Efficacy of Sensory and Motor Interventions for Children with Autism

Efﬁcacy of Sensory and Motor Interventions for Children with Autism

Grace T. Baranek1

Idiosyncratic responses to sensory stimuli and unusual motor patterns have been reported clinically in young children with autism. The etiology of these behavioral features is the subject of much speculation. Myriad sensory- and motor-based interventions have evolved for use with children with autism to address such issues; however, much controversy exists about the efficacy of such therapies. This review paper summarizes the sensory and motor difficulties often manifested in autism, and evaluates the scientific basis of various sensory and motor interventions used with this population. Implications for education and further research are described.

KEY WORDS: Sensorimotor therapies; evidence-based treatments; sensory integration.

PURPOSE, SEARCH PROCEDURES, AND lish language for the past 30 years with a focus on the SCOPE past 10 years. Also, manual searches of key references from articles were completed. In a few noted cases, The purpose of this paper is threefold: (1) briefly studies of individuals with related disorders (e.g., men- summarize the empirical literature with respect to sen- tal retardation), or slightly older children (over 8 years) sory and motor development/abnormalities in children were included provided the results had implications for with autism, (2) evaluate the scientific basis of sensory young children with autism. and motor interventions used with children with autism, and (3) describe implications of these findings for edu- LITERATURE REVIEW OF SENSORY AND cation and further research. Subject headings and key- MOTOR DEVELOPMENT words were searched for terms related to sensory and motor deficits (e.g., arousal, sensory reactivity/pro- Empirical studies about sensory and motor devel- cessing, habituation, posture, praxis, gross/fine/oral opment in children with autism are limited compared to motor development, etc.), and categorical terms spe- studies of other aspects of development. Those that exist cific to sensory and motor interventions (e.g., sensory often suffer from a variety of methodological limitations; integration, prism lenses, etc.). Searches were con- however, these studies provide both a foundation of sci- ducted using MEDLINE, CINAHL, & PSYCINFO data- entific knowledge critical for understanding the early bases to find empirical studies specific to children with development of children with autism and guidance for autism spectrum disorders (i.e., autistic disorder, per- intervention planning. Empirical evidence converges to vasive developmental disorder) available in the Eng- confirm the existence of sensory and motor difficulties for many children with autism at some point in their early development (Adrien et al., 1987; 1992; 1993; Baranek, 1 The Clinical Center for the Study of Development and Learning, Room 111 Medical School Wing E—CB # 7120 University of North 1999; Dahlgren & Gilberg, 1989; Hoshino et al., 1982; Carolina at Chapel Hill, Chapel Hill, NC 27599-7120; e-mail: Ohta, Nagai, Hara, & Sasaki, 1987; Ornitz, Guthrie, & [email protected] Farley, 1977; Scharre & Creedon, 1992), although much 397 0162-3257/02/1000-0397/0 © 2002 Plenum Publishing Corporation 398 Baranek variability is present in specific symptoms or patterns in the absence of known peripheral dysfunction (e.g., expressed. These types of behaviors appear neither uni- hearing acuity, visual defect) per se. Thus some reversal nor specific to the disorder of autism; however, searchers have suggested shifting toward the investi- qualitative aspects of these patterns have not been well gation of more complex levels of information studied, and prospective, longitudinal investigations that processing including attentional control mechanisms systematically document developmental trajectories from and executive functions to help explain some of these infancy through childhood are yet to be accomplished. unusual sensory features or motor deficits (e.g., Lin- The majority of evidence stems from parental reports, coln et al., 1995; Minshew, Goldstein, & Seigel, 1997; which themselves are prone to some biases and method- Wainwright & Bryson, 1996). Given that many con- ological weaknesses. Empirical data from retrospective ventional educational environments are sensorily com- video studies (Adrien et al., 1992; Baranek, 1999) and plicated and unpredictable, interventions likely need to clinical evaluations (Gillberg et al., 1990) are emerging consider the individualized sensory processing needs to suggest that patterns of sensory and motor features of children demonstrating such difficulties to optimize in autism may differ qualitatively from those in other de- successful participation in such programs. velopmental disorders. Furthermore, unusual sensory- With respect to developmental milestones, uneven- perceptual features appear to be manifest quite early in ness between domains is often reported, indicating a rel- the development of children with autism (i.e., by 9 to ative sparing of general motor skills compared to 12 months of age). Though not well understood, sensory language or social skills in children with autism (Klin, processing and motor patterns may be related to core fea- Volkmar, & Sparrow, 1992; Stone et al., 1999). How- tures, development of other aberrant behaviors, and later ever, not all children with autism demonstrate prowess prognosis; thus these patterns have implications for early in motor skills and considerable variability exists (Amato diagnosis and intervention. & Slavin, 1998; DeMyer et al., 1972; Jones & Prior, Unusual sensory responses (e.g., hypo- and hyper- 1985; Johnson, Siddons, Frith, & Morton, 1992; Ohta responses; preoccupations with sensory features of ob- et al., 1987; Ornitz et al., 1977; Rapin, 1997; Rinehart, jects, perceptual distortions; paradoxical responses to Bradshaw, Brereton, & Tonge, 2001). Many demonstrate sensory stimuli) have been reported in 42 to 88% of atypical features (e.g., low muscle tone, oral-motor prob- older children with autism in various studies (Kientz & lems, repetitive motor movements, dyspraxia) or test in Dunn, 1997; LeCouteur et al., 1989; Ornitz et al., 1977; the delayed ranges on standardized motor assessments Volkmar, Cohen, & Paul, 1986), indicating that these particularly as the complexity of tasks increases. are common concerns in this population. Percentages Whether or not these difficulties are purely motoric is vary depending on how specifically items were sam- unclear, because other areas also affect test-taking abil- pled. Auditory processing problems are particularly ities. Furthermore, delayed motor development is not noted, with one study (Greenspan & Weider, 1997) pur- a unique characteristic of children with autism, because porting that 100% of subjects demonstrated these dif- it is often associated with the level of mental retarda- ficulties. Visual spatial skills are often more advanced tion in general. However, because more than 75% of than other areas of development, although individual children with autism have concomitant mental retarda- differences are noted. Sensory processing abilities also tion, the presence of motoric concerns, regardless of appear to be uneven and of a fluctuating nature in whether they are primary or secondary to autism, still autism, such that both hyper- and hypo-responses are has substantial implications for individualized educa- evident in the same child. These aberrant sensory re- tional interventions. Developmental motor delays, al- actions are thought to reflect poor sensory integration though only minimally different during infancy, may and/or arousal modulation in the central nervous sys- become magnified with progressive age (e.g., Ohta tem, although the underlying nature of these symptoms et al., 1987). Especially during early foundational years, remains speculative (e.g., neurological structures and motor skills provide a means for learning important systems involving the cerebellum, limbic system, cor- skills in other domains (e.g., social skills, academics) tical mechanisms, etc). Both patterns of under- and and thus motor-related difficulties may need to be ad- over-arousal have been reported (Hutt, Hutt, Lee, & dressed in the educational curricula or through related Ounsted, 1964; James & Barry, 1984; Kinsbourne, therapy services. At least one study (Perez & Sevilla, 1987; Kootz & Cohen, 1981; Kootz, Marinelli, & 1993) demonstrates a predictive relationship between Cohen, 1982; Rimland, 1964; Zentall & Zentall, 1983). motor skills in children with autism to functional out- Children with autism tend to show these abnormal sen- comes in other domains such as vocational and leisure sory responses to both social and nonsocial stimuli and skills 5 years later. Efﬁcacy of Sensory and Motor Interventions for Children with Autism 399

Motor planning deficits are an area of particular in- used to define the parameters for the review: (a) re- terest, given that several studies point out that both medial interventions that target specific sensory or younger and older children with autism may demonstrate motor components per se, broader performance out- difficulties with aspects of praxis (Adams, 1998; Jones comes that are thought to be the result of the sensory- & Prior, 1985; Rinehart et al., 2001; Rogers et al., 1996; motor treatment, or both; (b) compensatory skills Smith & Bryson, 1998; Stone et al., 1990). These diffi- training approaches; and/or (c) task/environmental culties are certainly exaggerated in tasks that require ex- modifications targeted for sensory and motor difficul- ecution of a social imitation, either motor or object ties. A variety of other interventions exist that may be related, but may also be present in nonimitated simple at least partially aimed at improving sensory and motor goal-directed motor tasks (e.g., reaching, grasping, and skills; however, only those interventions that have a placing) (Hughes & Russell, 1993). Motor planning primary basis in sensory processing or motor theories deficits are sometimes mistaken for general clumsiness; were included. Traditional behavioral interventions and however, Rinehart et al. (2001) separated out two com- psychopharmacological treatments were excluded. ponents of action and found that highly functioning chil- Likewise, comprehensive educational models (e.g., dren with autism spectrum disorders, ages 5 to 19 years, Greenspan—DIR Model; TEACCH) were excluded, had intact movement execution but atypical movement even though these programs frequently include a sen- preparation during a simple motor reprogramming task sory processing or motor development component. Al- compared with typical, IQ-matched controls. Specifi- though a variety of professionals (e.g., occupational cally, children with autistic disorder were characterized therapists, adaptive physical educators, physical ther- by a lack of anticipation during movement preparation apists, speech pathologists, etc.) may utilize various phases, findings suggestive of difficulties in motivational sensory-motor strategies listed within the context of aspects of behavior or attention for action. Children with any given child’s individualized educational plan, this Asperger’s disorder showed slower preparations for paper is not intended to be a critique of the efficacy of movement at phases in which movement should be those related services. Sensory-motor interventions are optimal—implicating additional dysfunction in the not presumed to represent the full scope of therapeutic/ frontal-striatal system, according to the authors. educational services offered by specialized profes- Although it is possible that the formulation of motor sionals. Sensory or motor treatments often are used as plans is deficient, it is also possible that simple motor an adjunct to a more holistic intervention plan. For planning is intact but that the use of externally guided vi- example, the occupational therapist provides therapeu- sual feedback is diminished, affecting the quality of motor tic interventions aimed at improving a child’s occupa- performance, postural stability, and the lack of effective tional performance (e.g., play, school functional skills, sequencing of actions (Masterton & Biederman, 1983; self-care) within the educational context. Remediation Gepner, Mestre, Masson, & de Schonen, 1995; Smith & of sensory or motor deficits (as well as other compo- Bryson, 1998; Stone et al., 1990; Kohen-Raz, Volkmar, nents including cognitive or psychosocial functions) & Cohen, 1992). Thus, perceptually challenging tasks may occur if indicated, but only within the larger con- that require smooth integration of visual with vestibular- text of occupational performance problems within the proprioceptive information, for example, may be partic- learning environment. Compensatory interventions and ularly difficult to perform and could result in poor quality environmental adaptations are also utilized and often of motor performance on complex tasks. These findings preferred because of the more immediate effects on taken together with evidence that motor imitation skills meaningful participation. in young children with autism predict later expressive To judge the validity of each category of inter- language skills and play skills (Stone et al., 1997), have ventions, first a description of each intervention ap- significant implications for educational interventions and proach is provided that includes the underlying future research. assumptions of the intervention, proposed mechanisms thought to be responsible for the therapeutic changes purported, and the service delivery model utilized to EFFICACY OF SENSORY AND MOTOR provide the intervention. In addition, the scientific evi- INTERVENTIONS dence (i.e., efficacy research from peer-reviewed sources) of each intervention category is reviewed. Because “interventions for sensory and motor Table I provides a comparative summary of the empir- deficits” are not synonymous with “sensory and motor ical studies reviewed for children with autism spectrum interventions,” the following inclusion criteria were disorders. Table 1. Sensory and Motor Intervention Studies

Treatment Int. Ext. Intervention Outcomes Study category a n Subjects val.b val.c Gend Design elements speciﬁcations measured Findings

