Sensory and Motor Development
There are fewer empirical studies about sensory and motor development in children with autism than studies of other aspects of development. However, the evidence converges to confirm the existence of sensory and motor difficulties for many children with autism at some point in their development, although there is much variability in the specific symptoms or patterns expressed (Dawson and Watling, 2000). In this chapter, the terms “sensory” and “sensory-perceptual” are used to refer to responses to basic sensations and perceptions, including touch, taste, sight, hearing, and smell. Because much of the research is based on parental reports or natural observation, the characterization of these behaviors as sensory-perceptual is based on inference, which, in the long run, must be tested.
Unusual sensory-perceptual reactions appear to be manifest in some children with autism as early as the first year of life (Baranek, 1999a; Dawson and Watling, 2000). These types of behaviors appear neither universal nor specific to the disorder of autism, and there are no longitudinal studies systematically documenting developmental trajectories of these behaviors from infancy through childhood. However, though not well understood, sensory processing and motor patterns may be related to other aberrant behaviors and core features of the disorder; thus, these patterns may have implications for early diagnosis and intervention.
CONSTRUCTS AND DEFICITS
The majority of infants with autism attain basic motor milestones essentially “on time” (Johnson et al., 1992; Lord et al., 1997; Rapin, 1996a). Although parental report of motor delays in infants with autism was low (9%), reports of motor delays and clumsiness in more complex skills increased to 18 percent during the preschool and school years (Ohta et al., 1987).
Some recent evidence suggests that, based on parents’ reports, sensory-perceptual abnormalities may be among some of the first signs of autism. Lack of responsiveness to certain sounds, hypersensitivity to the taste of foods, and insensitivities to pain are more commonly seen in infants with autism than in typical infants or infants with other developmental disorders (Hoshino et al., 1982). More recently, retrospective parents’ reports of the presence of unusual sensory behaviors (e.g., strange response to sounds, atypical interest in visual stimuli, overexcitement when tickled, unusual visual behavior), and some play behaviors (e.g., play limited to hard objects), discriminated between children with autistic spectrum disorders and typical children during infant and toddler ages (Dahlgren and Gillberg, 1989; Gillberg et al., 1990).
Converging evidence, based on retrospective home video studies, demonstrates very early nonspecific sensory and motor difficulties in infants later diagnosed with autism. Stereotypic behaviors, under- and overreactions to auditory stimuli, unusual postures, and unstable visual attention were found to be characteristic of infants with autism, compared with those with other developmental disorders or with typical children. In addition, autistic symptoms observed during the first year persisted into the second year of life (Adrien et al., 1992, 1993). In another study, poor responsiveness to visual stimuli in the environment, excessive mouthing behaviors, decreased responsiveness to sound (e.g., name being called), and aversion to social touches were found to be characteristic of infants with autism. However, unusual motor posturing and repetitive motor behaviors were not more common in children with autism than in other children, and visual fixations, reduced level of affective range, and stereotyped object play were more generally characteristic of the group with other developmental disorders than of the children with autism, contrary to original hypotheses (Baranek, 1999b). Other researchers using retrospective videotape analyses have not found early sensorimotor abnormalities in children with autistic spectrum disorders (Mars et al., 1998; Osterling and Dawson, 1999; Werner et al., 2000).
Sensorimotor Deficits in Autism
Several recent studies comparing children with autism and children with other developmental disorders have concluded that the prototypical developmental profile for children with autism is one of motor skills that are relatively more advanced than social skills, even when all are delayed (DeMyer et al., 1972; Klin et al., 1992; Stone et al., 1999). Early hand-eye coordination significantly predicted later vocational skills and independent functioning, while earlier fine motor skills predicted later leisure skills (Martos-Perez and Fortea-Sevilla, 1993).
Although the basic motor skills of children with autism are often reported to be an area of relative strength, numerous studies also provide evidence that motor problems may sometimes be quite significant. Specific deficits have been reported, including in motor imitation, balance, coordination, finger-to-thumb opposition, speech articulation, and the presence of hypotonia. No significant differences were found in tactile perception or gait, beyond that accounted for by cognitive level (Jones and Prior, 1985; Rapin, 1996b; Stone et al., 1990).
