groups of exposed populations. Often, there is no other way to identify exposed cases.

The reports from Chess’ group did not consider the timing of exposure to rubella that resulted in autism, but critical periods for various rubella sequelae have been studied by others. Ueda, Nishida, Oshima, and Shepard (1979) investigated a sample of cases in which the time of appearance of the rash associated with the disease was known, and they described the outcome for eye defects, deafness, mental retardation, and heart malformations. They concluded that the period from the second to the fifth week postconception was associated with severe cognitive limitation, and cases with multiple defects had been exposed within the first 8 weeks. Because the autism cases reported by Chess and others all had multiple defects, it is likely that they had been exposed during the first 8 weeks postconception.

Fortunately, because of the development of a vaccine against rubella, we should no longer face the prospect of epidemics of this disease followed by epidemics of offspring with severe birth defects. As long as vaccination rates remain near 100%, this cause of autism has been eliminated.


Evidence and Relative Risk

In a clinic for children with Fetal Alcohol Syndrome (FAS) or Fetal Alcohol Effects (FAE), Nanson (1992) identified six who scored as severely autistic on the Childhood Autism Rating Scale (Schopler et al., 1988) in a sample of 326 cases. Each was also diagnosed with FAS, which requires facial dysmorphology, growth retardation, and dysfunction of the central nervous system. Subsequent papers (Harris et al., 1995; Aronson et al., 1997) have supported the conclusion that ASDs can coexist with FAS and perhaps with FAE. The additional cases described in these papers all shared the facial anomalies characteristic of FAS. The relative risk, based on the first study and again using Chakrabarti and Fombonne’s data for autistic disorder, is over 10. Fombonne (2002) has pointed out several difficulties that make the ethanol studies hard to interpret. None is based on the kinds of diagnostic measures for autism that are commonly used today. Furthermore, FAS and FAE are very common, so it is likely that they would occur in concert with autism occasionally by chance. However, the relative risk of 10 is still high, even with those caveats.

Rather than depending completely on the epidemiological findings, a teratologist might ask other questions to determine whether an association of ethanol exposure with autism makes biological sense. For example, “Is the critical period for FAS similar to any critical period associated with autism?” The critical period for FAS is thought to be in the third to fifth week postconception (Sulik et al., 1986). This period overlaps with the one we have just discussed for rubella, making it seem reasonable that ethanol exposures leading to FAS are occurring at a developmental stage that is sensitive to autism outcomes, as well. Another question might be, “Do children with FAS and autism share any dysmorphic features with autism after other exposures?” Yes, the facial dysmorphologies that characterize FAS (epicanthal folds, short palpebral fissures, underdeveloped maxillary region [Jones & Smith, 1973]) include features observed in autism subsequent to valproate exposure (to be discussed below). This suggests that the critical periods for autism after ethanol exposure and valproate exposure are similar. Thus, from an embryological standpoint, an FAS–autism association makes sense.

FAS and ASD Compared

Could it be that the behavioral phenotypes of ASDs and FAS are so similar that the autism diagnoses within FAS represent diagnostic substitution? First, let us consider the behavioral characteristics attributed to cases exposed to ethanol in utero. In the earliest description of FAS, based on eight severe cases in very young children (Jones et al., 1973), cognitive limitation was evident in all. Subsequent studies (e.g., Streissguth, et al., 1991) have examined a broader range of degrees of injury and included adolescents and adults, allowing a more detailed examination of cognitive abilities. The average IQ in this sample was 68, but almost half of the subjects fell in or near the normal range. Academic testing revealed that the subjects had particular difficulty with arithmetic, as compared to reading and spelling. Vineland (Sparrow et al., 1984) scores on daily living skills averaged age 9 years although the chronological age was 17 years. Socialization scores were the most affected, averaging 6 years of age. The patients in this study were not just deficient in adaptive behaviors, but exhibited a remarkable level of maladaptive behaviors. Of 31 subjects, none had a maladaptive behavior score in the insignificant range. The most common maladaptive behaviors cited were “poor concentration and attention, dependency, stubbornness or sullenness, social withdrawal, teasing or bullying, crying or laughing too easily, impulsivity, and periods of high anxiety.”

After many years of recording caregivers’ descriptions of the behavior of patients with prenatal alcohol injury, Streissguth, Bookstein, Barr, Press, and Sampson (1998) attempted to determine the descriptors most characteristic of offspring exposed to ethanol and evaluate descriptors for the strength of their association with other items on the list. The result is a 36-item checklist, the Fetal Alcohol Behavior Scale, with some predictive value for identifying people exposed to maternal drinking.

This set of behaviors is a list of features of FAS/FAE, just as the diagnostic behaviors of autism are a list of features of ASDs. Starting with the behavior most highly correlated with the scale as a whole, the list includes: (1) overreacting to situations, (2) chatting with little content, (3) bringing up unusual topics, (4) demanding attention, and (5) not aware of consequences of actions. Of 36 items, only a few sound similar to the behaviors seen in autism. Many of the behaviors are more like the opposite of autism (e.g., likes to talk, interrupts, wants to be the center of attention, physically loving, overly friendly).

Bishop, Gahagan, and Lord (2007) have compared children with ASD diagnoses to children with FAS/FAE diagnoses, in an attempt to determine the core features of autism that separate this diagnosis from other disabilities. They used the AutismDiagnostic Inventory Revised (ADI-R) (Lord et al., 1994) and the Autism Diagnostic Observation Schedule (ADOS) (Lord et al., 2000) to evaluate sample groups of children with autism, pervasive developmental disorder, and FAS/FAE who were matched for age and full-scale IQ. The results are presented as sets of items on which the groups show great differences and ones on which they are more similar. ADI items from all domains were among those that discriminated between the groups. For example, subjects with autism were much more likely to have problems with pointing to express interest, using gestures, and imitative social play. These are from the communications domain. From the social domain, difficulties on items such as range of facial expressions, sharing enjoyment, and offering comfort were more than twice as common in children with autism as in those with FAS/FAE. The one behavior from the restricted and repetitive behavior domain with reported difficulties strongly favoring the autistic group was hand and finger mannerisms.

Direct observation with the ADOS also showed striking differences across domains for children with autism compared to the Fetal Alcohol spectrum. These included problems with eye contact, directed facial expressions, amount of reciprocal social communication, and unusual sensory interests. The authors have pointed out that even categories of behavior for which the groups seem to have similar degrees of difficulty may reflect different kinds of behavior. For example, abnormalities of group play with peers could be scored for a child who ignores the approaches of peers, or for one who is loud, bossy, and demanding (the second list is borrowed from the Fetal Alcohol Behavior Scale, described above). Bishop et al. (2007) suggest that reduced propensity for and frequency of social interactions seem to be core features of autism, but the quality of such interactions may be disturbed in too many ways to be as useful for diagnosis.

In summary, it seems unlikely that investigators could mistake the behavioral characteristics of FAS for those of autism. It seems more likely that a subset of FAS/FAE cases exhibits a different behavioral phenotype consistent with an autism diagnosis. It remains to be seen whether that behavioral pattern arises from ethanol exposure alone, perhaps with slightly different timing, or on a different genetic background, from the more typical pattern, or whether the subset results from dual etiologies, with ethanol leading to the physical features by which FAS is diagnosed and some other factor leading to the behaviors diagnostic of autism. This author favors the view that high doses of ethanol early in pregnancy increase the risk of autism spectrum disorders.