Ayres & SI-classical 10 Mild-severe III IV III Retrospective case-control 2ϫ wk for one year Language, awareness of en- Subjects with hyper- Tickle (1980) autism; of treatment responders vironment, purposeful ac- responsivity to tactile 3–13 yr versus nonresponders tivities, self-stimulation., and vestibular stimuli social and emotional had better outcomes behavior than hypo-responsive subjects Case-Smith & SI-classical 5 Autism/MR; IV II I Single-subject AB design; 30 min treatment Mastery play, engaged be- General improvements in Bryan (1999) 4–5 yr 3 wk baseline, 10 wk sessions plus haviors, peer and adult mastery play, engaged treatment teacher consulta- interactions behaviors, and adult tion interactions. No significant changes in peer interactions Linderman & SI-classical 2 Autism; 3–4 yr IV IV I Single-subject AB design; 1 hr/wk in clinic Functional behavior Increase in social inter- Steward 2 wk baseline, 7 or measures varied between actions, response to (1999) 11 wk treatment subjects (e.g., social movement and affec- interactions, tion, and approach to communication) new activities Larrington SI-based 1 Autism/severe IV IV III Descriptive case study Varied sensory treat- Variety of behaviors (e.g., Positive effects reported (1987) MR; 15-yr- ment (vest./prop/ SIB, play skills, social in many domains of old boy oral motor) treat- interactions) behavior ment in school and group home; 2 yr Reilly et al. SI-based 18 Autism 6–11 yr III I III Counter-balanced alternating 2 sessions 30 min Quantity and quality of Fine motor activities (1983) treatments design; random vestibular treat- vocalizations measured elicited more variety assignment to activity ment versus 2 ses- during treatment only. of speech and mean order sions 30 min fine (No pre-post measures). length of vocalizations motor treatment and decreased autistic (alternating); pro- speech than the vided within a 3 vestibular treatment wk Stagnitti, SI-based 1 Sensory Defen- IV IV II Descrriptive case study Brushing, joint com- Anecdotal reports of tactile Improvements in all areas Raison, & (Sensory siveness and pression 3–5ϫ/day tolerance, affect, activity following treatment; Ryan Diet) Possible for 2 wk (other level, temper tantrums at benefits faded by (1999) Autism spec- techniques were home, school/community 5 months post-treat- trum; 5-yr-old also used); re- ment; 6- and 9-mo as- boy peated again at 5 sessments (following mo post treatment 2nd phase of treatment) showed sensory defen- SS-touch siveness cured Edelson et al. pressure 12 Autism 4–13 yr I I III RCT with placebo; Pre- and Hug Machine (pres- Behavioral (tension, anxi- Decreased tension and (1999) post-treatment measures sure) versus ety, hyperactivity) and anxiety for treatment placebo (no pres- physiological (galvanic group. Subjects with sure) 2ϫ 20-min skin response) initially higher anxiety sessions per week; level had better out- 6 wk total comes Field et al. SS-touch 22 Autism; mean I I III RCT with alternative treat- Touch therapy ver- Touch aversions, off-task Both groups showed (1997) therapy age 4.5 yr ment; Pre- and post-treat- sus one-to-one behaviors, orientation to positive changes; ment measures quiet play (read) sounds, stereotypic be- treatment group im- 2ϫ 15-min ses- havior, teacher report proved in response to sions per week; sounds, stereotypic be- 4 wk total havior, and social be- McClure & SS-touch 1 Autism/MR; IV IV III Descriptive case report Elastic (pressure) Self-stimulatory behaviors, haviors Holtz-Yotz pressure 13-yr-old boy wrappings ad- self-abuse, social inter- Generally positive out- (1990) in psychiatric ministered over 4 actions comes in all areas unit treatment ses- during treatment sions Ray et al. SS-vestibular 1 Autism; 9-yr- IV IV II Descriptive case report; Self-initiated Time spent vocalizing (1988) old male Pre-, mid-, and post-treat- vestibular stimu- Increased vocalizations ment measurement lation; 5-min ses- during treatment; most sions 2ϫ week prominent in first over 4 wk week Zisserman SS-touch 1 Autism; 8-yr- IV IV III Descriptive case report Pressure garments Self-stimulatory behavior Decrease in self-stimula- (1991) pressure old female (vest and gloves) (e.g., hand slapping) tory behaviors with worn in class- gloves, slight (non- room significant) decrease with vest

Bettison (1996) AIT (Berard) 80 Autism or As- I I III RCT with alternative treat- Treatment group Aberrant behavior (ABC), Similar gains noted in perger (sound ment; pre- and post mea- had AIT sensory problems, sound both groups. 75% of hypersensi- sures at 1, 3, 6, 12 mo (filtered/modu- sensitivity, IQ, language, subjects improved in tivity), ages lated music); audiometric tests. 1st month. Much vari- 3–17 yr control group had ability and attenuation structured listen- of improvements over ing treatment course of 12-mo fol- (unprocessed low-up. Two subjects music), 2 ϫ 30 had significant adverse min per day, effects 10 days Brown (1999) AIT (Berard) 2 Autism; 3.5 and IV IV III Descriptive case reports AIT 2 ϫ 30 min Descriptions of sensory, General improvements 5 yr per day, 10 days motor, and functional noted in a variety of behaviors domains (e.g., attention and speech) Gillberg et al. AIT (Berard) 9 Autism (all III III IV Pre- and post-design AIT 30 min daily, Pre and post-tests of autism No significant differ- (1997) with MR); (parents/raters aware 10 days symptoms on CARS and ences on CARS and 3–16 yr of treatment) ABC ABC; sensory problems declined slightly Link (1997) AIT (Berard) 3 Autism; ages 6, IV IV III Descriptive case reports AIT 30 min daily, Descriptions of hearing acu- One child improved; one 7, and 15 yr; (pre- and post-measures) 10 days ity, sound sensitivity, and child had mixed re- 2 nonverbal behavior at home/school sults; one child deteri- reported by caregivers orated following a seizure during AIT and was D/C on day 5

continued Table 1. Sensory and Motor Intervention Studies (Continued)

Treatment Int. Ext. Intervention Outcomes Study category a n Subjects val.b val.c Gend Design elements speciﬁcations measured Findings

Mudford et al. AIT (Berard) 16 Autism (low II I I Balanced cross-over experi- AIT versus placebo Aberrant behavior (ABC, Several drop-outs due to (2000) functioning); mental design (parents (disengaged Nisonger; observational problems tolerating ages blind to treatment condi- headphones), recordings), IQ, adaptive treatment; placebo 5–14 yr tions) 30 min 2ϫ/day, behavior (Vineland), lan- slightly more benefi- (M ϭ 9.4) 10 days; 3–5 mo guage measure (Reynell) cial (less aberrant be- baseline; treat- havior) than AIT for ment and control subjects that com- phases in random pleted treatment order for all subjects Neysmith- AIT (Toma- 6 Severe Autism; II III IV Pre- and post-design with 20 (30-min) ses- CARS scores (from inde- Pre-treatment, all sub- Roy (2001) tis) ages 4–11 yr independent evluation sions repeated pendent ratings of video- jects were “severely (repeated across subjects) for 4–8 blocks. taped play observations autistic” on CARS; Treatment phases and teacher/parent inter- After treatment 3 followed by 3–8 views) (younger) children had weeks of no reduced severity of treatment/evalua- autism; 3 older chil- tion phases dren had few changes Rimland & AIT (Berard) 445 Autism; pri- III II III [Part 1: Compared 3 AIT AIT 2ϫ 30 min per Audiograms, behavioral re- No significant differ- Edelson mary or sec- devices using RCT.] Part day for 10 days; actions to sound, parent ences among 3 AIT (1994) ondary dx; 2: Retrospective case one of three AIT reports of sound sensitiv- devices. Slightly im- 4–41 yr control (treatment group device conditions ity, aberrant behavior, proved sound sensitiv- (M ϭ 10.7) compared to no treat- (EERS, Audio- and behavior problems ity & acuity. Decreased ment, control group of 9 kinetron, or checklists audiogram variability subject from a previous Audio Tone En- associated with im- pilot study). Follow-up hancer). proved behavior. from 1 to 9 mo in 191 Age and degree of subjects. sound sensitivity not related to behavioral improvement. Rimland & AIT 18 Autism; I I III RCT w/ alt. treatment; Pre-, AIT versus alter- Audiograms; parental re- Treatment group had de- Edelson 4–21 yr mid-, and post-treatment native treatment ports of hypersensitivity creased auditory prob- (1995) measures at 2 wk and at (unprocessed to sound and aberrant be- lem behaviors and 1, 2, 3 mo post-treatment. music) 2ϫ 30 havior (stereotypy, hyper- aberrant behaviors 3 (raters and parents blind min per day for activity, speech, etc) months post-treatment. to conditions) 10 days No significant changes in sound sensitivity; however, subjects were not necessarily hyper-sensitive pre- treatment Zollweg, Palm, AIT 30 Autism/MR; I I III RCT w/ alt. treatment; pre- AIT vs. alt. treat- Audiograms, sensitivity and No significant difference & Vance ages 7–24 yr; and post-treatment at 1 ment. (un- loudness tolerance; aber- in audiological or be- (1997) M ϭ 14 yr, 5 wk, and at 1, 3, 6, and 9 processed music) rant behaviors (ABC) havioral outcomes be- mo (included mo post-treatment (dou- 2ϫ/day for 10 tween the groups; 1 AUT ble blind) days, 30 min/ control group showed Ͻ8 yr) session slightly less aberrant behavior at 6 mo post-treatment Carmody et al. Prisms 24 Autism; Ages III II III Cross-over design (within Three brief opto- A single session assessment Head posture, body pos- (2001) 3–18 yr.; subjects) with three con- metric assess- period with 60–90 sec ture, facial expression, (Median ϭ 8) ditions. ment conditions trials for each task: and ball catching in- [Independent ratings from for each subject: Recorded spatial orient- creased with correct video used to validate ini- no lens (baseline) ing behaviors (i.e., head (facilitating) prism tial optometric ratings.] versus base-up or and body posture, facial lenses as compared to base-down lenses expression) while seated incorrect and habitual (randomly or- watching TV and ball lenses dered treatment) catching performance for each subject while standing Kaplan et al. Prisms 14 Autism; ages III II III Cross-over design (within Performance de- A single-session assessment Overall posture and per- (1996) 4–15 yr; subjects) with three con- scribed under period with 60–90 sec formance was better (Mean ϭ ditions. three assessment trials for each task. with correct prism 8 yr) [Independent ratings by two conditions: no Recorded ratings on vi- lenses for group as a authors made later from lens (baseline) sual-spatial orientation whole video to validate initial versus base-up or behaviors (i.e., head posi- optometric ratings.] base-down lenses tion, body posture, facial (randomly or- expression) during vari- dered treatment) ous lab tasks (i.e., watch for each subject TV seated and standing on balance board; ball catch game). Kaplan et al. Prisms 18 Autism/PDD; I II III Double-blind cross-over de- Subjects matched/ Measured behavior prob- Improved behavior in (1998) ages 7–18 yr; sign with placebo control randomly as- lems (Aberrant Behavior treatment condition (Mean ϭ signed; 1⁄2 re- Checklist); visual-spatial (short term). No sig- 11.5 yr) ceived placebo orientation and attention nificant differences be- (39% with (clear lens); 1⁄2 pre-, mid, and post-treat- tween tx and control strabismus) got treatment ment for both treatment conditions for postural (ambient prism and control conditions. orientation and atten- lens) conditions tion. No significant for 3–4 mo; then differences between conditions were children with and reversed without strabismus; trend for more behavior problems with placebo lenses.

continued Table 1. Sensory and Motor Intervention Studies (Continued)

Treatment Int. Ext. Intervention Outcomes Study category a n Subjects val.b val.c Gend Design elements speciﬁcations measured Findings

Neman et al. Patterning 66 Institutionalized I I III RCT with alt treatment Doman-Delcato Cognitive performance, No dramatic changes in (1974) subjects with (physical activities with method patterning developmental tests for individuals; no signifi- MR; (M ϭ sensory stimulation and treatment 2 vocabulary, visual- cant changes in IQ or 15 yr) attention) and no-treat- hr/day 5ϫ per perceptual, and motor motor skills. Signifi- ment control group week for 2.5 mo, skills cant increases for then 7ϫ per wk treatment group in two for 4 mo categories most related to treatment skills taught; visual compe- tence and mobility. No treatment control fared worst on most measures Bridgman et al. Patterning 12 Various DD— III II III Prospective cohort design Doman-Delcato Developmental profile, IQ Treatment group slightly (1985) CP, MR, (nonrandom group as- method patterning with Bayley mental scale better in language and autism/ signments); pre, mid- 8 hr per day; con- or Stanford-Binet (mea- socialization at 3 mo seizure (M ϭ (3 and 6 mo) and post- trol group re- surements not uniformly but findings were 8 yr) treatment measurements ceived Special applied) short-lived and no sig- (at 10 mo) Education ser- nificant differences vices; 10 mo found at 10 mo. Kern et al. Exercise 7 Autism 4–14 yr II II III Repeated reversal design 5–20 min structured Percent of self-stimulation Consistent decrease in (1982) (ABAB); 45 reversals; individual jog- as well as number of cor- self-stimulatory behav- Multiple pre- and post- ging sessions; rect responses on acade- iors and improvements treatment measures variable settings mic (matching task) and in academic and play (e.g., clinic and ball playing (catching responses following home) and vari- task). treatment able duration of treatment 4–17 days Kern et al. Exercise 3 Autism 7–11 yr II II IV Counter-balanced, alternat- 15 min vigorous ac- Time sampling of stereo- Reduction in self-stimu- (1984) ing treatments design. tivity (jogging) typed behaviors latory behaviors after Baseline and multiple followed by 15 vigorous jogging post-treatment measures min mild activity exercise. (ball playing) and return to 1st condition; condition order reversed on second day Levinson & Exercise 3 Autism (low II II IV Counterbalanced, alternat- 15 min group Heart rates, individualized Reduction in self-stimu- Reid functioning) ing treatments design; walking (mild measures of self-stimula- latory behaviors after (1993) 11 yr multiple pre-, during, exercise) for tory behaviors vigorous exercise post-treatment measure- 5 sessions versus treatment. Effects wear ments 15 min individual off after 90 min jogging (vigorous exercise) ϫ 4 sessions across 9 wk Watters & Exercise 5 Autism; ages II II IV Alternating treatments de- Compared three Rates of self-stimulatory Decrease in self-stimula- Watters 9–11 yr sign with three pre-task conditions: 8–10 behaviors and academic tory behaviors follow- (1980) conditions in randomized min jogging ver- performance during a lan- ing exercise condition order. Pre- and post-treat- sus 15 min TV guage training session ment measures viewing versus post treatment varied academic tasks; given 1 session per day, 1–4 ϫ per week (total weeks varied)