Imitation skills have been a focus of study in autism. They have been consistently found to be impaired in children with autism, and deficits in imitation were found in more than 60 percent of a large longitudinal cohort (Rapin, 1996b). Imitation of body movements was more impaired than object imitation skills in young children with autism; imitation of body movements predicted later expressive language skills, and imitation of actions with objects was associated with later play skills (Stone et al., 1997). Specific gesture imitation was deficient in children with autistic spectrum disorders, although it did not account for all of the motor coordination deficits. Vocabulary size and accuracy of sign language in autistic children correlated highly with their performance on two measures of apraxia and with their fine motor age scores (Seal and Bonvillian, 1997). In addition, praxis deficits may also be present in children with autism during goal-directed motor tasks that do not require imitation (Hughes and Russel, 1993; Smith and Bryson, 1998). Deficits in oral-motor praxis, including poor range of movements, isolation of movement, and awkward execution, were also noted in children with autism given both verbal and imitative prompts (Adams, 1998; Rapin, 1996b)
Adolescents with Asperger’s Disorder and high-functioning autism showed average to above average performance in simple motor tasks, but had impairments in skilled motor tasks (Minshew et al., 1997). Both groups showed similar problems with coordination (Ghaziuddin et al., 1994). the performance of children with autism on goal-directed motor tasks was better in purposeful contexts than in nonpurposeful conditions (Hughes and Russel, 1993; Rogers et al., 1996; Stone et al., 1997).
Muscle Tone, Postural Stability, and Motor Control
Although one study by Jones and Prior (1985) found no significant differences in muscle tone between older children with autism and mental-age-matched typical children, other researchers have reported such differences in children with autism spectrum disorders (Rapin, 1996b). Children with autism were posturally more unstable than typical children, and they were less sensitive to visually perceived environmental motion. They also displayed unusual reactions to vestibular tasks (Gepner et al., 1995; Kohen-Raz et al., 1992). Children with autism relied on proprioceptive feedback over visual feedback to modulate goal-directed motor actions, including reaching and placing objects under conditions that required adaptation to the displacement of a visual field by prisms. This finding might be indicative of a perceptual deficit resulting in poor visual control and visual sequential processing (Masterton and Biederman, 1983). Although vestibular mechanisms may be generally intact and postural responses adequate under some conditions, postural mechanisms may be more compromised in children with autism when integration of visual-proprioceptive, vestibular functions, and motor skills is required.
Prevalence of Atypical Sensory Responses and Motor Stereotypies
Standardized behavioral examinations demonstrated that the overwhelming majority of children with autistic spectrum disorders displayed atypical sensorimotor behaviors at some point during the toddler or preschool years, including both heightened sensitivities or reduced responsiveness across sensory modalities, and motility disturbances such as stereotypies (Ermer and Dunn, 1998; Kientz and Dunn, 1997; Rapin, 1996b). Unusual sensory and motor behaviors included but were not limited to failing to respond to sounds (81%), heightened sensitivity to loud noises (53%), watching hands or fingers (62%), and arm flapping (52%) (Volkmar et al., 1986). Hand-finger mannerisms, whole body mannerisms other than rocking, and unusual sensory interests, as recorded on the Autism Diagnostic Interview, discriminated children with autism from those with other developmental delays (Le Couteur et al., 1989; Lord et al., 1994). A pattern of atypical sensory modulation and motor behaviors, including rubbing surfaces, finger flicking, body rocking, and absence of responses to stimuli, was present in almost 60 percent of one cohort (in 15% to a severe degree) (Rapin, 1996b). This pattern similarly distinguished children with autistic spectrum disorders from children with other developmental disorders, even those with very low developmental levels (Adrien et al., 1987; Rapin, 1996b).