Evidence and Relative Risk

The description of autism among people exposed to thalidomide in utero (Strömland et al., 1994) is part of a remarkable study that offers a very specific time of origin for the injuries that resulted in autism. More extensive reporting of the ophthalmological results of the study can be found in work by Miller (1991) and by Miller & Strömland (1991). Work by Strömland & Miller (1993) provides a detailed description of the physical anomalies that characterized the patients. Of about 100 cases in the Swedish registry of thalidomide victims, Miller and Strömland were able to enroll 86 and personally examine them for a host of ophthalmological measures, for physical malformations, and for cranial neurological dysfunctions. In the course of their investigation, they noted four cases with obvious mental retardation and psychiatric disturbances. With the assistance of a psychiatrist, the cases with mental retardation were diagnosed as having autistic disorder, using the criteria of DSM III-R. Four cases in a sample of 86 suggest a relative risk of almost 30. However, the authors collected other data that show that the risk is even greater than this.

Each of the 86 subjects was examined for physical malformations. Strömland and Miller (1993) summarized earlier work on the specific stages of development when thalidomide results in particular physical effects. The earliest defects are those of the ears (starting at day 20 postconception) and cranial nerves, followed by those of the upper limb (starting at day 24) and then the lower limb (starting at day 27). Among the 86 cases studied in Sweden, only 17 had been injured in the earliest part of the critical period for thalidomide embryopathy, as evidenced by ear anomalies and cranial nerve dysfunctions without limb anomalies (Strömland & Miller, 1993). Each of the cases with autism had ear anomalies and cranial nerve dysfunction, indicating early exposure. One also had an upper limb effect, and none had malformations of the lower limb. Therefore, the period of injury that results in an autism outcome must be between days 20 and 24. If we use the 17 cases injured early as our denominator for our relative risk calculation, rather than including the whole thalidomide-exposed sample, then the relative risk for autism after exposure during the critical period is 140 times the risk in the general population.

For readers not familiar with the field of teratology, the finding of a very narrow window when a particular defect can arise is not unusual, but typical, of birth defects in general. In most cases we don’t have supporting data to define a critical period as specific as the one in this study, but it is clear that such periods exist and that they depend on the coincidence of a particular exposure with particular events in development.

The critical period identified in this study is the time when the neural tube is closing and the first neurons are forming (Bayer et al., 1993). Those neurons are the motor neurons that make up the cranial nerve motor nuclei. The cranial nerve dysfunctions observed in the thalidomide study are evidence that some of these earliest-forming neurons were affected by the teratogen, offering another line of support to the critical period determined from physical malformations. Duane syndrome (a type of strabismus with lack of innervation from the abducens nerve to the lateral rectus muscle of the eye and reinnervation of the muscle by the oculomotor nerve) was seen in two of the cases with autism and in 24 cases altogether. Four others had some limitation of abduction. One case with autism had gaze paresis, as did six others in the total sample. Six patients had esotropia. Three of the patients with autism had palsy of the VIIth cranial nerve (the facial nerve) and 14 from the rest of the sample were affected. Abnormal lacrimation (lack of emotional tearing and/or crying to gustatory stimuli) was present in a total of 17 cases, including two of the cases with autism (Strömland et al., 1994; Strömland & Miller, 1993).

Mechanism of Action

The mechanism of action of thalidomide is not well-known. Recent hypotheses about mechanisms have focused on explaining the limb anomalies and may not explain the effects on the nervous system. It has long been known that the thalidomide-induced limb anomalies seen in primates and rabbits do not occur in rodents (Shumacher et al., 1968), but recent studies suggest that central nervous system (CNS) effects do occur in rodents (e.g., Myazaki et al., 2005; Hallene et al., 2006). Thus, it is possible that the mechanistic pathways involved in thalidomide teratogenicity may differ for different terata.

A recent series of studies in chick embryos indicates that thalidomide’s disruption of limb development may have its origin in disturbances of the expression of bone morphogenetic proteins, leading to overexpression of the gene Dickkopf1, which inhibits Wnt signaling (Knobloch et al., 2007). Because thalidomide is a potent antiangiogenic agent (Folkman, 1995), another hypothesis has focused on the effects of the drug on angiogenic factors, such as insulin growth factor1 and fibroblast growth factor 2 (Stephens et al., 2000). Another idea is that the drug’s induction of oxidative stress alters limb outgrowth by way of its effect on nuclear factor kappa B, a redox-sensitive transcription factor (Hansen et al., 2002). The authors have shown that thalidomide effects on this pathway differ between sensitive species (e.g., rabbit) and insensitive species (e.g., rat).

In summary, the Swedish thalidomide study, although carried out in humans, offers the kind of specificity and powerful conclusions usually seen only in animal studies. This is the result of many features of the study; for example, the fact that thalidomide’s effects include very rare anomalies made it relatively easy to identify cases. Additionally, medical record keeping in Sweden is so thorough that exposures could be documented retrospectively. The discovery of autism among the subjects was serendipitous, but it occurred because the investigators devoted so much time to interacting with each subject and family, completing exhaustive examinations of each case. It is also a tribute to the scientific acumen of these expert clinicians that they did not confine their curiosity to their specialty (pediatric ophthalmology). It was the additional data on somatic defects and cranial nerve dysfunction that allowed them to understand the meaning of the autism cases.

Valproic Acid

Evidence and Relative Risk

All the antiseizure medications are teratogenic to some degree (reviewed in Holmes, 2002). They are frequently given in combination. The reader should be aware that these drugs tend to interfere with one another. The result is that patients taking more than one medication tend to be on higher doses of each drug than patients who are taking only one.

The first report of autism in a child exposed in utero to valproic acid (Christianson et al., 1994) appeared in the same issue of Developmental Medicine and Child Neurology as Strömland et al. (1994). The paper discusses two sibling pairs with classic features of Fetal Valproate Syndrome (Di Liberti et al., 1984; Jager-Roman et al., 1986; Ardinger et al., 1988; Kozma, 2001), such as epicanthal folds, hypertelorism, broad nasal bridge; long upper lip with flat filtrum, ear malformations, and “pinched” finger tips. One of the four children tested positive for autism. Developmental delay, and especially delay in expressive language, had been reported earlier (Ardinger et al., 1988) but those cases were not tested for pervasive developmental disorder. Soon, another case of autism in Fetal Valproate Syndrome was described (Williams & Hersh, 1997), then five more (Williams et al., 2001). One of the five had a performance IQ of 100 and a verbal IQ of 81, suggesting that the autism risk in Fetal Valproate Syndrome is not dependent on the mental retardation sometimes seen in the syndrome.

More recent studies have examined larger populations of children exposed to valproic acid in utero and offer information on the risk of autism after exposure to this teratogen. In the first (Moore et al., 2000), 52 cases were ascertained through a parent support group for children diagnosed with a fetal anticonvulsant syndrome and five came from referrals to the local genetics service. The main purpose of the study was to document and compare the dysmorphologies of the various anticonvulsant exposures. In essence, the dysmorphologies of the valproate cases were similar to those described in the earlier case reports. Many other data were collected directly, as well, but behavioral diagnoses were ascertained from parental reports. As one might guess from the method of ascertainment, there was a very high degree of developmental delay or behavioral problems among children in the sample, with only four of the families reporting no problems.