KEY: aTreatment Category Codes ● SI, Sensory Integration Therapy (classical) ● SI-based, Sensory-Integration Based Approach ● SS, Sensory Stimulation Technique ● Patterning, Sensorimotor Patterning ● AIT, Auditory Integration Training ● Prism, Ambient Prism Lenses (Visual Therapy) ● Exercise, Exercise Therapy bInternal Validity Classification Criteria: ● I: Prospective study comparing treatment to alternative or placebo (e.g., RCT) where evaluators are blind to treatment status. ● II: Multiple baseline, ABAB, reversal/withdrawal with measurement of outcome blind to treatment conditions, or prepost design with indep. evaluation. ● III: Prepost or historical designs or multiple baseline, ABAB, reversal/withdrawal (not “blind”). ● IV: Other (e.g., single subject [AB or ABA] designs without multiple baselines; case study reports). cExternal Validity Classification Criteria: ● I: Random assignment of well-defined cohorts and adequate sample size for comparisons. ● II: Nonrandom assignment, but well-defined cohorts with inclusion/exclusion criteria and documentation of attrition/failures. In addition, adequate sample size for group designs or replication across three or more single subjects. ● III: Well-defined population of three or more subjects in single-subject designs or sample of adequate size in group designs. ● IV: Other. dGeneralization Classification Criteria: ● I: Documented changes (i.e., generalization) in at least one natural setting outside of treatment setting. ● II: Generalization to one other setting or maintenance beyond experimental intervention in natural setting in which intervention took place. ● III: Intervention occurred in natural setting or use of outcome measures with documented relationship to functional outcome. ● IV: Not addressed or other. 406 Baranek

Sensory Integration Therapy environmental adaptations are often provided in tandem with the direct intervention. Cost varies depend- Description and Assumptions ing on the location, frequency, and duration of the Sensory integration (SI) therapy, based on the treatment, but it is comparable to hourly rates of other work of Dr. A. Jean Ayres, is intended to focus directly therapy services. Equipment is generally low-tech, but on the neurological processing of sensory information can be moderately expensive. The feasibility of doing as a foundation for learning of higher-level (motor or classical SI in a school setting is low because of the academic) skills. Some of the neurological assumptions need for specialized equipment and the “pull-out” upon which this model is based (i.e., hierarchically or- model that may conflict with inclusionary principles. ganized nervous system) have received criticism as being outdated; recent theorists are reconceptualizing Empirical Studies this theory (Bundy & Murray, 2002). The assumption that sensory experiences have an effect on learning is Three studies investigated interventions that less controversial, although the mechanisms through matched the criteria of classical SI therapy. Two utilized which this occurs are somewhat ambiguous and often prospective AB designs with several subjects and ade- debated. Disruptions in subcortical (sensory integra- quate controls to look at SI treatment efficacy (Case- tive) functions are treated by providing controlled ther- Smith & Bryan, 1999; Linderman & Steward, 1999); one apeutically designed sensory experiences for a child to utilized a retrospective design to identify predictors of respond to with adaptive motor actions. Through so- positive outcomes within a group of children with autism matosensory and vestibular activities actively con- receiving SI (Ayres & Tickle, 1980). Although outcome trolled/sought out by the child, the nervous system is measures included aspects of proximate (sensory pro- thought to be able to better modulate, organize, and in- cessing/modulation) functions and/or broader outcomes tegrate information from the environment, which in turn (e.g., play skills, social interactions), none of the studies provides a foundation for further adaptive responses directly attempted to remediate deficits in praxis. and higher-order learning. Other necessary components Case-Smith & Bryan (1999) studied five boys of the classical SI model include a child-centered ap- across a 3-week baseline phase and a 10-week inter- proach, providing a just-right challenge (scaffolding), vention that consisted of a combination of classical SI facilitating progressively more sophisticated adaptive treatment and consultation with teachers. Independent motor responses, and engaging the child in affectively coding of videotaped observations of free play indi- meaningful and developmentally appropriate play in- cated that three of the five boys demonstrated signi- teractions. The child’s focus is intended to be placed ficant improvements in mastery play, and four of five on the occupation of play (intrinsically motivated) and demonstrated less “nonengaged” play. Only 1 subject not on cognitive-behavioral strategies or repetitive had significant improvements with adult interactions, drills—as is the focus of other sensorimotor and be- and none changed in level of peer interaction. Outcome havioral approaches. Treatment goals may center on measures more directly related to intrinsic features of improving sensory processing to either (a) develop bet- the intervention (e.g., individual mastery play) ap- ter sensory modulation as related to attention and be- peared more improved than measures that were not havioral control, or (b) integrate sensory information directly addressed in treatment (e.g., peer interaction). to form better perceptual schemas and practic abilities Although it is possible that the positive results could as a precursor for academic skills, social interactions, be attributed to factors other than the intervention it- or more independent functioning. self (i.e., maturation, caregiving effects), the authors note that the behaviors did not change systematically across all outcome measures. However, because sen- Service Delivery Model and Approach sory processing variables were not directly assessed, it SI therapy is classically provided utilizing a direct is not known whether the positive outcomes are directly one-on-one intervention model in a clinic environment due to improvements in sensory processing mechanisms that requires specialized equipment (e.g., suspended per se, as would be purported by sensory integration swings). Treatment plans are designed individually and theory. It is also possible that the improvements evi- carried out by a trained therapist (OT) approximately denced are a function of other components of the in- 1 to 3 times per week, 1-hour sessions. Duration typi- tervention (e.g., play coaching, motivational strategies). cally entails several months and in some cases years. Linderman & Steward (1999) also utilized a sin- Consultative services, home/school programs, or task/ gle subjects AB design with two 3-year-olds with per- Efficacy of Sensory and Motor Interventions for Children with Autism 407 vasive developmental disorder (mild autism) to track categories are stronger than in other areas, at least for the functional behavioral changes in the home envi- SI studies compared to no treatment conditions; how- ronment associated with classical SI (clinic-based ever, effects appeared to be equivocal when compared 1 hour per week for 7 to 11 weeks). Subject 1 (who was with alternative treatments. noted to have tactile hypersensitivity) demonstrated gains in all intended outcomes (social interactions, ap- Other Sensory Integration-Based Approaches proach to new activities, response to holding). Subject 2 (who had both hypo-responsiveness to vestibular and Description and Assumptions hyper-responsiveness to tactile sensations) made gains This section groups several types of approaches in activity level and social interaction, but not in func- that are based on the foundations of SI theory, but de- tional communication. This study purports that SI may viate from classical SI in one or more criteria: (a) so- result in some functional gains that generalize to natu- matosensory and vestibular activities are provided but ralistic contexts. However, given the limitations inher- suspended equipment is not used; (b) treatment is more ent in this single subject design, it is also possible that adult-structured or passively applied (i.e., not con- other co-occurring interventions (e.g., education), mat- ducive to child-directed play), and/or (c) treatment is urational effects, and parent participation in evaluation more cognitively focused than found in classical SI. procedures may have also contributed to the positive For example, structured perceptual-motor training ap- outcomes, thus limiting definitive conclusions about proaches may fall into this category. Although these the efficacy of the treatment. were popular 10 to 30 years ago, they are less often The last study by Ayres and Tickle (1980) utilized used today with children with autism. a retrospective (within group) design to explore variables One example of a popular modernized version of an that predicted positive or negative outcomes following SI-based program is the “Sensory Diet” (or sometimes one year of SI therapy in 10 children (mean age 7.4 years) referred to as a sensory summation approach), in which with autism. They found that subjects who tended to the child is provided with a home or classroom program have average to hyper-responsive patterns to the vari- of sensory-based activities aimed at fulfilling the child’s ous stimuli (e.g., touch, movement, gravity, and air sensory needs. A schedule of frequent and systematically puff) showed better outcomes than those with a hypo- applied somatosensory stimulation (i.e., brushing with a responsive pattern. Numerous limitations (i.e., small sam- surgical brush and joint compressions) is followed by a ple size, variability of the outcome measures used, lack prescribed set of activities designed to meet the child’s of control over maturational effects, retrospective nature sensory needs are integrated into the child’s daily rou- of the study) make conclusive statements difficult; how- tine. Another example that melds together aspects of sen- ever, the study raises the possibility that differences in sory integration theory with a cognitive-behavioral outcomes may be partially dependent on specific subject approach is the “Alert Program,” in which a child (usu- attributes including patterns of sensory processing. ally with a higher functioning level and verbal abilities) Numerous efficacy studies of SI exist with other is given additional cognitive strategies to assist with populations (e.g., children with MR or LD) (e.g., Arendt, his/her arousal modulation. Maclean, & Baumeister, 1988; Ayres, 1972; Clark et al., 1978; Hoehn & Baumeister, 1994; Humphries et al., Service Delivery Model and Approach 1990; 1992; 1993; Kanter, Kanter, & Clark, 1982; Po- latajko et al., 1991; 1992; Wilson et al., 1992). Although These models often utilize a direct intervention (one- these are excluded from this review, it is noteworthy on-one or group). Consultative/collaborative models are that these studies have sparked much controversy, given common; caregivers under the supervision of a thera- the lack of consistent empirical support. Some difficult pist may carry out school programs. Treatments gen- issues surround the nature of the intervention and erally require less equipment than classical SI and can whether or not these studies were truly representative often be provided in an inclusive setting. Cost is com- of SI therapy versus some other variation of sensori- parable to SI, depending on the length and duration of motor or perceptual-motor treatments. Several meta- intervention. analyses (Ottenbacher, 1982; Vargas & Camilli, 1999) Empirical Studies are also available; these report that effect sizes vary from low to moderately high depending on the populations No studies of specific perceptual-motor treatments studied, recency of the studies, and specific parameters for autism were found. Two studies (Larrington, 1987; measured. Outcomes in psychoeducational and motor Reilly, Nelson, & Bundy, 1983) using structured 408 Baranek sensorimotor interventions based on SI principles were was cured of his sensory defensiveness, and autistic found. Despite documenting numerous positive out- symptoms appeared to resolve. These anecdotal claims comes, the descriptive case study by Larrington (1987) appear unfounded given that no systematic data were provides limited internal validity as a result of multi- collected and that adequate methodological controls ple methodological weaknesses, probable maturation were not instituted (e.g., a variety of competing treat- effects, and a focus on an older age-group (teenager) ments were implemented during the 9-month period). that has limited generalizability to young children. The No empirical studies on the Alert Program were more rigorous study by Reilly et al., (1983) utilized a found for children with autism or related populations. It counter-balanced alternating treatments design with may be that some of these programs have been too re- 18 children diagnosed with autism (ages 6 to 11 years) cently developed to have been subjected to empirical to measure the comparative effects of two interventions tests; more likely they are utilized by clinicians who re- on production of vocalizations. The authors expected port outcomes via individualized education plans (IEPs) that structured vestibular-based activities would facil- and not through the peer-reviewed research literature. itate higher amounts of language in their subjects during the intervention than the table-top fine motor activities. Sensory Stimulation Techniques Order of the two treatment conditions was randomly assigned, and each subject received two 30-minute ses- Description and Assumptions sions of each type. No significant differences were These approaches are varied and usually involve found with respect to function of speech, articulation, passively providing one type of sensory stimulation total language, or rate of vocalizations during the in- through a circumscribed modality (e.g., touch pressure, tervention sessions. In contrast to the hypotheses, sig- vestibular stimulation) with a prescribed regimen. Some- nificant differences in favor of the fine motor group times these techniques are incorporated within the were found for variety of speech, length of vocaliza- broader sensory-integration–based programs described tions, and amount of autistic speech. Limitations of above; other times they are used in isolation. The as- these findings included: (a) the children’s vocalizations sumptions vary with the treatment, but most are based had been previously reinforced by teachers during fine on neurophysiological principles stipulating that a given motor activities in the classroom—which may have sensory experience may provide facilitatory or inhibitory proved in favor of the alternative therapy; (b) a short influences on the nervous system that result in behav- duration of the intervention period may not have been ioral changes such as arousal modulation. An example sufficient to provide an adequate sampling of behav- of a commonly used technique is “deep pressure” (i.e., iors, and (c) outcome measures outside of the treatment firm touch pressure providing calming input), which can sessions were not provided, thus it is not known if the be applied via therapeutic touch (e.g., massage; joint vocalizations generalized to other contexts. compression), or an apparatus (e.g., Hug Machine, pres- Although the Sensory Diet interventions are com- sure garments, weighted vests). Vestibular stimulation, monly used, only one study of a child suspected of hav- another example, is often used to modulate arousal, fa- ing an autistic spectrum disorder was found (Stagnitti, cilitate postural tone, or increase vocalizations. Auditory Raison, & Ryan, 1999). This case report described a integration training and visual therapies will be discussed 5-year-old boy with severe sensory defensiveness who in separate sections. underwent a treatment program consisting of brushing (i.e., sensory summation) followed by joint compres- Service Delivery Model and Approach sions an average of 3 to 5 times daily for 2 weeks. The program included integration of appropriate sensory ac- The service delivery models vary, depending on tivities interspersed throughout the child’s daily activ- the intervention provided. Both direct and consultative ities and routines (i.e., sensory diet) and was carried out approaches are used, although the direct intervention by the parents at home under the supervision of a thera- approach is more often described in research. Cost pist trained in these methods. Following initial im- varies depending on the equipment used (e.g., minimal provements, the treatment program was repeated several cost of brushes vs. moderate to significant costs of Hug months later when the child’s behaviors seemed to again Machine) and staff time allocated. deteriorate. Post-treatment parental reports suggested Empirical Studies improvements in tolerance of tactile stimulation, fewer temper tantrums, an increase in activity level, and bet- Five studies specific to children with autism were ter coordination. The authors concluded that the child found in this category. Three of the studies utilized case Efficacy of Sensory and Motor Interventions for Children with Autism 409