Some studies report pronounced individual differences and suggest subtypes based on patterns that include unusual sensory or motor behaviors, in addition to social and communicative differences (Eaves et al.,
1994; Greenspan and Wieder, 1997; Stone and Hogan, 1993; Wing and Gould, 1979). In general, attempts have not yet been made to replicate these findings across studies. However, the variability in motor and sensory processing symptoms in children with autism, like other domains of development, may be related to developmental factors such as age or developmental stage. For example, few stereotypies are reported by parents of very young children with autism. Repetitive behaviors and mannerisms became more common in the same children after age 3 (Cox et al., 1999; Lord, 1995; Stone et al., 1997). Others have found no differences in sensory manifestations across ages or autism severity levels in school-aged children with autism (Kientz and Dunn, 1997). Maturational factors may affect sensory responses differently at varying developmental periods in children with autism.
Sensory and Arousal Modulation
Some of the unusual sensory processing and motor patterns seen in autism have been thought to result from problems in arousal modulation or habituation that result in withdrawal, rejection, or lack of response to sensory stimuli. Both physiological overarousal to novel events and underarousal and slower rates of habituation have been reported in children with autism (Hutt et al., 1964; James and Barry, 1984; Kinsbourne, 1987; Kootz and Cohen, 1981; Kootz et al., 1982; Rimland, 1964; Zentall and Zentall, 1983). A pattern of sensory rejection of external stimuli was associated with higher levels of arousal on measures of blood pressure, heart rate, and peripheral vascular resistance in children with autism, which was greatest in lower functioning children; this finding was attributed to problems in filtering and modulating responses to novelty (Cohen and Johnson, 1977; Kootz et al., 1982). It has been theorized that unpredictable and complex tasks may increase arousal modulation difficulties in both social and nonsocial situations (Dawson and Lewy, 1989).
Some studies suggest that physiological abnormalities relate to the bizarre behavioral symptoms seen in children with autism, particularly the need to preserve sameness. The children may be more sensitive to the environment and may use behavioral strategies, such as avoiding environmental change and social interaction, as methods of reducing further disorganizing experiences. In particular, tactile hypersensitivies were found to be related to behavioral rigidities (Baranek et al., 1997). Other studies have found no evidence of overarousal, and some have found evidence of underarousal (James and Barry, 1980; Corona et al., 1998). The overall circadian regulation of cortisol production, a physiologic marker of response to stress, was not found to be significantly different in autistic children; however, a tendency toward cortisol hypersecretion during school hours was found, and it appeared to be an environmental stress response (Richdale and Prior, 1992).
Visual and Auditory Systems
The presence of unusual visual or auditory behaviors has been consistently reported in children with autism (Dahlgren and Gillberg, 1989; Gillberg et al., 1990). Even though visual-spatial skills (e.g., completing complex puzzles) are a relative strength, many children with autism demonstrate unusual visual-spatial behaviors, such as visual stereotypies, atypical interest in visual stimuli, or unusual visual gaze behavior. Many children with autism exhibit exaggerated sensitivity to common environmental noises, such as dishwashers, hairdryers, and garbage disposals. This hypersensitivity is also evident when children are in a busy or crowded area, such as an amusement park.
There is little rigorous research on intervention techniques for the sensory symptoms of children with autistic spectrum disorders. In general, the quality of research in the existing assessments of the efficacy of sensorimotor interventions for autistic spectrum disorders has been relatively strong in external validity and the selection and definition of samples (see Figure 1–2 in Chapter 1) in comparison with studies in other areas. Because this area is small, the few published studies that included random assignment represent a relatively high proportion of the literature. Criteria for internal validity, including the use of prospective methods and evaluation of blind procedures, were also met for a higher proportion of published studies in the sensorimotor area than any other area than communication (as shown in Figure 1–1 in Chapter 1), though, as in other areas, 50 percent of studies did not meet fairly minimal standards in this area (see Box 1–1 in Chapter 1). Studies that included measures of generalization were very rare (see Figure 1–3 in Chapter 1). The limited consideration of generalization is also of concern, but one that holds true, though to a lesser extent, for other areas of research. Overall, it is clear that high quality research can be done in this area, but that it is very rare, and many widely publicized treatments have not received careful, systematic study. Thus, the following discussion of these methods must be primarily descriptive (see Dawson and Watling  and Goldstein  for a recent review and commentary).