Among the whole group, 34 had been exposed to valproate alone and 12 to valproate plus one or more other drugs. Among these 46 cases, there were three previously diagnosed with autism and two previously diagnosed with Asperger’s Syndrome. There was one case with an autism diagnosis in which the individual had been exposed to carbamazepine and diazepam. Assuming that the reported clinical diagnoses were accurate, the five in 46 rate of ASDs appears to be very high. Because of the Asperger’s cases, we must calculate the relative risk using Chakrabarti and Fombonne’s value of 62.6/10,000 for pervasive developmental disorder (PDD) as the denominator, so the relative risk is about 17. However, in this study, with its subjects selected from a group already known to have behavioral problems, we have no way to estimate the true risk of ASDs among all children exposed to valproate in utero.

Rasalam et al. (2005) carried out a population-based study using records over a 20-year period to identify all local children known to have been exposed to antiepileptic drugs in utero. Their mothers had been referred to Aberdeen Maternity Hospital because of their high-risk pregnancies. There were 398 mothers who had 626 exposed children at the facility. From this pool, 159 mothers with 260 children agreed to participate in the study. Families were interviewed by a trained research nurse. The structured interview included questions about behavioral and social issues.

From the responses, 26 children were identified for an extensive review of their medical records, and 14 of these had indications of ASDs. The investigators accessed this group’s complete records, including reports from specialists who had examined their behavior. A child psychiatrist then studied these notes for items specified in the DSM IV criteria for autism, and found that 12 met the criteria for ASD. Among those who had been exposed to valproate alone, five of 56 were positive and among those exposed to valproate alone or in combination with other drugs, nine of 77 were positive. Eight qualified for a diagnosis of autism and one for a diagnosis of Asperger’s Syndrome. Comparing this rate to Chakrabarti and Fombonne’s rate for PDD, the relative risk is about 19.

There may be a remaining selection bias in this study, even though it was a serious attempt to assess the whole exposed population. It seems possible that families of children with behavioral difficulties might be more likely to agree to recruitment than families of children developing normally. Unfortunately, we don’t know whether this was the case, nor do we know how many children of the initial 626 children exposed to antiseizure medication were exposed to valproate. We don’t know how many were lost to follow-up (their appearance in the unenrolled group does not represent a refusal of recruitment). However, assuming that the proportion exposed to valproate was the same in the unenrolled cases as in those who were enrolled, then even if we include those lost to follow-up and assume that there were no cases of ASDs in the unenrolled group, the risk associated with valproate would still be substantial. In fact, it is 7.72 with a confidence interval of 3.90 to 15.29. This is significantly different from 1.00 (no effect) at p = < .0001.

Now let us consider some other characteristics of the children evaluated in this study. For some variables, the authors describe the children with ASD diagnoses subsequent to exposure to antiseizure medications as a group, without separating out those exposed to valproic acid, but since most of the cases (9/12) had been exposed to valproate, these generalizations are of interest. The mean IQ of those with ASDs was below average, but most cases were in the normal range. No child had any evidence of regression or loss of skills. Results specific to the children exposed to valproate included: head circumference above average, a male to female ratio of 3:6, and an assortment of birth defects (e.g., strabismus, pyloric stenosis, hypospadias).

Children with autism after in utero exposure to valproate share the many dysmorphologies of Fetal Valproate Syndrome with exposed children who do not meet the diagnostic criteria for autism. Some of these features appear in other environmental etiologies as well. For example, epicanthal folds and maxillary hypoplasia are typical of Fetal Alcohol Syndrome as well as Fetal Valproate Syndrome. Strabismus and ear anomalies characterize children exposed to thalidomide as well as those exposed to valproate. These common features suggest that the critical periods for valproate teratogenicity must be similar to those for ethanol and thalidomide.

Mechanism of Action

The mechanism of action of valproic acid in teratogenesis cannot be described completely at this time, but at least two mechanisms have been identified. First, valproate is a direct inhibitor of histone deacetylase (HDAC) (Phiel et al., 2002). The HDACs have a role in the folding of chromatin, and their inhibition can unfold chromosomes, making some genes more available for transcription. Other HDAC inhibitors are also teratogenic, so there is good reason to think that this mechanism is one of the ways valproic acid alters development. It has been proposed that increases in Wnt gene signaling after HDAC inhibition are one of the mechanisms by which valproic acid leads to birth defects (reviewed in Wiltse, 2005).

A second mechanism, which may be related to the first (see Gurvich et al., 2005), is the ability of valproate to drive the expression of the gene, Hoxa1 (Stodgell et al., 2006). This gene is critical to the development of the early embryo when the neural plate is developing into the neural tube, and it plays a special role in hindbrain development (Chisaka et al., 1992). Its expression is dependent on levels of retinoic acid and both high and low levels of retinoic acid are extremely teratogenic, presumably because of their effect on the levels of Hoxa1 expression (e.g., Means & Gudas, 1995; White et al., 2000).

A number of analogs of valproic acid have been synthesized and tested for antiseizure activity and teratogenicity (e.g., Hauck & Nau, 1989; Bojic et al., 1996). These studies have demonstrated that the antiseizure property of the drug and the teratogenic property do not share the same mechanism. Stodgell and others (2006) made use of these analogs in their studies of gene expression in rat embryos after exposure to valproate. They found that compounds with high teratogenicity had strong effects on Hoxa1 expression, although those with no teratogenic action had no effect on the expression of the gene. Thus, the effect on Hoxa1 may be one mechanism of action of the drug. These effects on the gene’s expression can be seen only during its normal period of expression (the time of neural tube closure) and very brief periods before and after normal expression.

Animal Model

More than the other environmental factors discussed above, valproate has proven useful in the creation of animal models relevant to autism (reviewed in Arndt et al., 2006). Rats exposed to the drug around the time of neural tube closure exhibit several neuroanatomical features similar to ones reported in histological studies of brains from people with autism. For example, the rat model has low Purkinje cell counts and small cerebellar volume (Ingram et al., 2000). The same features have been reported in human autism cases (e.g., Bauman & Kemper, 1985; Ritvo et al., 1986; Bailey et al., 1998; Whitney et al., 2008; Courchesne et al., 1988).

The rat model has also been found to express a variety of behavioral abnormalities (e.g., Schneider et al., 2001; Schneider & Przewlocki, 2005; Stanton et al., 2007; Schneider et al., 2007; Markram et al., 2008). Many of the behavioral effects are in general categories of behavior that are related to autism (e.g., nociception, social behavior) but a few are in very specific behaviors known to be affected in the same way as in autism. One is a decrement in prepulse inhibition of the acoustic startle response. This was first reported in subjects with ASDs by McAlonan and colleagues (2002) and more recently by Perry and colleagues (2007). The same depression of prepulse inhibition has been demonstrated in rats exposed to valproate around the time of neural tube closure (Schneider & Przewlocki, 2005). This measure of sensory gating in the CNS could be related to failures of inhibition as seen in repetitive thoughts and actions in autism.

Eyeblink conditioning is a second very specific behavior affected similarly in both human autism and the valproic acid rat model. This basic paradigm of association learning, in which a tone is paired repeatedly with a stimulus to the eye that causes a blink until the blink occurs to the tone alone, has been studied in many neurological and psychiatric conditions. For example, children with FAE or dyslexia are severely impaired in acquisition of the conditioned eyeblink response (Coffin et al., 2005), as are people with Alzheimer’s disease (Woodruff-Pak & Papka, 1996a). People with Huntington’s disease have normal acquisition of the learned response, although the timing of their blinks is abnormal (Woodruff-Pak & Papka, 1996b). In contrast to all the other conditions that have been studied, autism results in an enhancement of conditioning (Sears et al., 1994; Arndt et al., 2006). The same enhancement has been demonstrated in the rat model exposed to valproate (Stanton et al., 2001). Thus, the rat exposed to valproate around the time of neural tube closure has both anatomic and behavioral parallels to autism.