study descriptions (McClure & Holtz-Yotz, 1990; benefit more from these types of sensory-based inter- Ray, King, & Grandin, 1988; Zisserman, 1991), and ventions than those who are under-responsive. two provided more rigorous and controlled methodo- Both McClure and Holtz-Yotz (1990) and Zisser- logies with randomization of subjects to two alterna- man (1991) provide interesting but methodologically tive treatment conditions (Edelson, Goldberg, Edelson, weak case reports using variations of touch pressure Kerr, & Grandin, 1999; Field et al., 1997). One study garments to diminish self-stimulatory behaviors. reviewed the effects of vestibular stimulation (Ray Neither study provided a functional analysis of the self- et al., 1988), and four studies investigated the effects stimulatory behaviors prior to the interventions, but both of somatosensory stimulation on a variety of behaviors. assumed the function of the behaviors was sensation All of these studies report effects of some type of touch seeking and/or arousal modulation based and hypo- pressure/deep pressure. thesized that providing a sensory substitute would yield Although several limitations (e.g., a tendency to calming effects and decreased stereotypy. Zisserman use numerous measures with small sample sizes) exist (1991) described clinical improvements using pressure in the majority of these studies, they yield some useful gloves (overall 46% decrease) but not for pressure vest information that may guide educational planning. The (11.8% decrease). Additionally, no carry-over effects two prospective controlled studies provide preliminary could be demonstrated once the gloves were removed. evidence that touch pressure may have a calming ef- Similarly, McClure & Holtz-Yotz (1990) using elastic fect on children with autism (Edelson et al., 1999; Field wraps to effect behavioral changes (i.e., social inter- et al., 1997). Field et al. (1997) measured the effects actions, self-stimulation) reported some positive effects of touch therapy (massage) on attentiveness and re- in an institutionalized 13-year-old boy with autism/ sponsivity in 22 preschool children with autism. Mas- mental retardation; however, investigator biases, co- sage was provided for 15 minutes per day, 2 days per interventions (medication, physical agent modalities), week for 4 weeks (i.e., eight sessions). Results indi- and poor reliability/validity for the measures used sig- cated that both groups showed positive changes with nificantly limited conclusive statements. The final case respect to all the observational variables (i.e., touch study by Ray et al. (1988) describes the effects of aversion, off-task behavior, orienting to sounds; stereo- vestibular stimulation on speech sounds in a 9-year-old typies) post-treatment; the touch therapy group demon- low-functioning boy with autism and dyspraxia. Self- strated significantly greater changes in responsiveness controlled vestibular stimulation (i.e., spinning in circles to sound and stereotypies and significant improvements 5 minutes for 17 days across a 4-week period). Speech on measures of social communication. Although hy- sounds were greater in the treatment (17% time) than the pothetically links were made to changes in autonomic pre- (2% time) or post-intervention (1.3% time) phases. (vagal) activity, these claims are unsubstantiated, be- Because no measures of vestibular processing were di- cause no physiological measures were taken. rectly collected, it is unclear whether the increased Edelson et al., (1999) investigated the efficacy of speech production was directly linked to vestibular pro- the Hug Machine—a touch pressure device designed cessing changes. The child was simultaneously receiv- by Temple Grandin to decrease arousal and anxiety by ing other SI treatments in school, which may have self-administration of lateral body pressure—in 12 chil- confounded results. dren with autism that had varying levels of anxiety. The treatment group (Hug Machine twice for 20 minutes Auditory Integration Training (AIT) and Related per week for 6 weeks) showed a significant reduction Acoustic Interventions on a tension scale and a marginally significant change Description and Assumptions on anxiety. Physiological measures (galvanic skin responses) were not significantly different between AIT is based on the concept that electronically groups overall; however, variability in the experimen- modulated/filtered music provided through earphones tal group increased over the course of treatment. The may be helpful in remediating hypersensitivities and authors concluded that children in the treatment group overall auditory processing ability that is thought to be with higher initial levels of arousal/anxiety were more problematic in children with autism. Although exact neu- likely to benefit from the intervention. Although the rological mechanisms underlying AIT and other listen- small sample size and marginally significant findings ing therapies are not known, various hypotheses have limit the conclusions in this study, there is some con- been proposed (e.g., improved functioning of the retic- vergence of findings with the Ayres and Tickle (1980) ular activating system, reorganization of the cerebellar- study that children with hyper-responsive patterns may vestibular system, modiﬁcation of brain serotonin levels). 410 Baranek

AIT is said to “massage” the middle ear (hair cells in the of 10 hours) by a professional trained in the techniques. cochlea) and enhance auditory perception. Methods and Fees can range from $1000 to $3000 for the 2-week equipment vary depending on the specific philosophical treatment. The equipment itself is highly technical and approach (i.e., Tomatis or Berard); the most commonly expensive. The Tomatis method is performed in simi- used approach for children with autism in the United lar ways but often with repeated blocks of intervention States is the Berard method, developed by the French with longer overall durations, sometimes spanning otolaryngologist during the 1960s–1990s. In the Berard years of treatment. The Porges acoustic intervention method, a modulating and filtering device (e.g., Ears Ed- uses five sessions at 45 minutes each. Occasionally, ucation & Retraining System [EERS], Audiokinetron or such auditory treatments are provided within a school Audio Tone Enhancer [ATE]) accepts music input from setting if the trainer brings the equipment to the school. CDs and transforms sounds by (a) randomly modulating high and low frequencies and (b) filtering out selected Empirical Studies frequencies in accordance with the child’s performance on an audiogram (test of auditory thresholds to a series Nine studies were found using various methods of of frequencies that measures hearing ability). Sound fre- AIT with children with autism. Eight of these studies quencies that are 5 to 10 dB different from their adjacent were utilized the Berard AIT method (Bettison, 1996; frequencies (i.e., “peaks”) are filtered out for the listen- Brown, 1999; Gillberg, Johansson, Steffenburg, & ing sessions. Volume level for the left ear is sometimes Berlin, 1997; Link, 1997; Mudford, et al., 2000; Rim- reduced in order to stimulate language development in land & Edelson, 1994; 1995; Zollweg, Palm, & Vance, the left hemisphere. Although improved sound modula- 1997). The Tomatis AIT approach was used in one tion is one goal of treatment, other goals include en- study (Neysmith-Roy, 2001) for the population of in- hancement of generalized behaviors (attention, arousal, terest. An additional study of the Tomatis AIT method language, social skills, etc.). demonstrating no positive benefits for children with The Tomatis method is similar, but integrates a significant learning disabilities is also published (Ker- psychodynamic with a psychophysiological perspec- shner et al., 1990) but is not included in this review. tive. In the passive phase, the individual listens to fil- Although Porges’ acoustic intervention is currently un- tered sounds of the maternal voice, as well as prepared dergoing some scientific experiments with children music through a modulating apparatus (i.e., Electronic with autism, no published studies in peer-reviewed Ear that attenuates low frequencies and amplifies higher journals were found. frequencies). The earphones have an attached bone con- Of the nine AIT studies reviewed in total, three ductor to facilitate sounds through vibration and air used randomized, controlled methods with either no conduction methods. Later, in the active phase, the sub- treatment or an alternate treatment/placebo, (Bettison, ject is introduced to language and audio vocal exercises 1996; Rimland & Edelson, 1995; Zollweg et al., 1997), that provide feedback of his/her own voice through two utilized other methodological designs with various headphones to reinforce more normal auditory percep- levels of control (Neysmith-Roy, 2001; Mudford, et al., tion and overall quality of life. 2000), one was a pre-post open trial (Gillberg et al., Variations of listening programs applied to chil- 1997), two were descriptive case reports (Brown, 1999; dren with autism include Stephen Porges’ Acoustic Link, 1997), and one compared various types of AIT Intervention. This treatment is based on Porges’ (1998) devices (Rimland & Edelson, 1994). These are sum- Polyvagal Theory—a phylogenetic theory of autonomic marized below. nervous system control that is the substrate for emo- Bettison (1996) utilized a randomized controlled tional and affective experiences. By providing filtered trials design with AIT as the treatment condition and sounds of the human voice (as opposed to any filtered an alternative structured listening task for the control music), the listening stimulation is designed to alter- condition for 80 children with autism (ages 3 to 17 years) natively challenge and relax the middle ear muscles to with concomitant sound sensitivities. At 1 month post- enhance speech perception. treatment, both groups demonstrated significant but equal amounts of improvement for all measures show- ing nonspecific treatment effects. Significant improve- Service Delivery Model and Approach ments were maintained for the audiogram scores only; The Berard AIT treatments are usually provided patterns of change across other behavioral measures individually in a small sound-quiet room for 30 min- (behavior checklist; development, sensory checklist, utes 2 times per day for 10 to 20 days (i.e., minimum and sound sensitivity questionnaire) was so variable Efﬁcacy of Sensory and Motor Interventions for Children with Autism 411

that it is difficult to interpret clinically, even though some blind to treatment conditions. Results of audiological as- of the changes were positive. By 12 months, most of sessments demonstrated that there was a small magni- the improvements seen reverted to initial post-treatment tude change of 200 Hz on pure tone thresholds between levels taken at 1 month. It is not clear what aspect of the two groups at only one of seven frequencies tested. the interventions was responsible for the outcomes— This finding was likely spurious given the number of maturation, practice effects, or other aspects of the comparisons and was not deemed clinically significant treatment may have also influenced the results. More- by the authors. No significant group differences were over, this study documents negative side effects for two found with respect to loudness discomfort levels at any subjects who had documented psychiatric problems. frequency for any of the follow-up time points; in fact, Rimland & Edelson (1995) present some positive only one subject demonstrated an increased tolerance be- effects with 18 subjects (4 to 21 yrs of age) in a ran- tween the pre-AIT and the 9 month, post-AIT condition. domized controlled trial using AIT (EERS) versus un- Behavioral data indicated that both groups were reported filtered music. Results indicated that the treatment to improve over time; slight differences favored the con- group demonstrated fewer auditory problem behaviors trol group at 6 months post-treatment for overall scores and aberrant behaviors than the control; these im- on the Aberrant Behavior Checklist (ABC), but groups provements were maintained across 3 months. How- had similar levels of improvement on this measure at ever, neither the pure tone discomfort test nor the 9-months after treatment. No significant differences be- hearing sensitivity questionnaire provided evidence of tween groups were noted for any of the subscales of the reduction in sound sensitivity post-treatment. Treat- ABC at any of the post-treatment time points. Although ment and control groups did not significantly differ in both groups were reported to improve over time, notable their hearing acuity at any of the six points tested. Posi- fluctuations across the subscales of aberrant behavior on tive results were not limited to those with hyper- post-treatment assessment phases were not easily inter- sensitive hearing. In a large scale study, these same pretable. These findings are similar to those of Bettison authors (Rimland & Edelson, 1994) investigated the (1996) suggesting that factors other than AIT (e.g., extra differential effectiveness of three types of auditory attention; caregiver expectations, etc.) could be respon- filtering devices for 445 subjects with primary or sible for reductions in aberrant behaviors. secondary diagnoses of autism (CA ϭ 4 to 41 years; Mudford et al. (2000) used a balanced cross-over M ϭ 10.73 years). Because no differences were found design to assess the effects of AIT (twenty 30-minute across the three devices, data were collapsed across treat- sessions) with 16 children with autism ages 5 to 14 ment conditions. Results indicated a significant but small years. Four were girls; all of the subjects were low- improvement in hearing sensitivity, as well as decreased functioning. Parents and AIT practitioners were blind variability in audiograms for some subjects, mostly to the treatment conditions that included either AIT during the first 3 months. Lower-functioning individu- (modulated music) without the filtering component, or als seemed to make greater gains. However, there was the placebo control (headphones disengaged, music no significant relationship between level of sensitivity playing in background of room). Baseline and follow- pre-AIT and the behavioral outcomes post-treatment. up assessments were used for each condition, includ- Zollweg, Palm, & Vance (1997) used a well- ing standardized measures of aberrant behavior, controlled, double-blind design with an alternative treat- cognitive functioning, adaptive levels, language levels, ment to study the effects of AIT on 30 low-functioning, and observational data. Notably, several children multihandicapped, institutionalized residents. Nine of dropped out of the study as a result of inability to tol- the 30 residents had autism; only 1 of the 9 with autism erate the procedures and/or severe behavioral difficul- was a child under the age of 8 years. Participants were ties. Of the 16 children completing the study, there were assigned randomly to treatment (filtered/modulated no positive findings in favor of the AIT condition, de- music) or control (unfiltered/unmodulated music) groups spite a liberal alpha level used in the statistical tests. and provided with twenty 30-minute sessions by trained In fact, small but significant differences on measures practitioners who were blind to the treatment conditions. of aberrant behavior favored the control condition for Participants were allowed to engage in other activities this low-functioning group. (e.g., eating, toy play, magazine browsing) simultane- Neysmith-Roy (2001) describes outcomes of the ously with treatment to facilitate cooperation. Subjects Tomatis AIT method with six children with severe were assessed for hearing sensitivity and tolerance autism ranging from 4 to 11 years. Several blocks of for loud sounds pre- and post-treatment (1- 3-, 6-, and “Intensives” (i.e., twenty 30-minute sessions repeated 9-month time periods) by a licensed audiologist, also from four to eight times) were each followed by a 3- to 412 Baranek