Sensory Integration Therapy
By focusing a child on play, sensory integration therapy emphasizes the neurological processing of sensory information as a foundation for learning of higher-level skills (Ayres, 1972). The goal is to improve subcortical (sensory integrative) somatosensory and vestibular functions by
providing controlled sensory experiences to produce adaptive motor responses. The hypothesis is that, with these experiences, the nervous system better modulates, organizes, and integrates information from the environment, which in turn provides a foundation for further adaptive responses and higher-order learning. Other components of the classical sensory integration model include a child-centered approach, providing a just-right challenge (scaffolding) with progressively more sophisticated adaptive motor responses and engaging the child in meaningful and appropriate play interactions.
There is a paucity of research concerning sensory integration treatments in autism. In one retrospective study, children with autism who had average to hyperresponse patterns to sensory stimuli tended to have better outcomes from sensory integration therapy than did those with a hyporesponsive pattern (Ayres and Tickle, 1980). Recently, some children with autistic spectrum disorders studied prospectively during sensory integration therapy showed significant improvements in play and demonstrated less “non-engaged” play. Only one child had significant improvements with adult interactions, and none had improved peer interactions (Case-Smith and Bryan, 1999).
Other approaches based on sensory integration therapy include the “sensory diet,” in which the environment is filled with sensory-based activities to satisfy a child’s sensory needs. The “alert program’” (usually with higher-functioning individuals) combines sensory integration with a cognitive-behavioral approach to give a child additional strategies to improve arousal modulation. No empirical studies of these approaches were identified for children with autism or related populations.
Sensory stimulation techniques vary but usually involve passive sensory stimulation; they are incorporated within the broader sensory integration programs or used in isolation. The underlying proviso is that a given sensory experience may facilitate or inhibit the nervous system and produce behavioral changes, such as arousal modulation. Examples of this approach include “deep pressure” to provide calming input by massage or joint compression or using an apparatus such as a weighted vest. Vestibular stimulation, another example, is often used to modulate arousal, facilitate postural tone, or increase vocalizations. These interventions have also not yet been supported by empirical studies.
Auditory Integration Therapy
Auditory integration therapy for autism has received much media attention in recent years. Proponents of auditory integration therapy suggest that music can “massage” the middle ear (hair cells in the cochlea), reduce hypersensitivities and improve overall auditory processing ability. Two philosophical approaches to auditory integration therapy
exist: Tomatis and Berard (the latter is more common in the United States). In both approaches, music is input through earphones with selected frequencies filtered out. Although improved sound modulation is one goal of treatment, other behaviors, including attention, arousal, language, and social skills, are also hypothesized to be enhanced. In children with learning disabilities given the Tomatis approach to therapy, no positive gains were noted in comparison with a placebo approach (Kershner et al., 1990). In a pilot study of the Berard auditory integration therapy method, children with autism demonstrated fewer auditory problems and aberrant behaviors than children who received no treatment, and there was no evidence of a reduction in sound sensitivity after treatment (Rimland and Edelson, 1995).
More recent studies noted no differences in responses to auditory integration therapy in children with autism or controls (Best and Miln, 1997; Gillberg et al., 1997). One study noted significant but equal amounts of improvement on all measures for children with autism and for a control group who listened to music (Bettison, 1996). In that study, treatment effects were not related to auditory integration therapy but may have been related to general auditory desensitization or simply to placebo effects. A recent review noted that for children treated with auditory integration therapy, objective electrophysiologic measures failed to demonstrate differences in hearing sensitivity between children with autism and controls, thereby questioning the overall premise of auditory integration therapy (Gravel, 1994).
A variation of auditory training programs applied to autism includes acoustic intervention: by using human voice instead of music, in theory, the stimulation alternatively challenges and relaxes the middle ear muscles to improve speech perception (Porges, 1998). Although acoustic intervention is currently undergoing some scientific experiments in children with autism, no empirical data are available to support this approach.
In summary, auditory integration therapy has received more balanced investigation than has any other sensory approach to intervention, but in general studies have not supported either its theoretical basis or the specificity of its effectiveness.