Evidence and Relative Risk

Misoprostol is a prostaglandin with various legitimate uses in medicine. In South America, it is used illegally by the poor as an abortifacient. Exposure causes uterine contractions, which may or may not expel the conceptus. When the drug fails and the conceptus comes to term there is a risk of birth defects consistent with ischemia in utero (Gonzalez et al., 1998). One of these is Moebius sequence, a bilateral or unilateral diplegia of the facial nerve and the abducens nerve. Investigations of familial cases of Moebius sequence indicate that anomalies of multiple genes can produce the deficit (Verzijl et al., 1999). There is also evidence that environmental factors play a role (Lipson et al., 1989). Ischemia was suggested as a possible cause of Moebius sequence long before the teratology of misoprostol was recognized (Bavinck & Weaver, 1986). The rate of autism is high among idiopathic cases of Moebius sequence (e.g., Johansson et al., 2001).

Bandim and colleagues (2003) recruited children diagnosed with Moebius sequence in Brazil, queried their mothers regarding misoprostol use, and tested the Moebius cases for autism. In the 23 cases under study, 14 had been exposed to misoprostol, according to the mothers’ statements. Among these, three met the criteria for an autism diagnosis under DSM IV. Of the nine children whose mothers denied misoprostol use, two met the criteria for an autism diagnosis. Thus, the relative risk for autism in misoprostol-induced Moebius sequence is 128. Notice that this incredible risk does not apply to misoprostol exposure in general, but only to exposures with dose and timing adequate to cause Moebius sequence. Unfortunately, at this time, we have no way to tell whether misoprotol exposure has occurred unless it has caused an obvious but rare birth defect such as Moebius sequence. The results in cases not exposed to misoprostol support earlier findings that idiopathic cases of Moebius sequence have a high rate of autism. Indeed, the 2/9 rate is just what other studies would predict. For example, the rate in the study by Johansson and colleagues (2001) was about 25%.

Why is Moebius sequence related to autism? The fact that ischemia of the brain stem appears to be the etiology of the misoprostol cases suggests that it may be the location of injury that links the two conditions. All the mothers who reported using misoprostol in studies by Bandim et al. (2003) exposed the embryo in the sixth week postconception, so the timing is slightly later than that discussed for other teratogens that increase autism risk. Further, the proposed mechanism, ischemia, seems unrelated to the mechanisms proposed for other factors. But we know from the neurological symptoms of valproate-exposed cases and thalidomide-exposed cases that brain stem nuclei were affected and animal studies show that valproate exposure can injure the developing cranial nerve motor nuclei (e.g., Rodier et al., 1996). Thus, location of injury appears to be the one obvious common feature shared by Moebius sequence and autism associated with environmental etiologies.

What Can We Learn from Environmental Etiologies?

Early Origins and Neurobiology

The fact that all the environmental factors recognized thus far share critical periods in embryonic life (the first 9 weeks postconception) is an important contribution to our understanding of autism etiologies. Table 48-1 summarizes the critical periods for all the known environmental risk factors. At the time these studies began to appear, few scientists were thinking that autism’s origins might occur so early in development. But in retrospect it seems obvious that very early injuries are the ones with the potential to explain the histology and connectivity observed in the brains of people with autism. For example, low numbers of Purkinje cells have been reported by many investigators, as discussed above, and Ingram et al. (2000), demonstrated that this condition could be reproduced in rats exposed to valproic acid during neural tube closure. Interestingly, the exposure used in this study occurred before the Purkinje cells form, so they were not injured directly. This reminds us that insults to a developing brain can have consequences for later-forming structures.

Table 48–1. Exposure period associated with an autism outcome for five known environmental risk factors


Risk Factor

Critical Period



< 8 weeks


2 to 5 weeks


20 to 24 days

Valproic acid

3 to 4 weeks


6th week

The deep nuclei of the cerebellum have been studied in a number of brains from people with autism (Bauman & Kemper, 1994; Bailey et al., 1998). The globose and emboliform nuclei were especially altered, although the dentate nucleus was less affected. These nuclei form just before the Purkinje cells (Bayer et al., 1993), again suggesting an early injury.

Rodier et al. (1996) reported a case of autism in which the superior olive was virtually absent. This very early-forming complex of nuclei (Bayer et al., 1993) is the first point in the auditory pathway where information from both ears converges on the same neurons, and plays a role in sound localization. Recently, experts on the structure examined five brains from people with ASDs and found that the medial superior olive was abnormal in every case (Kulesza & Mangunay, 2008). The normal strict orientation of the neurons was deranged, even in cases where the number of neurons appeared to be unaffected.

Early loss of neurons can alter the connections formed subsequently by other neurons. Duane syndrome, as seen in thalidomide-exposed subjects (Strömland & Miller, 1993) is a good example. Many theories of the neurobiology of autism focus on abnormalities of connections (e.g., Just et al., 2004; Kana et al., 2007), and it is easy to see how projections might go to the wrong places when tissue that provides axonal guidance to them is missing or altered. In the case reported by Rodier et al. (1996), fibers that normally go up or down the neuroaxis were seen going in all directions in the region where they would typically form a capsule around the facial nucleus. In its absence, they appear to have lost their orientation. In Bailey et al. (1998), one case of autism exhibited an extra brain stem tract never seen in controls. All these findings fit well with the idea of a disruption of development in the early embryo.

Early Origins and Early Developmental Genes

In retrospect, we now know that a number of the genes thought to play a role in autism are ones whose main or only period of expression occurs in embryonic life. These include Engrailed 2, which organizes aspects of early cerebellar development (Joyner et al., 1991; Kuemerle et al., 1997). Several alleles of the human gene have been shown to be associated with autism, although the causal variants have not been identified (Gharani et al., 2004; Benayed et al., 2005).

Individuals who inherit two copies of a truncating mutation of HOXA1 have a high rate of autism, as well as deafness, vascular malformations, large head size, Duane syndrome, and facial hypotonia (Tischfield et al., 2005; Bosley et al., 2007; Bosley et al., 2008). This is called the Bosley-Salih-Alorainy syndrome (BSAS). Homozygous inheritance of another truncating mutation of the same gene (Holve et al., 2003; Bosley et al., 2008) leads to many of the same symptoms, but not autism.

Several theories of the mechanisms of action of valproate involve the Wnt gene pathway, and there is some evidence that WNT2 might be a susceptibility gene for autism (Wassink et al., 2001). The involvement of early developmental genes in the etiology of autism makes the early critical periods found for teratologic risk factors seem logical, rather than surprising.

Communalities Across Etiologies

It would be exciting if all the environmental risk factors shared similar mechanisms of action, but they do not. Although their critical periods are similar, they are not exactly the same. Yet this review demonstrates that patients with these different etiologies do share many characteristics. Table 48-2 summarizes some characteristics shared across different conditions leading to autism. For example, there are some common facial features shared across several risk factors, such as malformed ears, epicanthal folds, and maxillary hypoplasia. There are dysfunctions of cranial nerves, such as strabismus and Moebius sequence.