8-week unsystematic evaluation period spanning 6 to across measures seen in several studies (e.g., Bettison, 21 months, depending on the individual child’s pro- 1996; Zollweg et al., 1997) are difficult to interpret. gram. Pre- and post-measures of autistic symptoms on Adverse effects (i.e., seizures, behavioral problems) the Childhood Autism Rating Scale (CARS) were and drop-out rates require special attention given that recorded on videotapes of two separate play conditions these are reported in several AIT studies. and parent/teacher reports during both the intervention Methodologically stronger studies (Bettison; 1996; and evaluation phases. Videos were randomized and Mudford et al., 2000; Zollweg et al., 1997) demon- coded by two raters who were blind to the conditions strated that improvements in behavior often were not of the study for each subject. Reductions in autistic significantly different between treatment (AIT) and symptoms overall including an improvement in pre- control (unfiltered/unmodulated music) conditions linguistic behaviors were noted in three of the six boys (Bettison, 1996; Zollweg et al., 1997). In two studies (50%), who also happened to be among the youngest (Mudford et al., 2000; Zollweg et al., 1997) the con- subjects, but no improvements were seen in the other trol group showed a subtle advantage in reduction of three (older) boys. Audiological/physiological mea- aberrant behaviors at at least one point in time. Simi- sures were not employed because the boys were unable lar improvements for subjects in both treatment and to cooperate for the testing. control conditions suggest that effects are due to fac- A case report by Brown (1999) on two preschool- tors that are peripheral to the treatment (e.g., extra at- age siblings with autism describes improvements in tention; caregiver expectations; compliance training) sensory-motor functions, attention, social interest, praxis, and/or other co-occurring treatments (i.e., behavioral eye control, and speech for two children who underwent and educational interventions) may be influencing out- AIT for two 30-minute sessions for 10 days. Similary, comes. Future replication studies need to better control Link (1997) presented a case report for three children for such effects. Although the hypothesis that music in with autism, documenting mixed findings. Sound sen- general (filtered or unfiltered) may have a beneficial sitivity (using unvalidated measures) was noted to be effect on behavior is plausible, it is yet to be tested em- unchanged for two of the three children and inconclu- pirically. Gravel (1994) provided a detailed critique of sive for the third. Informal behavioral observations AIT studies, stressing the importance of distinguishing were found to be improved for one child (across par- between statistical and clinical significance. She states ent and teacher ratings) and mixed a second child. One that small differences between frequencies (i.e., 5 dB) child unfortunately suffered a psychomotor seizure dur- or fluctuations in hearing sensitivity in some children ing the AIT procedure and was removed from the pro- with autism may be attributable to attentional/behavioral tocol on day 5. Parent and teacher reports on day 5 difficulties that preclude reliable responses on behav- indicated deterioration in behaviors possibly related to ioral audiometry. Furthermore, electrophysiological the seizure. Both studies were weak methodologically measures (ABRs) fail to demonstrate differences in such that no systematic data was collected and pre- and hearing sensitivity between children with autism and post-measures lacked independence, as well as valid- controls—a finding that may challenge the overall ity and reliability data; thus findings are speculative. premise of AIT. In summary, the nine available studies using various methods of AIT with children with autism demonstrate mixed results. Studies with sample selection Visual Therapies biases and unmasked evaluation processes would tend Description and Assumptions to favor outcomes for the treatment groups; however, this was not always the case. Some studies, using var- A variety of visual therapies including but not lim- ious types of methodologies, reported improvements in ited to oculomotor exercises, ambient prism lenses, and either behavior or audiological measures (e.g., Brown, colored filters (i.e., Irlen lenses) have been applied to 1999; Link, 1997; Neysmith-Roy, 2001; Rimland, children with autism. These visual or behavioral opto- 1994; 1995) at some points in time for some children. metric therapies and are aimed at improving visual pro- However, it is important to note that if changes are seen cessing or visual-spatial perception that may be related in behaviors without concomitant changes in hearing to autistic symptomatology (e.g., unusual visual stereo- sensitivity, an alternative mechanism than that pro- typies, coordination problems, strabismus, attention, posed by AIT theory may be responsible for the im- etc.). In particular, the ambient (i.e., visual-spatial sys- provements. Furthermore, significant fluctuations of tem) as opposed to the focal (i.e., visual acuity system) behavior over time, and inconsistency of performance is hypothesized to be dysfunctional in children with Efﬁcacy of Sensory and Motor Interventions for Children with Autism 413

autism. Thus prism lenses transform the light through sures were unchanged, at least in the context of the lab- an angular displacement of 1 to 5 degrees (base up or oratory tasks. Children were not engaged in simulta- base down), producing a shift in the field of vision. neous visual training tasks that are usually prescribed These are thought to lead to more stable perception and in conjunction with prism lenses, which may have lim- improved behavior or performance. ited long-term effects. Treatment effects did not appear related to whether or not the children had a prediag- Service Delivery Model and Approach nosed strabismus; however, a slight trend toward more behavioral problems during the clear lens (placebo) Visual therapies and corrective lenses are pre- condition for the strabismus subgroup was reported. scribed individually through a licensed optometrist. In Unfortunately, the above results were less encour- some cases, programs are carried out by parents or other aging than those from the original pilot study (Kaplan professionals. Various visual treatments may also be et al., 1996) and a more recent study (Carmody et al. combined, as in the case of ambient lenses and oculo- 2001), both of which demonstrated significantly bet- motor exercises. Costs may be moderate and include ter performance on orientation and visual-spatial per- the optometrist’s fees for the initial and follow-up eval- formance when children with autism were wearing uations and cost of prescriptive eyewear. correct (“facilitating”) prism lenses. Carmody et al. (2001) assessed responses to prism lenses for a conve- Empirical Studies nience sample of 24 children (2 girls, 22 boys) with Anecdotal data are plentiful; however, empirical autism, ages 3 to 18 years, in Hong Kong. Optometric studies regarding the efficacy of visual therapies specific evaluations revealed that 18 children had normal visual to children with autism are limited. No published stud- acuity, 3 were far-sighted, and 3 were near-sighted. All ies were found on the use of Irlen lenses or independent children were assessed on measures of spatial orienta- use of oculomotor therapies for children with autism. tion/spatial management abilities with two functional Three studies (all by the same group of researchers) in- tasks (i.e., television viewing and ball catching) during vestigating prism lenses were found (Carmody, Kaplan, each of three visual assessment conditions: (1) habit- & Gaydos, 2001; Kaplan, Carmody, & Gaydos, 1996; ual viewing (no lenses), (2) eyeglasses with prisms using Kaplan, Edelson, & Seip, 1998). base-up condition, or (3) eyeglasses with prisms using The study by Kaplan et al. (1998) used the most base-down conditions. Both prism conditions used a mild rigorous methodological design (double-blind cross- displacement of the visual scene (5.6 degrees); however, over). It was an extension of the earlier pilot investi- one condition was judged to be the child’s favored or gation (Kaplan et al., 1996) and used 18 children with correct condition (facilitating lenses) and one was judged autism (CA ϭ 7 to 18 years; M ϭ 11.53 yrs). Thirty- to be an incorrect condition (interfering lenses) based on nine percent had strabismus. All subjects were pre- observational measures during the trials by the initial scribed prism lenses with modification determined rater. Assessments were videotaped. The full perfor- individually for the visual direction (base up or base mance assessments were then reviewed by an indepen- down) and angular displacement. Five subjects were lost dent rater, and results were analyzed statistically to as a result of non-compliance with the eyewear. The re- validate whether or not the facilitating condition was su- maining 18 were matched and randomly assigned to one perior. Results validated the initial rater’s assessment that of the two conditions (placebo lenses or treatment with head position, body posture, facial expressions, and ball prism lenses) for 3 months; then treatment conditions catching skills significantly improved in the “facilitating were switched for the second phase (4 months). Pre-, lens” condition as opposed to the habitual or interfering mid-, and post-treatment measures were taken for pos- lens conditions. tural orientation, attention, and visual-spatial perfor- Differences in the ages of the samples and the rel- mance tasks, as well as ratings of behavioral problems. ative strength of the methodological designs across the Results indicated short-term positive effects—behavioral three studies could have contributed to the inconsis- 1 findings were most apparent at the mid-evaluation (1 ⁄2 tent findings. The Kaplan et al. (1996) and the Car- or 2 months), with less improvement at follow-up (3 or mody et al. (2001) studies used considerably weaker 4 months). Performance on orientation and visual- designs with relatively short trial periods across the spatial tasks was not significantly different between three assessment conditions (i.e., no lenses, incorrect conditions. Given these findings, it is unclear what spe- prism lenses or correct prism lenses). Replication stud- cific mechanisms were responsible for the behavioral ies by independent investigators, using well-controlled improvements because the visual performance mea- designs and longer-term follow-up are needed. Likewise, 414 Baranek outcome measures used in naturalistic contexts are From a clinical perspective, it appears that this treat- needed to demonstrate generalizability of these findings. ment is rarely applied in any systematic way for use with children with autism unless other accompanying and significant neuromotor problems are present (e.g., Sensorimotor Handling Techniques cerebral palsy). No empirical studies were found in peer- Description and Assumptions reviewed journals for C-S. No empirical studies of sensorimotor patterning were found specific to autism. Several types of sensorimotor handling techniques However, two older empirical studies with children with (e.g., reflex integration, neurodevelopmental therapies, mental retardation were found (Bridgman, Cushen, patterning, etc.) have been applied to children with de- Cooper, & Williams, 1985; Neman et al., 1974). Both velopmental disabilities. Neurodevelopmental therapy studies provide similar findings and limitations. In the (NDT) is a specific sensorimotor (physiotherapy) more rigorous study (Neman et al., 1974), the pattern- treatment that originated with the Bobaths in England ing group out-performed the other groups in visual com- in the 1950s–1960s. Its focus is on normalization of petence and mobility. On other measures, the patterning muscle tone, integration of primitive reflexes, and fa- group out-performed the passive control but not the al- cilitation of more normal movement patterns through ternative intervention group; no measurable differences specific handling techniques. Sensorimotor patterning were found in IQ, and no case made dramatic im- is a remedial technique that uses a series of very struc- provement on any measure. Results from methodolog- tured, passively manipulated exercises to the limbs to ically weaker study (Bridgman et al., 1985) indicated reprogram the central nervous system. Originally de- that the treatment group slightly out-performed the con- signed by Doman and Delcato in the 1950s–1960s, this trol group on language and socialization, whereas the treatment is based on an older and simplistic recapitu- control group fared better in self-help skills. Given that lation theory. Developmental gross motor patterns that in both studies improvements were more pronounced in may have been “missed” (e.g., creeping or crawling) the earlier assessments, it appears that findings may re- are patterned passively for neurological reorganization. flect early enthusiasm of participants. High drop-out Although it is not used extensively for children with rates are particularly noted with this treatment method. autism, there has been a recent resurgence of interest, particularly as a last-resort therapy. Although craniosacral (C-S) therapy (Upledger, 1996) is not a sensori- Physical Exercise motor treatment per se, it’s similarities to other Description and Assumptions handling techniques combined with recent anecdotal data of its increased use with children with autism war- Although physical exercise is included in many ranted inclusion in this paper. This osteopathic treat- regular education curricula, it is not systematically or ment involves physical manipulation (e.g, repeated consistently utilized with children with autism. Health treatments using gentle and noninvasive traction and benefits of various exercise programs have been touted, decompressions) to alleviate restrictions in the cranio- including changes in physical as well as mental well- sacral system. being. Researchers have been interested in the application of physical exercise particularly as it effects Service Delivery Model and Approach maladaptive or self-stimulatory behaviors. Assumptions frequently made are that aerobic exercise would di- These interventions are administered individually minish stress or modulate self-stimulatory or hyper- by persons trained in the procedures through attendance active behavior through physiological changes related in specialized workshops. Contact with the child must to release of neurotransmitters such as acetylcholine or be direct physical contact. Interventions may be short beta-endorphins. or long term for NDT or C-S. Home programs for patterning therapies are eventually carried out by families Service Delivery Model and Approach and tend to be particularly long in duration (8 hours per day for several months or years). Cost of treatments is Physical education programs are provided easily dependent on frequency and duration (therapists’ fees). via individual or group methods by teachers within an educational curriculum given adequate physical space. Frequency or interventions vary from multiple times Empirical Studies per day to not at all. The majority of research studies No empirical studies specific to autism or related conducted utilize approximately 15 minutes of treat- disorders (e.g., mental retardation) were found for NDT. ment time per day. The cost of providing services is Efﬁcacy of Sensory and Motor Interventions for Children with Autism 415

usually modest, although larger physical spaces are compensatory approaches, etc.), no empirical studies often necessitated (e.g., running track or gym) in lieu on autism were found. Clinical observation suggests of expensive equipment (e.g., exercise bicycle). that providing developmental motor training (i.e., skills training in hierarchically sequenced developmental Empirical Studies stages) and compensatory teaching strategies are com- Four studies of the efficacy of physical exercise monly utilized approaches in many educational and for children with autism were found (Kern, Koegel, & therapeutic programs for children with various diag- Dunlap, 1984; Kern et al., 1982; Levinson & Reid, noses; thus it was surprising to find so little. However, 1993; Watters & Watters, 1980). All of these studied it appears that these strategies are being provided as effects of antecedent exercises on self-stimulatory/ part of a broader therapeutic consultation program or stereotyped behaviors using some variation of single- via comprehensive educational models; thus empirical subject designs with controls (e.g., alternating treat- data specific to these components may be obscured in ments designs). Two of these studies also measured the literature. Future reviews may wish to address these aspects of academic and play tasks in children with components more specifically. autism (Kern et al., 1982; Watters & Watters, 1980). All studies found some beneficial, albeit short-lived, effects of exercise for decreasing self-stimulatory be- SUMMARY OF EFFICACY OF SENSORY AND haviors and mixed findings for improving other simple MOTOR INTERVENTIONS cognitive/play tasks. Effects were greater for more in- tensive aerobic activity (relative to mild exertion). A The theoretical strength of many sensory and 1 maximum effect was noted at about 1 to 1 ⁄2 hours with motor interventions, particularly sensory integration attenuation over time. Limitations of these studies in- (SI), rests on empirical findings that children with cluded relatively small sample sizes, large variability in autism indeed have measurable deficits in various sen- the independent measures, and the potential confounds sory processing and motor functions. Some of the treat- of extra attention/social interaction that may have con- ments reviewed (e.g., sensorimotor handling) provide tributed to the beneficial effects of the program. The a questionable rationale for their use with children with Watters & Watters (1980) study also concluded that autism and have no empirical evidence to evaluate their there was no evidence to support that decreases in self- efficacy with this population. In particular, sensorimo- stimulatory behavior would automatically generalize to tor patterning is based on an older neurological theory improved academic performance. Two additional stud- that has been essentially disproven; several other pro- ies on general physical education effectiveness for chil- grams based on sensory integration theory also suffer dren with autism were found (Schleien, Heyne, & Berken, from partially outdated assumptions and are being mod- 1988; Weber & Thorpe, 1992). Weber & Thorpe (1992) ernized. Although the sheer volume of studies was low found that for older children (ages 11 to 15) greater across categories reviewed, it was encouraging that sev- learning/retention of gross motor skills occurred in a eral new studies were conducted in recent years. Find- task variation condition. Schleien et al., (1988) found ings from these studies were often mixed. Several that physical education activities in an integrated phys- studies in the area of SI, sensory stimulation, auditory ical education class with typical children did not ap- integration training, prism lenses, and physical exer- pear to significantly affect performance of motor or cise yielded some positive, albeit modest outcomes; play skills over a short term (9-week intervention). however, methodological constraints (e.g., use of small Short treatment length or poor sensitivity of motor mea- and convenience samples, weak/uncontrolled designs, sures may have limited the findings. It is also possible observer bias, etc.) limit conclusive statements and gen- that general physical exercise and recreational activi- eralizability of much of this work. In some areas, such ties do not necessarily generalize to improving social as AIT, a few well-controlled studies have been re- play without specific play skills being taught within the cently conducted but with little overwhelming support context of treatment. for the treatment. The biggest limiting factor is that many studies fail to directly link changes in the purported dysfunc- Other Categories of Sensory or Motor tional mechanism (e.g., auditory sensitivity, visual dis- Interventions tortions, vestibular dysfunction) to the functional Although searches were attempted for many other changes in behavior. Studies either provide outcome specific sensory-motor intervention categories (e.g., de- measures of the proximate sensory behaviors (e.g., au- velopmental motor therapies, motor skills training, ditory sensitivity, arousal, tactile defensiveness) or the 416 Baranek broader functional behaviors (e.g., social interactions, subtypes (e.g., hyper- versus hypo-responsive) and other play skills, academic performance), and rarely do they subject variables (e.g., age, developmental levels) may link both in systematic and measurable ways. A few affect prognosis for treatment outcomes. Though small preliminary studies of AIT and sensory stimulation sample sizes and retrospective designs limit generaliz- treatments have attempted this; however, the results are ability, there appeared to be some converging evidence still tenuous and inconsistent across studies, indicating to suggest that a hyper-responsive pattern (i.e., high anx- need for replications. Furthermore, it is still unclear iety, arousal, or sensitivities) may be more amenable to what specific processes may be responsible for the sensory techniques aimed at arousal modulation and re- gains reported, even in the studies that had reasonable sultant gains in performance. Physiological studies of controls and sample sizes. Are treatment effects truly arousal indicate that younger (or less mature) children reflective of the intervention, or are there other non- may have a higher tendency to display hypersensitive treatment effects (e.g., parent’s expectations, matura- reactions and reject sensory stimuli that interfere with tion, imposed structure, added attention, practice, etc.) other aspects of functioning. If so, one implication of that influenced the results? Other aspects of the tasks, these findings may be that perhaps beginning some not central to the treatment protocol, may also be pro- types of sensory-motor interventions at earlier ages ducing beneficial effects. For example, children in one would be more beneficial. Some studies have only been AIT study (Bettison, 1996) responded favorably under conducted with older children and adults (e.g., exercise both the treatment and the alternative condition, thus, treatments), and these results cannot be generalized to processed (filtered/modulated) music did not appear to preschool populations. We cannot know the answers to be the critical treatment component. Although music these types of questions until more systematic research “in general” may prove to be beneficial, an alternative with increased specificity of subject variables is con- explanation could involve the repeated demands for ducted to help distinguish various levels of response to compliance and structured attention to task that are not treatments; however, these findings indicate that when specific to auditory processing. Another AIT study provided, sensory and motor interventions need to be (Zollweg et al., 1997) demonstrated that both groups individualized for a given child with autism. improved but the control group outperformed the treat- A further concern of this area of intervention is that ment group on some measures, further questioning the most of the studies provide limited follow-up after efficacy of the treatment. Similarly, one of the core intervention, and so it is not known whether positive principles and strengths of classical sensory integration effects are sustained in the longterm. A few better- is that the therapy is child-directed and based in an in- controlled follow-ups have been included in some of the dividualized play context. Thus, purported gains in en- AIT studies; however, in those studies in which positive gagement and functional play skills may be influenced effects were noted initially, an attenuation of responses by play coaching techniques employed by an expert occurred over time (over the course of 9 to 12 months). therapist as much as they are by improvements in sen- These types of findings were also true for one study on sory processing per se. Only further research with ad- prism lenses and two on exercise treatments. This could equate specificity and controls (e.g., multiple baseline indicate that either the treatments were not useful in the conditions in single-subject designs) can tease apart the long term or, conversely, that more frequent application effective from the noneffective components of these of treatment is needed to maintain such effects. Repeated interventions. treatments were certainly useful with exercise therapies Given that autistic symptoms are manifested dif- where physiological and behavior changes were sus- ferently across development and that heterogeneity tained for approximately 90 minutes following each exists within the autism spectrum, it is likely that indi- treatment. The effects of treatment frequency, duration, vidualized patterns of reactivity may be associated with and intensity on both short- and long-term outcomes differential treatment outcomes irrespective of the in- need to be further addressed. tervention category reviewed. Although outcomes for Finally, issues of generalizability and feasibility individual children have been mixed, it is possible that of sensory and motor interventions need to be addressed significant individual differences in subject character- more fully. The majority of studies across sensory and istics may be masking significant group effects. That is, motor treatment categories, particularly those con- we don’t know which children will benefit the most ducted less recently, have not attempted to investigate from which treatments and under which specific condi- broader issues of generalizability. Two SI studies were tions. Several studies in the areas of SI and sensory conducted in naturalistic contexts and documented stimulation indicated that specific sensory processing some functional gains outside of the treatment context. Efﬁcacy of Sensory and Motor Interventions for Children with Autism 417

Others have been unable to prove that the intervention classical SI provides the strongest child-centered and has substantial effects on academic performance, beyond playful approach; this type of approach is often ap- the scope or context of the treatment per se. For exam- pealing to even the most unmotivated or disengaged ple, several exercise studies have demonstrated positive child. In the case of sensory stimulation treatments effects on primary reduction of aberrant behaviors, but (e.g., brushing, massage), AIT, patterning, and C-S, the more limited improvements in secondary effects on child must be able to tolerate various sensory applica- academic skills, play, and recreation. AIT studies tions or physical manipulations provided by the thera- demonstrate much variability across studies and lack of pist and usually in a restricted space. For some children replication for broader behavioral effects—in some with autism, however, structure and repetition are not cases, control treatments produced stronger reductions necessarily aversive, although passive application of in aberrant behaviors. Gains in areas of mastery play and stimulation may be; the effects on stress need to be doc- engagement as a result of SI (i.e., Case-Smith & Bryan, umented better in future studies. No negative side ef- 1999) are interesting to note; however, it is not sur- fects have been reported in most of the literature prising that behaviors less directly addressed in the con- reviewed, with the exception of AIT, in which two stud- text of therapy (e.g., peer interactions and functional ies documented increased behavior problems and some communication) show the least improvement. Without adverse health effects with some children. In addition, direct practice in generalizing to functional tasks and compliance has been an issue in several treatments, in naturalistic environments, the effects of therapeutic such as limited tolerance for eyewear or headphones, gains in sensory processing or motor components may causing drop-outs from treatment. be limited. Studies of motor development in typical Feasibility solutions, however, must go beyond children have demonstrated that behaviors emerge from ease of integration and cost effectiveness, they must the confluence of multiple dynamic systems; thus, task- also answer questions of best practice within the scope relevant information and perceptions of affordances in of educational goals. Because most of these interven- the environment can substantially alter movement pat- tions are used to augment comprehensive educational terns (Connolly & Elliott, 1989; Thelen & Ulrich, programs, it is important to know whether or not these 1991), and motor skills learned in one context may not treatments actually facilitate progress or hinder it by automatically generalize to naturalistic functional ac- taking away valuable instruction time. The effects of tivities (Case-Smith, 1995). In fact, several studies have specific sensory and motor interventions combined with found that performance of children with autism on goal- various types of educational models need to be further directed motor and imitation tasks appears to be better investigated in large outcome studies. At least one case in meaningful and purposeful contexts than in decon- study report of three preschool children (Schwartz textualized situations (e.g., Hughes & Russell, 1993; et al., 1998) indicates that there may be multiple means Rogers, et al., 1996; Stone, et al., 1997). Thus one to achieving promising outcomes in young children would expect that difficulties in applying newly ac- with autism. Various combinations of specific inter- quired sensory or motor components to functional tasks vention strategies (including SI in one case) were in- and/or generalizing motor skills learned out of context tegrated into the curriculum and produced positive would be magnified for children with autism. effects. Although classical SI therapy originated with The feasibility of carrying out specific sensory and a clinic based, noninclusive model, newer approaches motor interventions varies tremendously and is depen- are attempting to utilize a more naturalistic context with dent upon such variables as cost, qualifications of sensory-based activities integrated into the classroom needed professionals, and congruence with the philo- routine. However, it will be important to investigate to sophical orientation of the broader educational pro- what degree specific treatments can be “altered” to fit gram. With respect to cost, all treatments require an inclusive education model while still retaining their specialized training and varying amounts of staff time; essential therapeutic elements and purported benefits. however, some require much more technologically so- Comparisons of such treatments need to be systemati- phisticated and expensive equipment (e.g., AIT, Hug cally investigated in future efficacy research. Machine, prism lenses) than do others (e.g., massage, As Rogers (1998) eloquently summarized in her re- C-S). Some treatments (e.g., exercise therapies, some view of comprehensive educational programs, we must SI-based approaches, prism lenses) are certainly more keep in mind that a lack of empirical data does not infer easily administered within the context of inclusive ed- that the treatment is ineffective, but rather that efficacy ucational programs than other treatments (e.g., senso- has not been objectively demonstrated. Funding will rimotor patterning). Of all the treatments summarized, be critical to this increasingly urgent investigation. 418 Baranek

Relying on nonharmful but potentially ineffective treat- for core educational curricula. If and when uti- ments can squander valuable time that could be used in lized, these should be viewed as supplementary more productive educational or therapeutic ways. Given interventions integrated at various levels into the that at least some positive findings are noted with re- broader individualized educational program. spect to the sensory and motor interventions reviewed, 4. Specific task/environmental modifications for future research must move from the current level of sensory processing or motoric deficits tend to be small-scale, poorly controlled, unsystematic studies of described within the context of broader educa- effectiveness, to a level that demands scientific rigor tional approaches or in combination with spe- and well-controlled large-scale designs. Only such re- ciﬁc interventions more so than they are reported search can provide answers to important questions of in the empirical literature. For example, adapta- not only what is effective but with whom and under tions may take the form of changing performance what conditions. expectations, modifying classroom activities to minimize negative sensory reactions, perceptual distortions or motoric difficulties, teaching com- RECOMMENDATIONS FOR EDUCATION pensatory strategies, and/or maximizing the child’s strengths to bypass sensory and motor 1. Although not all children with autism display difficulties and facilitate fuller participation. sensory processing and motor dysfunction, Such adaptations for sensory processing or motor these types of difficulties are prevalent in the difficulties would be feasible in many educa- population and are reported to interfere with tional programs and could be used in tandem performance in many broader developmental with other interventions. and functional domains. Therefore, at a mini- 5. Given the limited scientific basis of many of the mum, “best practice” guidelines would indicate remedial sensory and motor intervention ap- that educational programs for young children proaches reviewed in this paper, a conservative with autism need to incorporate appropriately approach is recommended for prescribing spe- structured physical and sensory environments cific sensory or motor treatments. Best practice that accommodate these unique sensory pro- would suggest that decisions be made on an in- cessing patterns and provide opportunities for dividualized clinical basis by expert profession- developmentally appropriate sensory-motor als. If indicated, it behooves the professional to experiences within the context of functional provide treatments in shorter-term increments educational goals. (e.g., 6 to 12 weeks) and document progress in 2. Comprehensive educational programs may ben- a systematic manner. Treatments need to be dis- efit from consultation with knowledgeable pro- continued if effects are not apparent within an fessionals (e.g., occupational therapists, speech expected time frame or if negative reactions are and language therapists, physical therapists, documented. Certainly, regression in skills fol- adaptive physical educators, etc.) to provide lowing discontinuation of services merits special guidance about potential interventions for chil- attention. dren whose sensory processing or motoric diffi- 6. With respect to specific interventions, the fol- culties interfere with educational performance, lowing points are made: as well as to provide support for families strug- a. Most categories of sensory and motor inter- gling with these issues. It is important to note ventions, including SI, sensory stimulation that related services may provide many mean- approaches, AIT, and prism lenses have ingful interventions that go well beyond the shown mixed effects for children with autism scope of this paper. Thus sensory and motor in- primarily through uncontrolled, descriptive terventions are not synonymous with the terms studies; large-scale experimental studies are for professionals employing these interventions often lacking. Furthermore, beneficial effects (e.g, occupational therapy, physical therapy, or of sensory and motor treatments on atypical speech pathology); these terms should not be in- behaviors in some children have been shown terchanged when making decisions for provision to be short-lived in several studies that fol- of therapeutic services. lowed children longitudinally. 3. It is important to note that none of the specific b. There is no available empirical evidence to interventions reviewed claim to be a substitute support the use of sensorimotor handling Efﬁcacy of Sensory and Motor Interventions for Children with Autism 419

therapies for general educational purposes their relationships to broader behavioral and in young children with autism. educational outcomes. c. Potential risks of AIT (adverse side effects in b. Replication studies by independent investi- some cases, lack of safeguards for hearing gators are needed for all sensory and motor loss) need to be seriously weighed against the treatment categories that have shown posi- potential benefits for each individual case tive effects. undergoing such an intervention. c. Methodologically more rigorous designs (e.g, d. SI and SI-based approaches appear to be rel- multiple baseline single-subject designs; ran- atively safe and anecdotally have shown domized controlled trials) with better-defined some benefits in a few children with autism. intervention components (duration/course, However, given the mixed findings across stimulus type/intensity/frequency, contextual uncontrolled, small sample studies, these conditions) and more reliable/valid and sys- treatments need to be individually deter- tematic outcome measures are recommended mined and carefully monitored until better- to directly test effects of a given treatment. controlled replication studies are completed d. Studies that identify specific behavioral and with children with autism. physiological patterns (individual differ- e. Deciding how much to pull a child away from ences) that differentiate responders from his or her educational program is a difficult nonresponders to specific treatments are decision, and thus providing treatments within warranted. the context of an inclusive environment needs e. Studies linking the purported neurological to be addressed. mechanisms with both proximate measures 7. Given the variability in developmental profiles of the phenomenon and functional measures of children with autism, it should be expected of performance in broader domains (e.g., play that not all children benefit equally from sen- skills, social skills, academic performance, sory or motor interventions. There is not a independent functioning) are needed. “one-size-fits-all” treatment for a diagnosis of f. Studies providing information on earlier in- autism. Thus the indiscriminate use of any sen- tervention, preventive benefits of sensory sory- or motor-based intervention is unethical. and motor interventions, and/or long-term Comprehensive assessments should be the basis impact on educational programming and for service decisions, and, if necessary, sensory functional outcomes of children in natural- and motor interventions must be prescribed in istic contexts should be encouraged. an individualized manner consistent with the g. Studies documenting the relative contribu- functional goals for each child. tions of sensory or motor interventions 8. Although, in general, intervention strategies within comprehensive educational curricula need to be developmentally appropriate, reme- are needed because it is unknown whether diation of component-level sensory processing educational goals are facilitated or inhibited functions or developmental motor skills may not by these various interventions. automatically result in functional gains or generalize to relevant contexts. Thus, best practice would suggest that functional activities integrated into daily routines within naturalistic ACKNOWLEDGMENTS contexts increase retention and generalization of skills. I extend my gratitude to Alice Blair, Angela Suratt, 9. The paucity of available scientific research in and Lorin McGuire for their assistance in preparation this area leaves many questions yet to be an- of this manuscript. I also thank Ruth Humphry, Linn swered. However, these preliminary findings Wakeford, Cathy Lord, Sally Rogers, and Pauline Fil- may provide direction for future studies. The ipek for their helpful reviews. An earlier draft of this following are recommended: manuscript was commissioned by the National Acad- a. Cross-sectional and longitudinal studies are emy of Sciences, National Research Council’s Com- needed to document the developmental pro- mittee on Educational Interventions for Children with gressions of unusual sensory processing fea- Autism, and presented at the first committee meeting tures and qualitative motor functions and in December of 1999. 420 Baranek

REFERENCES Dahlgren, S. O., & Gillberg, C. (1989). Symptoms in the first two years of life. European Archives of Psychiatry and Neurologi- Adams, L. (1998). Oral-motor and motor-speech characteristics of cal Science, 238, 169–174. children with autism. Focus on Autism and Other Developmen- DeMyer, M., Barton, S., & Norton, J. A. (1972). A comparison of tal Disabilities, 13, 108–112. adaptive, verbal and motor profiles of psychotic and non- Adrien, J. L., Lenoir, P., Martineau, J., Perrot, A., Haneury, L., Lar- psychotic subnormal children. Journal of Autism and Childhood mande, C., & Sauvage, D. (1993). Blind ratings of early symp- Schizophrenia, 2, 359–377. toms of autism based upon family home movies. Journal of the Edelson, S. M., Goldberg, M., Edelson, M. G., Kerr, D. C., & American Academy of Child and Adolescent Psychiatry, 33, Grandin, T. (1999). Behavioral and physiological effects of deep 617–626. pressure on children with autism: A pilot study evaluating the Adrien, J. L., Ornitz, E., Barthelemy, C., Sauvage, D., & Lelord, G. efficacy of Grandin’s Hug Machine. American Journal of Oc- (1987). The presence or absence of certain behaviors associated cupational Therapy, 53, 145–152. with infantile autism in severely retarded autistic and nonautis- Field, T., Lasko, P. M., Henteleff, T., Kabat, S., Talpins, S., & Dowl- tic retarded children and very young normal children. Journal ing, M. (1997). Brief report: Autistic children’s attentiveness of Autism and Developmental Disorders, 17, 407–416. and responsivity improve after touch therapy. Journal of Autism Adrien, J. L., Perrot, A., Sauvage, D., Leddet, I., Larmande, C., and Developmental Disorders, 27, 333–339. Hameury, L., & Barthelemy, C. (1992). Early symptoms in Gepner, B., Mestre, D., Masson, G., & de Schonen, S. (1995). Pos- autism from family home movies: Evaluation and comparison tural effects of motion vision in young autistic children. Neuro- between 1st and 2nd year of life using I.B.S.E. Scale. Acta Pae- report, 6(8), 1211–1124. dopsychiatrica: International Journal of Child and Adolescent Gillberg, C., Ehlers, S., Schaumann, H., Jakobsson, G., Dahlgren, Psychiatry, 55, 71–75. S. O., Lindblom, R., Bagenholm, A., Tjuus, T., & Blidner, E. Amato, J., & Slavin, D. (1998). A preliminary investigation of oro- (1990). Autism under age 3 years: A clinical study of 28 cases motor function in young verbal and nonverbal children with referred for autistic symptoms in infancy. Child Psychology and autism. Infant-Toddler Intervention 8, 175–184. Psychiatry, 31, 921–934. Arendt, R. E., Maclean, W. E., & Baumeister, A. A. (1988). Critique Gillberg, C., Johansson, M., Steffenburg, S., & Berlin, O. (1997). of sensory integration therapy and its application in mental retar- Auditory integration training in children with autism: Brief re- dation. American Journal of Mental Retardation, 92, 401–429. port of an open pilot study. Autism, 1(1), 97–100. Ayres, J. (1972). Improving academic scores through sensory inte- Gravel, J. S. (1994). Auditory integrative training: Placing the bur- gration. Journal of Learning Disabilities, 5, 338–343. den of proof. American Journal of Speech and Language Pathol- Ayres, A. J., & Tickle, L. S. (1980). Hyper-responsivity to touch and ogy, 3, 25–29. vestibular stimuli as a predictor of positive response to sensory Greenspan, S., & Weider, S. (1997). Developmental patterns and out- integration procedures by autistic children. American Journal comes in infants and children with disorders in relating and of Occupational Therapy, 34, 375–381. communicating. A chart review of 200 cases of children with Baranek, G. T. (1999). Autism during infancy: A retrospective video autistic spectrum diagnoses. Journal of Developmental and analysis of sensory-motor and social behaviors at 9–12 months Learning Disorders, 1, 87–141. of age. Journal of Autism and Developmental Disorders, 29, Hoehn, T. P., & Baumeister, A. A. (1994). A critique of the appli- 213–224. cation of sensory integration therapy to children with learning Baranek, G. T., Foster, L. G., & Berkson, G. (1997). Tactile defen- disabilities. Journal of Learning Disabilities, 27, 338–350. siveness and stereotyped behaviors. American Journal of Occu- Hoshino, Y., Kumashiro, H., Yashima, Y., Tachibana, R., Watanabe, pational Therapy, 51, 91–95. M., & Furukawa, H. (1982). Early symptoms of autistic chil- Bettison, S. (1996). The long-term effects of auditory training on dren and its diagnostic significance. Folia Psychiatrica et Neu- children with autism. Journal of Autism and Developmental Dis- rologica Japonica, 36, 367–374. orders, 26, 361–374. Hughes, C., & Russell, J. (1993). Autistic children’s difficulty Bridgman, G. D., Cushen, W., Cooper, D. M., & Williams, R. J. (1985). with mental disengagement from an object: Its implica- The evaluation of sensorimotor patterning and the persistence of tions for theories of autism. Developmental Psychology, 29, belief. British Journal of Mental Subnormality, 31, 67–79. 498–510. Brown, M. M. (1999). Auditory integration training and autism: Two Humphries, T. W., Snider, L., & McDougall, B. (1993). Clinical eval- case studies. British Journal of Occupational Therapy, 62, 13–17. uation of the effectiveness of sensory integrative and perceptual Bundy, A. C., & Murray, E. A. (2002). Sensory Integration: A. Jean motor therapy in improving sensory integrative function in chil- Ayre’s Theory Revisited. In A. C. Bundy, E. A. Murray & S. dren with learning disabilities. Occupational Therapy Journal Lane (Eds.), Sensory integration: Theory and practice. Philadel- of Research, 13, 163–182. phia: F.A. Davis. Humphries, T. W., Wright, M., McDougall, B., & Vertes, J. (1990). Carmody, D. P., Kaplan, M., & Gaydos, A. M. (2001). Spatial ori- The efficacy of sensory integration therapy for children with entation adjustments in children with autism in Hong Kong. learning disability. Physical and Occupational Therapy in Pe- Child Psychiatry and Human Development, 31, 233–247. diatrics, 10, 1–17. Case-Smith J. (1995). The relationships among sensorimotor com- Humphries, T. W., Wright, M., Snider, L., & McDougall, B. (1992). ponents, fine motor skills, and functional performance in A comparison of the effectiveness of sensory integrative ther- preschool children. American Journal of Occupational Therapy, apy and perceptual-motor training in treating children with learn- 49, 645–652. ing disabilities. Journal of Developmental and Behavioral Case-Smith, J., & Bryan, T. (1999). The effects of occupational ther- Pediatrics, 13, 31–40. apy with sensory integration emphasis on preschool-age chil- Hutt, C., Hutt, S. J., Lee, D., & Ounsted, C. (1964). Arousal and dren with autism. American Journal of Occupational Therapy, childhood autism. Nature, 204, 909–919. 53, 489–497. James, A. L., & Barry, R. J. (1984). Cardiovascular and electro- Clark, F. A., Miller, L. R., Kucherawy, D. A., & Azen, S. P. (1978). dermal responses to simple stimuli in autistic, retarded and A comparison of operant and sensory integrative methods on normal children. International Journal of Psychophysiology, 1, developmental parameters in profoundly retarded adults. Amer- 179–193. ican Journal of Occupational Therapy, 32, 86–92. Johnson, M. H., Siddons, F., Frith, U., & Morton, J. (1992). Can Connolly, K., & Elliott, J. (1989). The emergence of a tool-using autism be predicted on the basis of infant screening tests? De- skill in infancy. Developmental Psychology, 25, 894–912. velopmental Medicine and Child Neurology, 34, 316–320. Efﬁcacy of Sensory and Motor Interventions for Children with Autism 421

Jones, V., & Prior, M. (1985). A comparison of adaptive, verbal McClure, M. K., & Holtz-Yotz, M. (1990). The effects of sensory and motor profiles of psychotic and non-psychotic subnormal stimulatory treatment on an autistic child. American Journal of children. Journal of Autism and Childhood Schizophrenia, 2, Occupational Therapy, 45, 1138–1145. 359–377. Minshew, N. J., Goldstein, G., & Siegel, D. J. (1997). Neuropsycho- Kanter, R. M., Kanter, B., & Clark, D. L. (1982). Vestibular stimu- logic functioning in autism: Profile of a complex information lation effect on language development in mentally retarded chil- processing disorder. Journal of the International Neuro- dren. American Journal of Occupational Therapy, 36, 46–41. psychological Society, 3, 303–316. Kaplan, M., Carmody, D. P., & Gaydos, A. (1996). Postural orien- Mudford, O. C., Cross, B. A., Breen, S., Cullen, C., Reeves, D., tation modifications in autism in response to ambient lenses. Gould, J., & Douglas, J. (2000). Auditory integration training Child Psychiatry and Human Development, 27 81–91. for children with autism: No behavioral benefits detected. Amer- Kaplan, M., Edelson, S. M., & Seip, J. L. (1998). Behavioral changes ican Journal on Mental Retardation, 105, 118–129. in autistic individuals as a result of wearing ambient transitional Neman, R., Roos, P., McCann, R. M., Menolascino, F. J., & Heal, prism lenses. Child Psychiatry and Human Development, 29, L. W. (1974). Experimental evaluation of sensorimotor pat- 65–76. terning used with mentally retarded children. American Journal Kern, L., Koegel, R. L., & Dunlap, G. (1984). The influence of vig- of Mental Deficiency, 79, 372–384. orous versus mild exercise on autistic stereotyped behaviors. Neysmith-Roy, J. M. (2001). The Tomatis method with severely Journal of Autism and Developmental Disorders, 14, 57–67. autistic boys: Individual case studies of behavioral changes. Kern, L., Koegel, L. R., Dyer, K., Blew, P. A., & Fenton, L. R. South African Journal of Psychology, 31, 19–28. (1982). The effects of physical exercise on self-stimulation and Ohta, M., Nagai, Y., Hara, H., & Sasaki, M. (1987). Parental per- appropriate responding in autistic children. Journal of Autism ception of behavioral symptoms in Japanese autistic children. and Developmental Disorders, 12, 399–419. Journal of Autism and Developmental Disorders, 17, 549–563. Kershner, J. R., Cummings, R. L., Clarke, K. A., Hadfield, A. J., & Ornitz, E. M., Guthrie, D., & Farley, A. H. (1977). The early devel- Kershner, B. A. (1990). Two year evaluation of the Tomatis lis- opment of autistic children. Journal of Autism and Develop- tening training program with learning disabled children. Learn- mental Disorders, 7, 207–229. ing Disability Quarterly, 13, 43–53. Ottenbacher, K. (1982). Sensory integration therapy: Affect of effect. Kientz, M. A., & Dunn, W. (1997). A comparison of the performance American Journal of Occupational Therapy, 36, 571–578. of children with and without autism on the Sensory Profile. Perez, J. M., & Sevilla, M. (1993). Psychological assessment of ado- American Journal of Occupational Therapy, 51, 530–537. lescents and adults with autism. Journal of Autism and Devel- Kinsbourne, M. (1987). Cerebral-brainstem relations in infantile opmental Disorders, 23, 653–664. autism. In E. Schopler & G. B. Mesibov (Eds.), Neurobiologi- Polatajko, H. J., Kaplan, B. J., & Wilson, B. N. (1992). Sensory in- cal issues in autism (pp. 107–125). New York: Plenum Press. tegration treatment for children with learning disabilities: Its Klin, A., Volkmar, F. R., & Sparrow, S. S. (1992). Autistic social status 20 years later. Occupational Therapy Journal of Research, dysfunction: Some limitations of the theory of mind hypothe- 12, 323–341. sis. Journal of Child Psychology and Psychiatry, 33, 861–876. Polatajko, H. J., Law, M., Miller, J., Schaffer, R., & Macnab, J. Kohen-Raz, R., Volkmar, F. R., & Cohen, D. (1992). Postural con- (1991). The effect of a sensory integration program on acade- trol in children with autism. Journal of Autism and Develop- mic achievement, motor performance, and self-esteem in chil- mental Disorders, 22, 419–432. dren identified as learning disabled. Occupational Therapy Kootz, J. P., & Cohen, D. J., (1981). Modulation of sensory intake Journal of Research, 11, 155–176. in autistic children: Cardiovascular and behavioral indices. Jour- Porges, S. W. (1998). Love and the evolution of the autonomic ner- nal of the American Academy of Child Psychiatry 20, 692–701. vous system: The Polyvagal theory of intimacy. Psychoneuro- Kootz, J. P., Marinelli, B., & Cohen, D. J. (1982). Modulation of re- endocrinology, 23, 837–861. sponse to environmental stimulation in autistic children. Jour- Rapin, I. (1997). Autism. New England Journal of Medicine, 337, nal of Autism and Developmental Disorders, 12, 185–193. 97–104. Larrington, G. G. (1987). A sensory integration based program with Ray, T. C., King, L. J., & Grandin, T. (1988). The effectiveness of a severely retarded/autistic teenager: An occupational therapy self-initiated vestibular stimulation in producing speech sounds case report. In Z. Mailloux (Ed.), Sensory Integration Ap- in an autistic child. Occupational Therapy Journal of Research, proaches (pp. 101–117). New York: Hawthorn Press. 8, 187–190. Le Couteur, A., Rutter, M., Lord, C., Rios, P., Robertson, S., Hold- Reilly, C., Nelson, D. I., & Bundy, A. C. (1983). Sensorimotor ver- grafer, M., & McLennan, J. (1989). Autism Diagnostic Inter- sus fine motor activities in eliciting vocalizations in autistic chil- view: A standardized investigator-based instrument. Journal of dren. Occupational Therapy Journal of Research, 3, 199–212. Autism and Developmental Disorders, 19, 363–387. Rimland, B. (1964). Infantile autism: The syndrome and its impli- Levinson, L. J., & Reid, G. (1993). The effects of exercise intensity cations for a neural theory of behavior. New York City: on the stereotypic behaviors of individuals with autism. Adapted Appleton Century Crofts. Physical Activity Quarterly, 10, (3), 255–268. Rimland, B., & Edelson, R. (May, 1994). The effects of auditory in- Lincoln, A. J., Courchesne, E., Harms, L., & Allen, M. (1995). Sen- tegration training on autism. American Journal of Speech and sory modulation of auditory stimuli in children with autism and Language Pathology, 18–24. receptive developmental language disorder: Event-related brain Rimland, B., & Edelson, S. (1995). Brief report: A pilot study of au- potential evidence. Journal of Autism and Developmental Dis- ditory integration training in autism. Journal of Autism and De- orders, 25, 521–539. velopmental Disorders, 25, 61–70. Linderman, T. M., & Steward, K. B. (1999). Sensory integrative-based Rinehart, N. J. Bradshaw, J. L., Brereton, A. V., & Tonge, B. J. occupational therapy and functional outcomes in young children (2001). Movement preparation in high-functioning autism and with pervasive developmental disorders: A single subject study. Asperger disorder: A serial choice reaction time task involving American Journal of Occupational Therapy, 53, 207–213. motor reprogramming. Journal of Autism and Developmental Link, H. M. (1997). Auditory integration training (AIT): Sound ther- Disorders, 31, 79–88. apy—Case studies of three boys with autism who received AIT. Rogers, S. J. (1998). Empirically supported comprehensive treat- British Journal of Learning Disabilities, 25, 106–110. ments for young children with autism Journal of Clinical Child Masterton, B. A., & Biederman, G. B. (1983). Proprioceptive versus Psychology, 27, 138–45. visual control in autistic children. Journal of Autism and Devel- Rogers, S. J., Bennetto, L., McEvoy, R., & Pennington, B. F. opmental Disorders, 13, 141–152. (1996). Imitation and pantomime in high-functioning adoles- 422 Baranek

cents with autism spectrum disorders. Child Development, 67, Upledger, JE. (1996). An overview of craniosacral therapy: Its ori- 2060–2073. gin and its applications for newborns and infants. Infants and Scharre, J. E., & Creedon, M. P. (1992). Assessment of visual func- Young Children, 9, 59–68. tion in autistic children. Optometry and Vision Science, 69, Vargas, S., & Camilli, G. (1999). A meta-analysis of research on sen- 433–439. sory integration treatment. American Journal of Occupational Schleien, S. J., Heyne, L. A., & Berken, S. B. (1988). Integrating Therapy, 53, 189–198. physical education to teach appropriate play skills to learners Volkmar, F. R., Cohen, D. J., & Paul, R. (1986). An evaluation of with autism: A pilot study. Adapted Physical Activity Quarterly, DSM-III criteria for infantile autism. Journal of the American 5, 182–192. Academy of Child Psychiatry, 25, 190–197. Schwartz, I. S., Sandall, S. R., Garfinkle, A. N., & Bauer, J. (1998). Wainwright, J. A., & Bryson, S. E. (1996). Visual-spatial orienting Outcomes for children with autism: Three case studies. Topics in autism. Journal of Autism and Developmental Disorders, 26, in Early Childhood Special Education, 18, 132–143. 423–438. Smith, I. M., & Bryson, S. E. (1998). Gesture imitation in autism I: Watters, R. G., & Watters, W. E. (1980). Decreasing self-stimula- Nonsymbolic postures and sequences. Cognitive Neuropsycho- tory behavior with physical exercise in a group of autistic boys. logy, 15, 747–770. Journal of Autism and Developmental Disorders, 10, 379–387. Stagnitti, K., Raison, P., & Ryan, P. (1999). Sensory defensiveness Weber, R. C., & Thorpe, J. (1992). Teaching children with autism syndrome: A paediatric perspective and case study. Australian through task variation in physical education. Exceptional Chil- Occupational Therapy Journal, 46, 175–187. dren, 59, 77–86. Stone, W. L., Lemanek, K. L., Fishel, P. T., Fernandez, M. C., & Wilson, B. N., Kaplan, B. J., Fellowes, S., Gruchy, C., & Faris, P. Altemeier, W. A. (1990). Play and imitation skills in the (1992). The efficacy of sensory integration treatment compared diagnosis of autism in young children. Pediatrics, 86, 267–272. to tutoring. Physical and Occupational Therapy in Pediatrics, Stone, W. L., Ousley, O. Y., Hepburn, S. L., Hogan, K. L., & Brown, 12, 1–36. C. (1999). Patterns of adaptive behavior in very young children Zentall, S. S., & Zentall, T. R. (1983). Optimal stimulation: A model with autism. American Journal of Mental Retardation, 104, of disordered activity and performance in normal and deviant 187–199. children. Psychological Bulletin, 94, 446–471. Stone, W. L., Ousley, O. Y., Littleford, C. D. (1997). Motor imita- Zisserman, L. (1991). The effects of deep pressure on self-stimulating tion in young children with autism: What’s the object? Journal behaviors in a child with autism and other disabilities. American of Abnormal Child Psychology, 25, 475–485. Journal of Occupational Therapy, 46, 547–551. Thelen, E., & Ulrich, B. D. (1991). Hidden skills. Monograph of the Zollweg, W., Palm, D., & Vance, V. (1997). The efficacy of auditory Society for Research in Child Development, 56 (pp. 1–98). integration training: A double-blind study. American Journal of Chicago: University of Chicago Press. Audiology, 6, 39–47.