A variety of visual therapies (including oculomotor exercises, colored filters, i.e., Irlen lenses, and ambient prism lenses) have been used with children with autism in attempts to improve visual processing or visual-spatial perception. There are no empirical studies regarding the efficacy of the use of Irlen lenses or oculomotor therapies specifically in children with autism. Prism lenses are purported to produce more stable visual
perception and improved behavior or performance by shifting the field of vision through an angular displacement of 1 to 5 degrees (base up or base down). Only one study investigated the use of prism lenses in children with autism, almost half of whom also had strabismus (Kaplan et al., 1996; Kaplan et al., 1998). Results indicated some short-term positive behavioral effects with less improvement at later follow-up. Performance on orientation and visual-spatial tasks was not significantly different between conditions. As with auditory integration therapy, studies have not provided clear support for either its theoretical or its empirical basis.
SENSORY AND MOTOR DEVELOPMENT AND EDUCATIONAL PROGRAMMING
Motor development plays an important role in learning—young children typically use motor skills to explore the environment, engage in social interactions, engage in physical activities, and develop basic academic skills, such as handwriting. Unusual sensory responses (e.g., hypo-and hyperresponses, preoccupations with sensory features of objects, paradoxical responses to sensory stimuli) are common concerns in children with autistic spectrum disorders. Given that most educational environments involve many sensory demands (e.g., the noise level of a regular classroom) and stimuli that may seem unpredictable (e.g., fire alarms), interventions may need to address the individualized sensory processing needs of children who have such difficulties. However, exactly how this should be done has not been addressed in scientific investigations.
Praxis is an area of particular interest; several studies note that both younger and older children with autism may demonstrate difficulties with aspects of motor planning. These difficulties are exaggerated in tasks that require execution of a social imitation, either motor or object related, but they may also be present in non-imitated, simple, goal-directed motor tasks. Such difficulties would affect many daily aspects of early childhood, such as games and sports (e.g., throwing a ball, riding a bicycle), crafts (e.g., using scissors), and performing gestures. Although it is possible that the formulation of motor plans is deficient in children with autistic spectrum disorders, it is also possible that simple motor planning is intact but that the use of externally guided visual feedback is diminished. If so, the quality of motor control, postural stability, and effective sequencing would all be affected.
Classical sensory integration therapy provides a child-centered and playful approach that is often appealing to even the most unmotivated or disengaged child. In the case of the other treatments based on sensory integration, a child must be able to tolerate various sensory applications or physical manipulations. For some children with autism, structure and
repetition are positive factors whereas passive application of stimulation may not be.
In general, interventions based in natural environments that teach or attempt to change behaviors in the context in which they would typically occur have been found to be most effective (see Chapters 10, 11, and 12). Thus, ways of helping children with autistic spectrum disorders cope with unusual sensory responses within their ordinary environments or modifications to these environments might be expected to have more effects than would specific, one-to-one therapies (e.g., individual sensory integration treatment or individual sensory diets) or group treatments with unique stimuli (e.g., auditory integration therapy). This is particularly likely, given the many questions that arise about the theoretical bases for these sensory interventions. However, even if the results of sensory approaches are not specific, children may benefit from techniques that elicit social engagement, attention, and the use of toys and other materials at home and within classroom settings.
FROM RESEARCH TO PRACTICE
There is no consistent evidence that sensory-based treatments have specific effects; in many cases, the theories underlying such approaches have not withstood careful consideration (Dawson and Watling, 2000; Goldstein, 1999). A lack of empirical data does not necessarily demonstrate that a treatment is ineffective, but only that efficacy has not been objectively demonstrated (Rogers, 1998). There were some nonspecific positive findings in the studies of interventions reviewed, and there is a need to address at least functional aspects of motor difficulties, particularly as they affect social, adaptive, and academic functioning.
Future research in these areas needs to include well controlled, systematic studies of effectiveness. Only such research can answer not only what is effective, but with whom and under what conditions. Because most sensory- and motor-oriented interventions augment comprehensive educational programs, it is critical to know whether or not these approaches facilitate progress as additional interventions or hinder it by taking away valuable instruction time. It will be important to investigate to what degree specific treatments can be altered to fit an inclusive education model while still retaining their essential therapeutic elements and purported benefits. Comparisons of such treatments need to be systematically investigated in future efficacy research.