Table 48–2. Craniofacial anomalies reported in individuals with autism subsequent to four environmental exposures and one genetic syndrome






Large head size




Moebius sequence or

facial hypotonia




Duane syndrome or other strabismus





Abnormal tearing



Epicanthal folds



Ear abnormalities




Hearing deficits



Hypoplasia of the midface or maxilla



In addition, some individuals exposed to environmental risk factors have symptoms also observed in cases with disruptions of early developmental genes. Hearing impairments are typical of rubella-induced autism, thalidomide-induced autism, and of both syndromes traced to homozygous inheritance of truncating mutations of HOXA1. The large head size reported in valproate-associated autism is also present in BSAS (Tischfield et al., 2005). Duane syndrome is typical of thalidomide-induced autism and of both syndromes arising from deficient expression of the HOXA1 protein.

Do idiopathic cases of autism share any of the symptoms reviewed for environmental exposure etiologies and early developmental gene etiologies? Obviously, they do. The high rate of minor craniofacial malformations in idiopathic ASDs has been recognized for many years (e.g., Steg & Rapoport, 1975; Walker, 1977). Because these can only arise during early stages of development, their presence suggests that many idiopathic cases have been subject to some early disturbance of development.

Strabismus, which we have seen to occur after exposure to thalidomide and valproate, and after homozygous inheritance of truncating mutations of HOXA1, has been noted to occur at elevated rates in idiopathic cases of autism as well (e.g., Sharre & Creedon, 1992). Similarly, the comorbidity of Moebius sequence or facial hypotonia with autism is seen in idiopathic cases (e.g., Johansson et al., 2001), as well as in cases associated with environmental risk factors and genetic syndromes reviewed here.

Large head size is a particularly interesting characteristic of a substantial number of people with ASDs that has received much attention. It appears that overgrowth of the brain and head begins soon after birth and tapers off in the second year of life (e.g., Courchesne et al., 2003; Dawson et al., 2007). One might speculate that postnatal overgrowth reflects some untoward influence on postnatal development occurring during the period of excessive growth. However, the fact that prenatal exposure to valproic acid also results in large head size (Rasalam et al., 2005), suggests that it may not be necessary to propose any postnatal events to account for this phenomenon. Indeed, the presence of large head size in Bosley-Salih-Alorainy syndrome proves that genetic activity altered only in the embryonic period can lead to large head size later in life. Even a single nucleotide polymorphism of HOXA1 has been shown to be associated with the quantitative trait of head size in autism (Conciatori et al., 2004). The same polymorphism influences head growth rates in typically developing children as well (Muscarella et al., 2007). Even though large heads occur after birth, the conditions that create them may be set in motion much earlier in development.

Does Cognitive Limitation Play a Part in Environmental Etiologies of Autism?

The question has often been raised as to whether the cognitive impairment associated with many teratogenic exposures creates the apparent association of ASDs with environmental factors. In some situations, as when sample cases are collected from institutional settings, the patients are almost certain to have cognitive deficits, but this may not be true of all autism cases with the same etiology. A good example comes from two studies of autism in Moebius sequence. In the first, subjects were recruited from institutional care, and all the autism cases had cognitive limitation (Johansson et al., 2001). In the second study, subjects were recruited from families attending an international meeting of support groups for people with a Moebius sequence diagnosis (McConnell et al., 2002). In this sample, the rate of autism was slightly higher than in the first study, but more than half of the autism cases had IQs in the normal range. Among the studies of environmental exposures discussed in this chapter, several have provided data on IQ. Rasalam et al. (2005) reported IQ data for all their subjects exposed to valproate. Although the mean IQ was significantly lower in subjects with ASDs than in those without, most subjects with ASDs had IQs within the normal range. The four thalidomide cases studied in Strömland et al. (1994) had cognitive limitation, but a fifth case, whom the investigators were unable to recruit, was known from her records to have autism and a normal IQ. Additional data from the Brazilian misoprostol study has been shared with the author, and suggests that three of seven autism cases tested for IQ fell in the normal range (personal communication from Liana Ventura, 2009). Taken together, the available data do not rule out a role for cognitive limitation as a contributing factor in the autism rates observed after environmental exposures, but they do indicate that cognitive limitation is not a requirement for the development of autism after environmental exposures.

In summary, the discovery of environmental factors that increase the risk of autism has given us information that helps to explain many histological findings in this spectrum of disorders. An early origin for autism fits well with the issues of connectivity that seem to be part of the neurobiology of ASDs. Cases with different environmental etiologies are related to one another in symptoms and the same symptoms can be seen in some cases with known genetic etiologies. Similarly, those same symptoms often occur in idiopathic cases. Rather than being a set of rare etiologies unrelated to the much greater number of cases of unknown origin, cases associated with environmental risks appear to represent different developmental pathways to similar phenotypes.


The discovery of environmental exposures that increase the risk of autism spectrum disorders has provided the field with new ideas regarding the etiology of autism and new approaches to understanding its neurobiology. The idea that autism arises in early stages of development is a strong message from studies of environmental risk factors. Animal models, such as the one based on early exposure to valproic acid, promise to add to our understanding of both anatomy and behavior in autism. Comparing cases whose autism arose after environmental exposures to idiopathic cases and to some genetic syndromes suggests that different etiologies often result in similar outcomes. Not only are the diagnostic behaviors similar across etiologies, but so are some neurological and somatic anomalies that occur with increased frequency in people with autism. Although we need ways to stratify cases to advance our understanding of the genetics of autism, we also need to recognize phenotypic features that are shared by different etiologies.

Challenges and Future Directions

  • It is no accident that the first five environmental risk factors identified are a virus and four drugs. It is much easier to identify these kinds of exposures and determine dose levels for them than for many potential teratogens. Can we recognize the association between terata and exposures to agents that are widespread in the environment? This is a much greater challenge, especially when individual doses are unknown.

  • Susceptibility genes that interact with environmental exposures should vary from one risk factor to another. Can we identify these genes?

Suggested Readings

Arndt, T. L., Stodgell, C. J., & Rodier, P. M. (2005). The teratology of autism. International Journal of Developmental Neuroscience, 23, 189–199.

Rasalam, A. D., Hailey, H., Williams, J. H., Moore, S. J., Turnpenny, P. D., & Lloyd, D. J., et al. (2005). Characteristics of fetal anticonvulsant syndrome associated autistic disorder. Developmental Medicine and Child Neurology, 47, 551–555.

Strömland, K., Nordin, V., Miller, M., Akerstrom, B., & Gillberg, C. (1994). Autism in thalidomide embryopathy: A population study. Developmental Medicine and Child Neurology, 36, 351–356.


Ardinger, H. H., Atkin, J. F., Blackston, R. D., Elsas, L. J., Clarren, S. K., Livingstone, S., et al. (1988). Verification of the fetal valproate syndrome phenotype. American Journal of Medical Genetics, 29, 171–185.Find this resource:

Ariens Kappers, J. (1957). Developmental disturbance of the brain induced by German measles in an embryo of the 7th week. Acta Anatomica, 31, 1–20.Find this resource:

Arndt, T. L., Chadman, K. K., Watson, D. J., Tsang, V., Rodier, P. M., & Stanton, M. E. (2006). Long delay eyeblink conditioning inautism [Abstract]. Birth Defects Research, Part A, 76, 323.Find this resource:

    Arndt, T. L., Stodgell, C. J., & Rodier, P. M. (2005). The teratology of autism. International Journal of Developmental Neuroscience, 23, 189–199.Find this resource:

    Aronson, M., Hagberg, B., & Gillberg, C. (1997). Attention deficits and autism spectrum problems in children exposed to alcohol during gestation: A follow-up study. Developmental Medicine and Child Neurology, 39, 583–587.Find this resource:

    Bailey, A., Le Couteur, A., Gottesman, I., Bolton, P., Simonoff, E, Yuzda, E., et al. (1995). Autism as a strongly genetic disorder: Evidence from a British twin study. Psychological Medicine, 25, 63–77.Find this resource:

    Bailey, A., Luthert, P., Dean, A., Harding, B., Janota, I., Mongomery, M., et al. (1998). A clinicopathological study of autism. Brain, 121, 889–905.Find this resource:

    Bandim, J. M., Ventura, L. O., Miller, M. T., Almeida, H. C., & Costa, A. E. (2003). Autism and Mobius sequence: An exploratory study of children in northeastern Brazil. Arquivos de Neuro-Psiquiatria, 61, 181–185.Find this resource:

    Bauman, M. L. & Kemper, T. L. (1985). Histoanatomic observations of the brain in early infantile autism. Neurology, 35, 866–874.Find this resource:

    Bauman, M. L. & Kemper, T. L. 1994. Neuroanatomic observations in autism. In M. L. Bauman & T. L. Kemper (Eds.), The neurobiology of autism (pp. 119–145). Baltimore, MD: Johns Hopkins University Press.Find this resource:

      Bavinck, J. N. B. & Weaver, D. D. (1986). Subclavian artery supply disruption sequence: Hypothesis of a vascular etiology for Poland, Kippel-Feil, and Moebius anomalies. American Journal of Medical Genetics, 23, 903–919.Find this resource:

      Bayer, S. A., Altman, J., Russo, R. J., Zhang, X. (1993). Timetables of neurogenesis in the human brain based on experimentally determined patterns in the rat. Neurotoxicology, 14, 83–144.Find this resource:

      Benayed, R., Gharani, N., Rossman, I., Mancuso, V., Lazar, G., Kamdar S, et al. (2005). Support for the homeobox transcription factor gene ENGRAILED 2 as an autism spectrum disorder susceptibility locus. American Journal of Human Genetics, 77, 851–868.Find this resource:

      Bishop, S., Gahagan, S., & Lord, C. (2007). Re-examining the core features of autism spectrum disorder and fetal alcohol spectrum disorder. Journal of Child Psychology and Psychiatry, 48, 1111–1121.Find this resource:

      Bojic, U., Elmazar, M. M., Hauck, R. S. & Nau, H. (1996). Further branching of valproate-related carboxylic acids reduces the teratogenic activity, but not the anticonvulsant effect, Chemical Research in Toxicology, 9, 866–870.Find this resource:

      Bosley, T. M., Alorainy, I. A., Salih, M. A., Aldhalaan, H. M., Abu-Amero, K. K., Oystreck, D.T., et al. (2008). The clinical spectrum of homozygous HOXA1 mutations. American Journal of Medical Genetics, 146, 1235–1240.Find this resource:

      Bosley, T.M., Salih, M. A., Alorainy, I. A., Oystreck, D.T., Nester, M., Abu-Amero, K. K., et al., (2007). Clinical characterization of the HOXA1 syndrome BSAS variant. Neurology, 69, 1245–1253.Find this resource:

      Chakrabarti, S. & Fombonne, E. (2001). Pervasive developmental disorders in preschool children. Journal of the American Medical Association, 285, 3093–3099.Find this resource:

      Chess, S., Fernandez, P., & Korn, S. (1978). Behavioral consequences of congenital rubella. Journal of Pediatrics, 93, 699–703.Find this resource:

      Chess, S. & Fernandez, P. (1980). Neurologic damage and behavior disorder in rubella children. American Annals of Deafness, 125, 998–1001.Find this resource:

      Chisaka, O., Musci, T. S., & Cappechi, M. R. (1992). Developmental defects of the ear, cranial nerves, and hindbrain resulting from targeted disruption of the mouse homeobox gene Hox1.6. Nature, 355, 516–520.Find this resource:

      Christianson, A. L., Chesler, N., & Kromberg. J. G. R. (1994). Fetal valproate syndrome: Clinical and neurodevelopmental features in two sibling pairs. Developmental Medicine and Child Neurology, 36, 357–369.Find this resource:

      Coffin, J. M,. Baroody, S., Schneider, K., & O’Neill, J. (2005). Impaired cerebellar learning in children with prenatal alcohol exposure: a comparative study of eyeblink conditioning in children with ADHD and dyslexia. Cortex, 41, 389–398.Find this resource:

      Conciatori, M., Stodgell, C. J., Hyman, S. L., O’Bara, M., Militerni, R., Bravaccio, C., et al. (2004). Association between the HOXA1 A218G polymorphism and increased head circumference in patients with autism. Biological Psychiatry, 55, 413–419.Find this resource:

      Courchesne, E., Carper, R., & Akshoomoff, N. (2003). Evidence of brain overgrowth in the first year of life in autism. Journal of the American Medical Association, 290, 337–344.Find this resource:

      Courchesne, E., Yeung-Courschesne, R., Press, G. A., Hesselink, J. R., & Jernigan, T. L. (1988). Hypoplasia of cerebellar vermal lobules VI and VII in autism. New England Journal of Medicine, 318, 1349–1354.Find this resource:

      Dawson, G., Munson, J., Webb, S. J., Nalty, T., Abbott, R., & Toth, K. (2007). Rate of head growth decelerates and symptoms worsen in the second year of life in autism. Biological Psychiatry, 61, 458–464.Find this resource:

      DiLiberti, J. H., Farndon, P. A., Dennis, N. R., & Curry, C. J. R. (1984). The fetal valproate syndrome. American Journal of Medical Genetics, 19, 473–481.Find this resource:

      Folkman, J. (1995). Clinical applications of research on angiogenesis. New England Journal of Medicine, 333, 1757–1763.Find this resource:

      Fombonne, E. (2002). Is exposure to alcohol during pregnancy a risk factor for autism? Journal of Autism and Developmental Disorders, 32, 243.Find this resource:

      Gharani, R., Benayed, V., Mancuso, L. M., Brzustowicz, L., & Millonig, J. H. (2004). Association of the homeobox transcription factor, ENGRAILED 2, with autism spectrum disorder. Molecular Psychiatry, 9, 474–484.Find this resource:

      Gonzalez, C. H., Marques-Dias, M. J., Kim, C. A., Sugayama, S. M. M., Da Paz, J. A., Huson, S. M., et al. (1998). Congenital anomalies in Brazilian children associated with misoprostol misuse in first trimester of pregnancy. Lancet, 351, 1624–1627.Find this resource:

      Gregg, N. M. (1941). Congenital cataract following German measles in the mother. Transactions of the Ophthalmologic Society of Australia, 3, 35–46.Find this resource:

        Gurvich, N., Berman, M. G., Wittner, B. S., Gentleman, R. C., Klein, P. S., & Green, J. B. A. (2005). Association of valproate-induced teratogenesis with histone deacetylase inhibition in vivo. FASEB Journal, 19, 1166–1187.Find this resource:

        Hallene, K. L., Oby, E., Lee, B. J., Santaguida, S., Bassanini, S., & Cipolla, M. (2006). Prenatal exposure to thalidomide, altered vasculogenesis, and CNS malformations. Neuroscience, 142, 267–283.Find this resource:

        Hansen, J. M., Harris, K. K., Philbert, M. A., & Harris, C. (2002). Thalidomide modulates nuclear redox status and preferentially depletes glutathione in rabbit limb versus rat limb. Journal of Pharmacology and Experimental Therapeutics, 300, 768–776.Find this resource:

        Harris, S. R., MacKay, L. L., & Osborn, J. A. (1995). Autistic behaviors in offspring of mothers abusing alcohol and other drugs: A series of case reports. Alcoholism: Clinical and Experimental Research, 19, 660–665.Find this resource:

        Hauck, R. S. & Nau, H. (1989). Asymmetric synthesis and enantioselective teratogenicity of 2-n-propyl-4-pentenoic acid (4-en-VPA), an active metabolite of the anticonvulsant drug, valproic acid. Toxicology Letters, 49, 41–48.Find this resource:

        Holmes, L. B. (2002). The teratogenicity of anticonvulsant drugs: A progress report. Journal of Medical Genetics, 39, 245–247.Find this resource:

        Ingram, J. L., Peckham, S. M., Tisdale, B., & Rodier, P. M. (2000). Prenatal exposure of rats to valproic acid reproduces the cerebellar anomalies associated with autism. Neurotoxicology and Teratology, 22, 319–324.Find this resource:

        Jager-Roman, E., Deichl, A., Jakob, S., Hartmann, A. M., Koch, S., Rating, D., et al. (1986). Fetal growth, major malformations, and minor anomalies in infants born to women receiving valproic acid. Journal of Pediatrics, 108, 997–1004.Find this resource:

        Johansson, M., Wentz, E., Fernell, E., Strömland, K., Miller, M. T, & Gillberg, C. (2001). Autistic spectrum disorder in Mobius sequence: A comprehensive study of 25 individuals. Developmental Medicine and Child Neurology, 43, 338–345.Find this resource:

        Jones, K. L. & Smith, D. W. (1973). Recognition of the Fetal Alcohol Syndrome in early infancy. Lancet, 836, 999–1001.Find this resource:

        Jones, K. L., Smith, D. W., Ulleland, C. N., & Streissguth, A. P. (1973). Pattern of malformation in the offspring of chronic alcoholic mothers. Lancet, 815, 1267–1271.Find this resource:

        Joyner, A. L., Herrup, K., Auerbach, B.A., Davis, C.A., & Rossant, J. (1991). Subtle cerebellar phenotype in mice homozygous for a targeted deletion of the En-2 homeobox, Science, 251, 1239–1243.Find this resource:

        Just, M. A., Cherkassy, V. L., Keller, T. A., & Minshew, N. J. (2004). Cortical activation and synchronization during sentence comprehension in high-functioning autism: Evidence of underconnectivity. Brain, 127, 1811–1821.Find this resource:

        Kana, R. K., Keller, T. A., Minshew, N. J., & Just, M. A. (2007). Inhibitory control in high-functioning autism. Biological Psychiatry, 62, 198–206.Find this resource:

        Kanner, L. (1943). Autistic disturbances in affective contact. Nervous Child, 2, 217–250.Find this resource:

          Knobloch, J., Shaughnessy, J. D., & Ruther, U. (2007). Thalidomide induces limb deformities by perturbing the Bmp/Dkk1/Wnt signaling pathway. FASEB Journal, 21, 1410–1421.Find this resource:

          Kozma, C. (2001). Valproic acid embryopathy: Report of two siblings with further expansion of the phenotypic abnormalities and a review of the literature. American Journal of Medical Genetics, 98, 168–175.Find this resource:

          Kuemerle, B., Zanjani, H., Joyner, A., & Herrup, K. (1997). Pattern deformities and cell loss in Engrailed-2 mutant mice suggest two separate patterning events during cerebellar development. Journal of Neuroscience, 17, 7881–7889.Find this resource:

          Kulesza, R. J., & Mangunay, K. (2008). Morphological features of the medial superior olive in autism. Brain Research, 1200, 132–137,Find this resource:

          Lipson, A. H., Webster, W. S., Brown-Woodman, P. D. C., & Osborn, R. A. (1989). Moebius syndrome: Animal model–human correlations and evidence for a brainstem vascular etiology. Teratology, 40, 339–350.Find this resource:

          Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., DiLavore, P. S., et al. (2000). The Autism Diagnostic Schedule-Generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30, 205–223.Find this resource:

          Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism Diagnostic Interview-Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24, 659–685.Find this resource:

          Markram, K., Rinaldi, T., La Mendola, D., Sandi, C., & Markram, H. (2008). Abnormal fear conditioning and amygdala processing in an animal model of autism. Neuropsychopharmacology, 33, 901–912.Find this resource:

          McAlonan, G. M., Daly, E., Kumari, V., Critchley, H. D., van Amelsvoort, T., Suckling, et al. (2002). Brain anatomy and sensorimotor gating in Asperger’s syndrome. Brain, 127, 1594–1606.Find this resource:

          McConnell, B., Drmic, I., Roberts, W., Miller, M. T., & Bryson, S. E. (2002, November 2). The co-occurrence of autistic spectrum disorders and Moebius syndrome: Data from the ADI-R and ADOS. Poster session at the International Meeting for Autism Research, Orlando FL.Find this resource:

            Means, A. L. & Gudas, L. J. (1995). The role of retinoids in vertebrate development. Annual Review of Biochemistry, 64, 201–233.Find this resource:

            Miller, M. T. (1991). Thalidomide embryopathy: A model for the study of congenital incomitant horizontal strabismus. Transactions of the American Ophthalmological Society, 89, 623–674.Find this resource:

            Miller, M. T. & Strömland, K. (1991). Ocular motility in thalidomide embryopathy. Journal of Pediatric Ophthalmology and Strabismus, 28, 47–54.Find this resource:

            Miyazaki, K., Narita, N., & Narita, M. (2005). Maternal administration of thalidomide or valproic acid causes abnormal serotonergic neurons in the offspring: Implication for pathogenesis of autism. International Journal of Developmental Neuroscience, 23, 287–297.Find this resource:

            Moore, S. J., Turnpenny, P., Quinn, A., Glover, S., Lloyd, D. J., Montgomery, T., et al. (2000). A clinical study of 57 children with fetal anticonvulsant syndrome. Journal of Medical Genetics, 37, 489–497.Find this resource:

            Muscarella, L. A., Guarnieri, V., Sacco, R., Militerni, R., Bravaccio, C., Trillo, S., et al. (2007). HOXA1 gene variants influence head growth rates in humans. American Journal of Medical Genetics, Part B, 144, 388–390.Find this resource:

            Nanson, J. L. (1992). Autism in fetal alcohol syndrome: A report of six cases. Alcoholism: Clinical and Experimental Research, 16, 558–565.Find this resource:

            Perry, W., Minassian, A., Lopez, B., Maron, L., & Lincoln, A. (2007). Sensorimotor gating deficits in adults with autism. Biological Psychiatry, 61, 482–486.Find this resource:

            Phiel, C. J., Zhang, F., Huang, E. Y., Guenther, M. G., Lazar, M. A., & Klein, P. S. (2001). Histone deacetylase is a direct target of valproic acid, a potent anticonvulsant, mood stabilizer, and teratogen. Journal of Biological Chemistry, 276, 36734–36741.Find this resource:

            Rasalam, A. D., Hailey, H., Williams, J. H., Moore, S. J., Turnpenny, P. D., Lloyd, D. J., et al. (2005). Characteristics of fetal anticonvulsant syndrome associated autistic disorder. Developmental Medicine and Child Neurology, 47, 551–555.Find this resource:

            Ritvo, E. R., Freeman, B. J., Scheibel, A. B., Duong, T., Robinson, H., Guthrie, D., et al. (1986). Lower Purkinje cell counts in the cerebella of four autistic subjects: Initial findings of the UCLA-NSAC autopsy research report. American Journal of Psychiatry, 146, 862–866.Find this resource:

              Rodier, P. M., Ingram, J. L., Tisdale, B., Nelson, S., & Romano, J. (1996). An embryological origin for autism: Developmental anomalies of the cranial nerve motor nuclei. Journal of Comparative Neurology, 370, 247–261.Find this resource:

              Scharre, J. E., & Creedon, M. P. (1992). Assessment of visual function in autistic children. Optometry and Visual Science, 69, 433–439.Find this resource:

              Sears, L. L., Finn, P. R., & Steinmetz, J. E. (1994). Abnormal classical eye-blink conditioning in autism. Journal of Autism and Developmental Disorders, 24, 737–751.Find this resource:

              Schneider, T., Labuz, D., & Przewlocki, R. (2001). Nociceptive changes in rats after prenatal exposure to valproic acid. Polish Journal of Pharmacology, 53, 531–534.Find this resource:

              Schneider, T. & Przewlocki, R. (2005). Behavioral alterations in rats prenatally exposed to valproic acid. Neuropsychopharmacology, 30, 80–89.Find this resource:

              Schneider, T., Ziolkowska, B., Gierik, A., Tyminska, A., & Przewlocki, R. (2007). Prenatal exposure to valproic acid disturbs the enkephalinergic system functioning, basal hedonic tone, and emotional responses in an animal model of autism. Psychopharmacology, 193, 547–555.Find this resource:

              Schopler, E. Reichler, R. J., DeVellis, R. F., & Daly, K. (1988). Towards objective classification of childhood autism: Childhood Autism Rating Scale (CARS). Journal of Autism and Developmental Disorders, 10, 91–103.Find this resource:

              Schumacher, H. J., Terpane, J., Jordan, R. L., & Wilson, J. G. (1968). The teratogenic activity of a thalidomide analogue, EM12, in rabbits, rats, and monkeys. Teratology, 5, 233–240.Find this resource:

              Sparrow, S. S., Bella, D. A., & Cicchetti, D. V. (1984). A manual for the Vineland. Circle Pines, MN: American Guidance Services.Find this resource:

                Stanton, M. E., Erwin, R. J., Rush, A. N., Robinette, B. L., & Rodier, P. M. (2001). Eyeblink conditioning in autism and a developmental rodent model [Abstract]. Neurotoxicology and Teratology, 23, 297.Find this resource:

                  Stanton, M. E., Peloso, E., Brown, K. L., & Rodier, P. M. (2007). Discrimination learning and reversal of the conditioned eyeblink reflex in a rodent model of autism. Behavioral Brain Research, 176, 133–140.Find this resource:

                  Steg, J. P., & Rapoport, J. L. (1975). Minor physical anomalies in normal, neurotic, learning disabled, and severely disturbed children. Journal of Autism and Childhood Schizophrenia, 5, 299–307.Find this resource:

                  Stephens, T. D., Bundy, C. J., & Fillmore, B. J. (2000). Mechanism of action in thalidomide teratogenesis. Biochemical Pharmacology, 59, 1489–1499.Find this resource:

                  Streissguth, A. P., Aase, J. M., Clarren, S. K., Randels, R. N., LaDue, R., & Smith, D. W. (1991). Fetal Alcohol Syndrome in adolescents and adults. Journal of the American Medical Association, 265, 1961–1967.Find this resource:

                  Streissguth, A. P., Bookstein, F. L., Barr, H. M, Press, S., & Sampson, P. D. (1998). A fetal alcohol behavior scale. Alcoholism: Clinical and Experimental Research, 22, 325–333.Find this resource:

                  Strömland, K. & Miller, M.T. 1993. Thalidomide embryopathy: Revisited 27 years later. Acta Ophthalmologica, 71, 238–245.Find this resource:

                  Strömland, K., Nordin, V., Miller, M., Akerstrom, B., & Gillberg, C. (1994). Autism in thalidomide embryopathy: A population study. Developmental Medicine and Child Neurology, 36, 351–356.Find this resource:

                  Sulik, K. K., Johnston, M. C., Daft, P. A., Russell, W. F., & Dehart, D. B. (1986). Fetal alcohol syndrome and DiGeorge anomaly: Critical ethanol exposure periods for craniofacial malformations as illustrated in an animal model. American Journal of Medical Genetics, Suppl. 2, 97–112.Find this resource:

                  Tischfield, M. A., Bosley, T. H., Salih, M. A., Alorainy, I. A., Sener, E. C., Nester, M. J., et al. (2005). Homozygous HOXA1 mutations disrupt human brainstem, inner ear, cardiovascular and cognitive development. Nature Genetics, 37, 1035–1037.Find this resource:

                  Ueda, K., Nishida, Y., Oshima, K., & Shepard, T. H. (1979). Congenital rubella syndrome: Correlations of gestational age at time of maternal rubella with type of defect. Journal of Pediatrics, 94, 763–765.Find this resource:

                  Verzijl, H. T., van den Helm, B., Veldman, B., Hamel, B. C., Kuyt, L. P., Padberg, G. W., et al. (1999). A second gene for autosomal dominant Mobius syndrome is localized to chromosome 10q, in a Dutch family. American Journal of Human Genetic, 65, 752–756.Find this resource:

                  Walker, H. A. (1977). Incidence of minor physical anomaly in autism. Journal of Autism and Childhood Schizophrenia, 7, 165–176.Find this resource:

                  Wassink, T. H., Piven, J., Vieland, V. J., Huang, J., Swiderski, R. E., Pietila, J., et al. (2001). Evidence supporting WNT2 as an autism susceptibility gene. American Journal of Medical Genetics, 105, 406–413.Find this resource:

                  White, J. C., Highland, M., Kaiser, M., & Claggett-Dame, M. (2000). Vitamin A deficiency results in the dose-dependent acquisition of anterior character and shortening of the caudal hindbrain of the rat embryo. Developmental Biology, 220, 263–264.Find this resource:

                  Whitney, E. R., Kemper, T. L., Bauman, M. L., Rosene, D. L., & Blatt, G. J. (2008). Cerebellar Purkinje cellsare reduced on a subpopulation of autistic brains: Astereological experiment using calbindin-D28k. Cerebellum, 7, 406–416.Find this resource:

                  Williams, P. G., & Hersh, J. H. (1997). A male with fetal valproate syndrome and autism. Developmental Medicine and Child Neurology, 39, 632–634.Find this resource:

                  Williams, P. G., King, J., Cunningham, M., Stephan, M., Kerr, B., & Hersh, J. H. (2001). Fetal valproate syndrome and autism: Additional evidence of an association. Developmental Medicine and Child Neurology, 43, 202–206.Find this resource:

                  Wiltse, J. (2005). Mode of action: Inhibition of histone deacetylase, altering WNT-dependent gene expression, and regulation of beta-catenin—developmental effects of valproic acid. Critical Reviews in Toxicology, 35, 727–738.Find this resource:

                  Woodruff-Pak, D. S., & Papka, M. (1996a). Alzheimer’s Disease and eyeblink conditioning: 750 ms trace vs. 400 ms delay paradigm. Neurobiology of Aging, 17, 397–404.Find this resource:

                  Woodruff-Pak, D. S., & Papka, M. (1996b). Huntington’s disease and eyeblink classical conditioning: Normal learning but abnormal timing. Journal of the International Neuropsychology Society, 2, 323–334.Find this resource: