14 Early Intervention in Autism Geraldine Dawson and Julie Osterling

ONE OF THE most exciting recent achieve- and communication. Other subtypes, such as ments in the field of autism is the ability to Asperger syndrome, are being systematically recognize this disorder at a very early age. A identified (American Psychiatric Association, young child with autism can now be recog- 1994). As our understanding of autism nized by difficulties in orienting to social increases, people with milder forms of autism stimuli, impoverished social gaze, and im- are more likely to be identified. As a result, pairments in the areas of shared attention estimates of the prevalence of PDDs are and motor imitation (Curcio, 1978; Dawson climbing. In 1996, it was estimated to be & Adams, 1984; Dawson, Meltzoff, & approximately 15:10,000 people, three times Osterling, 1995; Mundy, Sigman, Ungerer, the prevalence for autistic disorder. School & Sherman, 1986; Osterling & Dawson, administrators are increasingly realizing that 1994). These difficulties pertain to skills more resources are needed to serve young usually evident during the first year of life children with autism and other PDDs. Fortu- and are likely to be apparent by at least 1 nately, there now is a great deal known about year of age (Osterling & Dawson, 1994). As how best to serve preschool-age children with of 1996, efforts are under way to develop autism and related disorders. standardized methods for very early This chapter describes eight examples of identification of children with autism model early intervention programs for children (Lord, Storoschuk, Rutter, & Pickles, 1993). with autism in the United States that have been In the near future, it is likely that early active since the 1980s. These programs are intervention programs will focus on helping reviewed with two goals in mind. The first is to young infants with autism and their parents. address two questions related to the Since the 1980s, we also have gained a effectiveness of early intervention programs better understanding of the tremendous het- for children with autism1: 1) Are the programs erogeneity of the diagnostic category known effective? and 2) Is effectiveness related to the as pervasive developmental disorder (PDD) type or philosophy of the intervention or to (American Psychiatric Association, 1994). characteristics of the child, such as IQ or We now recognize that autistic disorder is verbal ability? This second question is but one subtype of the general category of particularly pertinent to issues related to PDD, this general category being second-generation research (see Guralnick, characterized by qualitative impairments in 1993, and Chapter 1). Based on the available social interaction

1It should be noted that the information on early intervention provided in this chapter is applicable to both children with autistic disorder and children with other diagnoses in the general category of PDD. Many of the intervention programs we review involved both children with autistic disorder and those with PDD, not otherwise specified (see American Psychiatric Association, 1994). outcome research, many children who receive early intervention program for children with early intervention make substantial autism. developmental gains and are able to be included in a general education classroom by the time they enter elementary school. EIGHT MODEL PROGRAMS: Furthermore, there exists little evidence that APPROACHES AND EFFECTIVENESS the philosophy of the program is critical for The eight preschool programs described in this ensuring a positive outcome as long as certain chapter have published detailed descriptions of fundamental program features are present. their philosophies and approaches. In addition, Finally, although it is likely that all children comprehensive intake and outcome data are with autism will benefit from early available for most programs. A brief intervention, it remains unclear whether rate description of each of the programs is provided of progress is related to child characteristics first. Next, the characteristics of the children such as IQ and language ability. entering the programs are reviewed, followed The second goal in reviewing these pro- by a summary of the outcome data. grams is to describe the common elements of existing early intervention programs for chil- Program Descriptions dren with autism. Despite having different Table 1 provides an overview of some of the philosophical backgrounds and approaches, basic program features. A more detailed de- seasoned clinicians and researchers dealing scription of each program follows. with children with autism are shaped by the Douglass Developmental Disabilities Center common experience of working with these The Douglass Developmental Disabilities children. Out of this common experience have Center program (Rutgers University; Han- come basic shared beliefs and methods for dleman & Harris, 1994; Harris, Handleman, helping children with autism. Unless one Kristoff, Bass, & Gordon, 1990) is based on approach is eventually found to be much more principles of applied behavior analysis and effective than another, these shared beliefs behavior intervention approaches. Three dif- and methods may come to be considered the ferent types of classrooms comprise the pro- most basic and essential features of an gram: the "prep" class, the small group class, and the integrated classroom, with children

Table 1. Program characteristics Program Program emphasis Hours Inclusion Staff/child 1-to-1 work per ratio weeka Douglass Developmentally sequenced 25 After 1–2 years 1:1 to 4:6:8b Yes Health Sciences Developmental/social-pragmatic 22.5 No 3:6 Yes LEAP Typical preschool—programming 15 Yes 3:6:10b Yes May Developmentally sequenced 30 After 1–2 years 1:1 to 3:6:7b Yes Princeton Individualized behavior programs 27.5 No 1:5 Yes TEACCH Structured teaching Varies Varies Varies Yes Walden Incidental teaching 30 Yes 1:3 Yes Young Autism Discrete trial training 40 Later in 1:1 Yes treatment aDoes not include hours of intervention received at home, except for the Young Autism Program, which provides services at child's home. bRatios are given in order as follows: number of staff/number of children with autism/number of typically developing children. typically progressing over approximately a 3- or psychiatrist during which they can discuss year period from the segregated, highly their child's development and behavior and structured prep classroom to an integrated other areas of concern. Monthly parent sup- preschool classroom. The prep classroom port group meetings are offered, and parents provides intensive one-to-one discrete trial are encouraged to observe and participate in training, based largely on Lovaas' model of the classroom as frequently as they wish. instruction (see Young Autism Program) and Learning Experiences . . . An Alternative includes both classroom and in-home in- Program for Preschoolers and Parents (LEAP) struction. The small group classroom maintains The LEAP program (Hoyson, Jamieson, & a 2:1 child–teacher ratio and focuses Strain, 1984; Strain & Cordisco, 1994) has two particularly on skills that will be needed to components: an integrated preschool program function in the integrated classroom. The in- (which includes both typically developing tegrated classroom is based in part on Strain's children and children with autism) and a LEAP model (described below) and serves behavior skills training program for parents. children with autism and those developing The curriculum emphasizes a blend of typical typically. A staff member visits each family at preschool activities (e.g., activity areas and home bimonthly. In addition, parents are in- weekly themes) and activities specifically vited to a monthly or bimonthly "clinic" designed for children with autism. The meeting at which all involved personnel primary goal of the curriculum is to expose discuss the child's progress and areas of children with autism to typical preschool concern. The program also offers parent and activities as much as possible and to adapt the sibling support groups. typical curriculum for the children with autism Health Sciences Center The Health Sciences only when needed. Independent play skills and Center program (University of Colorado; social interaction are facilitated by using peer Rogers & DiLalla, 1991; Rogers, Herbison, models and by prompting, fading, and Lewis, Patone, & Reis, 1986; Rogers & Lewis, reinforcing target behaviors. 1989) serves children with autism and PDD as The behavior skills training program well as children with varied behavior disorders. teaches parents techniques for managing their The curriculum is based on developmental child's behavior and teaching new skills. A principles and emphasizes play as an important family service coordinator works with the medium through which socioemotional, family to teach specific target behaviors in a communicative, and cognitive development are natural context, such as the home or com- facilitated. The program also emphasizes munity. Families are also invited to attend a eliciting positive affect within the context of monthly parent support group. social interaction to increase children's attention May Institute The May Institute (Boston; to others and motivation to engage in Anderson, Avery, Dipietro, Edwards, & Chris- relationships. Social relationships are facilitated tian, 1987; Anderson, Campbell, & Cannon, by assigning each child a primary teacher, by 1994) offers a developmentally sequenced fostering close proximity to and imitation of program based on principles of applied behavior peers, and by modeling and prompting specific analysis and behavior intervention approaches. social behaviors. For children who comprehend When children enter the program, they and their complex spoken language, teachers also provide families receive intensive (15 hours per week) in- verbal explanations in order to help the children home training for a period of 6 months. During better understand other people's emotions, this period, the in-home therapist and parents behavior, and perspectives. When appropriate, provide one-to-one instruction focusing on children are provided with occupational basic skills such as play, self-care, language, therapy, individual psychotherapy, or both. and the reduction of problem behaviors. After Parents in the program are offered 1-hour the children complete the home-based weekly consultation with a child psychologist treatment, or concurrent with this treatment, they attend one of the Institute's two are revised. Twice a month, each family is preschool programs, the "Step 1" classroom visited at home by a home programmer who helps or the integrated classroom. Most children the family implement behavior programs that the attend the Step 1 class for approximately 1 child has successfully achieved at school in year, where basic classroom skills such as order to maximize generalization of skills from following instructions, imitation, and school to home. Parent meetings are held working in groups are taught. The Step 1 regularly at the school. classroom contains only children with autism, and instruction usually takes place in highly Treatment and Education of Autistic structured, teacher-directed small groups. The and Communication-Handicapped Children integrated classroom includes children with (TEACCH) The TEACCH curriculum developmental disabilities (mostly autism) as (University of North Carolina at Chapel Hill; well as typically developing children. The cur- Lord & Schopler, 1994) emphasizes two basic riculum focuses on teaching skills that the principles: structuring the environment to children will need in a general kindergarten promote skill acquisition and facilitating classroom. independence at all levels of functioning. Children are taught new, developmentally Families are visited at home by their child's appropriate skills in a highly structured service coordinator once a month. During teaching environment in which one-to-one these visits, the parents and service coordi- instruction is offered. Once a skill is estab- nator discuss the child's progress and parents' lished, the children are taught to use the skill in areas of concern. Once a month, the program a less structured environment with less offers parent support groups and educational support from adults so that they can gradually discussion sessions led by program staff. The gain increasing independence. Independence May Institute also provides other services, is also fostered by the structured teaching such as respite care for families and outside environment, which relies heavily on visual referral information. cues such as the use of "start and finish Princeton Child Development Institute The boxes" to signal the beginning and end of Princeton Child Development Institute program activities. The TEACCH program, which (Princeton University; Fenske, Zalenski, Krantz, strives to provide the least restrictive teaching & McClannahan, 1985; McClannahan & Krantz, environment, is used throughout North 1994) is based on principles of applied behavior Carolina in a variety of different school analysis and behavior intervention approaches and settings and in a variety of different ways. serves only children with autism. Each child is Consultants visit each child's classroom on a first evaluated so that a set of individualized regular basis to provide information on how to behavior "programs" can be prescribed. Most adapt each classroom to the needs of each child children's first set of behavior programs targets and his or her family. basic skills such as following simple instructions, Walden Preschool The Walden Preschool motor and verbal imitation, toileting, and curriculum (Emory University School of matching skills. Over the course of the school Medicine; McGee, Daly, & Jacobs, 1994) fo- day, the children work in different activity areas cuses on language and social development and rooms and with different staff members in using an incidental teaching method (Hart & order to promote g e n e r a l i z a t i o n o f Risley, 1975; Kaiser, Yoder, & Keetz, 1992), the r e q u i r e d skills. Children are taught to use basic tenets of which involve giving the child picture schedules to assist them with the many opportunities to practice a skill during transitions they are required to make during the interactions between an adult and child that day. Picture schedules also assist the children in spontaneously arise in natural contexts such as learning to initiate activities, make choices, and free play. The classroom consists of a number become more independent. Periodically, the of "teaching zones" (usually four) that are children's progress is assessed and specific goals organized around different teaching goals and preschool, if possible). The third year emphasizes are designed to naturally attract children to appropriate expression of emotion, preacademic desired activities and materials. Each teaching tasks, and observational learning. Aggressive and zone is staffed by at least one teacher who self-stimulatory behaviors are handled by ignoring, adapts the teaching goals to each child's time-out, shaping, and (as a last resort) delivery of developmental level and facilitates the a loud "no" or slap on the thigh. acquisition of target skills. Children are free to move from one zone to another; one teacher's Characteristics of Children primary responsibility is to help children make Enrolled in the Programs transitions from one zone to another. The Table 2 provides an overview of the charac- preschool classroom includes children with teristics of children who participated in the autism and typically developing children. preschool programs. Although the assessment Parents are offered a number of options for measures varied across the programs, virtually how they would like to maintain contact with all programs administered on a regular basis the school, which range from home visits to some form of standardized assessment regular meetings with the teaching staff. designed to measure children's verbal and nonverbal abilities, autistic symptoms, and Young Autism Program The Young Autism adaptive behaviors. Some of the programs also Program (University of California at Los Angeles assessed other developmental domains, such as [UCLA]; Lovaas, 1987; Lovaas, Smith, & play, motor, and social skills. McEachin, 1989; McEachin, Smith-Tristram, & The average age of the children entering the Lovaas, 1993) is based on principles of applied programs was 3.5-4 years. This is the average behavior analysis and behavior intervention age at which an initial diagnosis of autism is strategies. In the initial phases of the program, first made (Siegel, Pliner, Eschler, & Elliott, children receive intensive one-to-one discrete trial 1988). Thus, it is likely that the children in this training for 40 hours per week. Parents are trained review began treatment at an age typical for to use the treatment procedures so that the inter- most children in the United States. The vention can take place during most of the child's children in the Young Autism Program at waking hours. In the first year, emphasis is placed UCLA, however, were typically younger than on reducing self-stimulatory and aggressive the children in the other programs; the behaviors, building compliance, imitation, and average age at intake for Lovaas' program was appropriate toy play. The second year emphasizes 2 years 8 months (McEachin et al., 1993), expressive and abstract language and interactive whereas the average age for the children in the play with peers. The children also are taught how other programs was 3 years 9 months. to function in a preschool group by enrolling them in a preschool program (in a general education

Table 2. Characteristics of children enrolled in the programs Program No. of children Mean age (months) Full-Scale Diagnosis (range) IQ(range) Douglass 36 50(30-62) 61 (36–105) Autism Health Sciences 49 46 70 (nonverbal) Autism or PDD LEAP 48 43 (30–64) 61 (6–119) Autism or PDD May 42 47 (36–62) 49 (37-71)a Autism or PDD Princeton 32 43 (30–58) 57 (36–83) Autism Walden 27 44 (30–66) 57 (29–91) Autism Young Autism 19 32 53 (30–82) Autism aIQ available for only 10 children. At intake, the typical Full-Scale IQ score for were able to function in a general education children in the programs fell in the mid-50s, setting. However, the nature of the general ranging from 49 to 70 on average, with most education placement and the extent to which programs reporting that a small number of their special support services were needed to main- children were not testable at intake. This range of tain a successful placement varied consider- functioning is consistent with that of the general ably. population of preschool children with autism. The The type of support services children re- beginning IQ levels of the children in different ceived when they were placed in a general programs were fairly similar. Most programs education setting also varied considerably. used the Stanford-Binet Scale (Terman & Merrill, Unfortunately, there often was little detailed 1960) to assess Full-Scale IQ, but some programs information regarding the extent and type of used measures that have a stronger nonverbal support children received. It is also important loading (e.g., Merrill-Palmer, McCarthy scales). to keep in mind that policies of a particular Because children with autism tend to perform school district likely influenced the extent to better on non-verbal tasks, it is difficult to compare which a given child was able to be included in a the intake IQs of children across the different pro- general education classroom. Consequently, grams. For instance, at intake, the children at children with equal levels of abilities may have the Health Sciences Center at the University of achieved different placement outcomes based on Colorado had an average IQ of 70, which is where they lived. For these reasons, it is not higher than the average IQs of the other pro- possible to meaningfully compare the different grams. However, IQ in this program was mea- programs' effectiveness based on the placement sured using primarily nonverbal IQ tests. data. Approximately half the programs accepted Six of the eight programs characterized child children with autistic disorder or PPD, not outcome in terms of specific developmental otherwise specified, whereas the other half gains made while in the program. The pro- admitted only children with a diagnosis of grams described the developmental gains of the autism. Most programs used Diagnostic and children in a variety of ways (e.g., IQ scores, Statistical Manual of Mental Disorders (DSM-III- developmental scores on standardized tests, R) criteria developed by the American observational measures taken in the class- Psychiatric Association (1994) in their initial room). Comparing the outcomes of different diagnostic evaluations, and some programs programs is problematic because of the used the Childhood Autism Rating Scale diversity of the measures used. (Schopler, Reichler,. DeVellis, & Daly, 1980) as an additional component in the diagnosis. General Conclusions Regarding the Effectiveness of Early Intervention Child Outcome Although a direct comparison of the effec- Outcome data are available for approximately tiveness of the different programs is not 150 children who completed one of the possible, combining information across pro- programs. All programs have been evaluated grams provides a general view of the overall using outcome data, but program directors progress of the 150 children who received described their children's outcomes in different early intervention. We can conclude that, de- ways. Some primarily emphasized the types of spite somewhat diverse intervention strategies post-preschool placements achieved by their and philosophical approaches, all of the pro- children, whereas others also described grams were quite effective in fostering positive specific developmental gains made by their school placements, significant developmental children. A summary of these data is provided gains, or both for a substantial percentage of in Table 3. Of the eight programs, six provided their students. Of the six programs that re- placement data for their graduating children. A ported placement data, four reported that substantial number of children in the programs approximately 50% of children were able to be Table 3. Child outcomes

Program Placement outcomes Developmental gains Douglass 3/21 children "fully integrated" A nearly 19-point increase in IQ after 1 16/21 in classes for children with neurological year of treatment communication disabilities Health Sciences Significant increases in language, cognitive, social, and motor skills Doubling of developmental rate in several areas, reaching typical rates Significant increases in language, cognitive, and motor skills Doubling of developmental rate in several areas LEAP Approximately 50% in "regular education classes" May 14/26 children "mainstreamed," but many required "specialized supports" such as an aide Princeton 12/24 children in "public schools" Nonverbal children identified by 3 years gained an average of 22-24 IQ points by 7 years of age TEACCH Those identified by 4 years gained 15- 19 IQ points by 9 years of age Walden 12/14 children "fully mainstreamed" Observational measures indicated that language use tripled Young Autism 9/19 children attended general first-grade classes By first grade, children gained an 8/19 children in classes for children with aphasia average of 20 IQ points, with 12 of 19 falling in the average range 47% of original 19 children in general classes at 13 years of age

integrated into a general classroom by the end ences in the philosophical approaches of the of the intervention. When outcome was assess- programs, these are the elements that most ed in terms of IQ, it was found that children program directors believe are essential, al- made, on average, an IQ gain of approximately though their methods for addressing each el- 20 points. Investigators did not report whether ement vary. The assumption is that, to a large a positive response to intervention was related extent, these elements emerged from many to specific child characteristics, such as IQ or hours of working with the unique challenges language ability. It should be noted, however, that a child with autism presents. In this that most children participating in the pro- sense, they can be considered the "tried and gram scored as having some level of mental true" features of early intervention for retardation (IQ less than 70) at the beginning children with autism; they are unlikely to re- of intervention. Thus, despite having autism flect an idiosyncratic viewpoint or one inves- and considerable cognitive delay, approxi- tigator's philosophical attitude. Therefore, mately half of the children responded very these are the elements that parents should positively to early intervention, and all or at reasonably expect a school system to provide least most children reportedly made significant for their child. gains. To what extent can the reported gains be Element One: Curriculum Content attributed to early intervention? Unfortunately, The curricula reviewed previously typically only one program published data on a control emphasize five basic skill domains. The first group of children with autism who did not skill domain focuses on the ability to attend to participate in the early intervention program elements of the environment that are essential for (Lovaas, 1987; Lovaas et al., 1989), and, to learning, especially to other people, and to com- date, there have been no true experimental ply with teaching demands. As was described studies that would require, at minimum, random previously, autism is characterized by a failure to assignment to different intervention groups and selectively attend to social stimuli, including outcome assessments conducted by people facial expressions, gestures, and speech naive with regard to intervention status. (Dawson et al., 1995). Thus, not only does the Despite its methodological limitations (e.g., child with autism have difficulty interpreting lack of random assignment to intervention vs. social stimuli, he or she is failing to attend to control group), the Lovaas study is very en- them in the first place. This is an especially couraging, especially in light of the 1993 difficult problem because knowledge of the findings that the positive effects of this world often comes directly from other people. intervention have been maintained through the Therefore, the first and often most difficult skill elementary school years (McEachin et al., 1993). to teach the young child with autism is to pay More specific answers to questions such as attention to other people. Children with autism whether one intervention approach is more also may have difficulty shifting their attention effective than another and what is the optimal from one stimulus to another (Courchesne et al., intensity of intervention will have to await the 1994), attending to more than one stimulus at a next generation of intervention research. time (Koegel & Schreibman, 1977), and sharing attention states with others (Mundy et al., 1986). Many programs address these attention pro- COMMON ELEMENTS OF THE blems as well (Lord & Schopler, 1994; Lovaas, PROGRAMS: THE "TRIED AND TRUE" 1987; Rogers & DiLalla, 1991). However, the This section describes several elements of programs vary in .how they increase attention early intervention for children with autism skills. The Young Autism Program emphasizes that are common to virtually all of the pro- discrete trial learning, whereas the Walden grams reviewed. Despite considerable differ- Preschool uses incidental teaching strategies in which staff structure the environment to in- rewarding responses to the child's smallest crease the child's chance of attending to appro- attempt to communicate, be it a glance, a vocal- priate stimuli. ization, or the slightest body movement (McGee The second skill domain is the ability to imitate et al., 1994). In this way, the child with autism is others, including both verbal imitation and motor able to achieve one of the most important and imitation. Verbal imitation often is easier for challenging early goals: grasping the concept of children with autism; some of these children communicative intent. Furthermore, it is recog- spontaneously echo what others say to them. A nized that the child with autism needs early couple of decades ago, echolalia was considered exposure to a range of communicative modal- pathological, and teachers would discourage ities, including the use of visual symbols, gest- echolalic speech. Based on careful longitudinal ures, written word, and speech (Prizant & research on the development of language in Wetherby, 1989). autism, we now know that echolalia is a critical The fourth skill domain is the ability to play first phase of early language acquisition for appropriately with toys. It is widely known that many of these children (Prizant & Wetherby, children with autism play with toys in idio- 1989). syncratic, repetitive ways (Jarrod, Boucher, & In contrast to verbal imitation, most children Smith, 1993) and that their imaginary play is with autism fail to spontaneously imitate the mo- often severely delayed or absent (Riguet, Taylor, tor actions of others (Curcio, 1978; Dawson & Benaroya, & Klein, 1981; Ungerer & Sigman, Adams, 1984). Motor imitation is an especially 1981). Intervention strategies for facilitating play important skill because it is a fundamental mech- development focus on increasing the child's anism for learning. Social skills, in particular, are ability to use toys in functionally appropriate often learned by imitating others. From a develop- ways and promoting the use of symbolic play. mental perspective, imitation is intimately linked Research has shown that children with autism to the development of representational thought often are capable of functional and symbolic play (Piaget, 1962). Furthermore, imitation serves when specific prompts are used (Lewis & many social functions, including providing a Boucher, 1988). Most intervention strategies context of mutuality in which communication and capitalize on this fact by using physical and verbal empathy may develop (Dawson & Lewy, 1989; prompts to facilitate toy play. Meltzoff & Gopnik, 1993; Uzgiris, 1981). Finally, the fifth skill domain is the ability to Approaches to teaching imitation range from socially interact with others, especially with peers direct teaching via operant conditioning methods (see Chapter 23). Interactive play with peers often (Lovaas, 1987) to facilitating spontaneous imita- is considered a more advanced skill domain for tion via structured social play (Rogers & DiLalla, young children with autism. Social skills are 1991). inherently complex and involve the core impair- The third skill domain is the ability to com- ments of children with autism. Some programs prehend and use language (see Chapter 16). As begin by teaching specific skills, such as turn with imitation, approaches to facilitating language taking and sharing, in the context of adult-child acquisition differ. All approaches, however, interaction and then systematically generalize recognize the motivational issues involved in these skills to child-child interactions. Many language acquisition in autism (Koegel & programs also recognize the effectiveness of using Mentis, 1985). Children with autism typically are typically developing peers as facilitators of social not motivated to communicate for the sake of behavior (Strain & Cordisco; 1994). sharing information and experiences (Tager- Flusberg, 1989, 1993). The child must be "tempt- Element Two: Need for Highly Supportive ed" to communicate by capitalizing on the natural Teaching Environments and Generalization desires and preferences of the child (Koegel & Strategies Johnson, 1989). Most early intervention strat- In most programs, core skills first are established egies involve providing immediate and in highly supportive teaching environments and then are systematically generalized to more com- attention to relevant information for learning and plex, natural environments. Core skills refer to success in learning. those skill domains that are inherently part of the An important issue that must be addressed, autistic syndrome and are critical for the acqui- however, is the fact that children with autism often sition of knowledge (also referred to as "pivotal show large discrepancies between their perfor- skills" by Koegel & Koegel, 1988). Not surp- mance levels in highly supportive versus more risingly, these are the skills listed previously, natural, complex environments. When specific which are the focus of most curricula: attention prompts and familiar cues are available, the child and compliance, imitation, communication, appro- with autism is capable of much higher levels of priate toy play, and social skills. In virtually all performance, even in skill domains that are consid- programs, these skills are first taught in highly ered core impairments in autism, such as toy play structured contexts in which the child interacts (Lewis & Boucher, 1988) and social responsive- directly with a trained therapist or teacher in an ness (Dawson & Galpert, 1990;Klinger & Dawson, environment that minimizes distraction and max- 1992). Indeed, increasing spontaneous perform- imizes attention to specific stimuli. Features of the ance of previously learned behaviors in complex, environment that serve to maximize attention to natural environments is one of the major hurdles relevant stimuli include repetition, predictability, faced by those working with individuals with and salience (e.g., by increasing the intensity of the autism. stimuli, by placing the stimulus directly in the Interestingly, typically developing children also child's visual field). Furthermore, explicit support show a significant discrepancy between their for a newly acquired behavior is provided by ano- competencies in different contexts. Fischer and ther person (e.g., by modeling the behavior or by colleagues refer to these two very different levels providing direct physical guidance or prompts). It of competence as functional versus optimal levels is important to note that all programs have a very (Fischer, Bullock, Rotenberg, & Raya, 1993). low staff-to-child ratio in order to provide this kind They argue that all behavior arises from collabor- of intensive teaching. Ratios typically are 1:1 or ation of person and context. High-support contexts 1:2, especially during the first part of the inter- evoke optimal levels of performance, whereas low- vention. support contexts evoke much lower, functional At first glance, advocating such an intensive, levels of performance. Thus, educators who are highly supportive teaching environment for child- assessing a child's abilities must think in terms of a ren with autism may seem counterproductive, range of behaviors and must carefully consider especially because it is now known that many context as an integral part of any competence they children with developmental disabilities are quite assess. capable of gaining skills in more complex environ- To understand why children with autism have an ments that typically involve small or even large unusually large discrepancy between their funct- groups of children. However, this may not be the ional and optimal performance levels, it may be case for a child with autism. We speculate that useful to consider how typically developing child- children with autism have such severe information- ren eventually come to be able to perform newly processing impairments that they have difficulty acquired behaviors in low-support environments. attending to and encoding critical information, Fischer et al. (1993) suggest that high-support con- such as gestures, language, and facial expressions, texts serve to activate complex neural systems when such information is presented in a highly through short-term memory processes. At first, a complex environment. The child with autism newly acquired skill is fragile because it is appears to need what has been referred to as an dependent on induced short-term memory com- augmented scaffold (Klinger &, Dawson, 1992). ponents that are not yet subject to long-term Scaffolding, a term long used in the developmental memory encoding. The input provided by the high- psychology literature, refers to the process support context pushes the neural system into a whereby parents and teachers naturally help to higher level state that is not available spontan- structure the child's environment to maximize eously. The state itself cannot yet be regenerated alone without appropriate contextual input. Even- Element Three: Need for tually, the child encodes the learned behavior (and Predictability and Routine perhaps aspects of the high-support environment One of the symptoms of autistic disorder is a strict as well) in long-term memory. As a result, the adherence to specific routines and need for same- child becomes able to spontaneously regenerate the ness in the environment. This feature of the behavior at a later time in the absence of immed- syndrome was recognized as early as 1943 in the iate context support. We suggest that it is such original writings of Kanner. Although all children, encoding and long-term memory processes that are especially those with developmental disabilities, impaired in children with autism and that such thrive on routine and predictability, the child with impairments help to explain the large discrepancy autism seems especially sensitive to changes in the between their optimal and functional performance environment and routine. Studies have shown that levels. There is ample evidence that children with children with autism become more socially respon- autism have not only attention problems but sive and attentive when information is provided in significant visual and auditory memory limitations a highly predictable manner and, conversely, that as well (Ameli, Courchesne, Lincoln, Kaufman, & their behavior is severely disrupted when the same Grillon, 1988; Boucher & Warrington, 1976; stimuli are presented in an unpredictable manner ( Klinger & Dawson, 1995; Minshew & Gold-stein, Dawson & Lewy, 1989; Ferrara & Hill, 1980). In 1993). the classroom, children with autism have difficulty Several strategies can be used to minimize with transitions from one activity to another and discrepancies between optimal and functional with any unanticipated change from expectations, levels of performance in children with autism. The such as a substitute teacher or field trip. It is not programs reviewed previously typically provide a yet understood why children with autism have high-support environment for the acquisition of such a strong need for predictability and routine. It new behaviors and then gradually fade the high has been theorized, however, that it stems, in part, level of contextual support and systematically gen- from difficulties in arousal modulation (Dawson & eralize the newly acquired behaviors to more com- Lewy, 1989) and from impairments in memory plex, natural environments. The high-support en- (Klinger & Dawson, 1995) and in processing of vironment is achieved in a variety of ways, ranging temporal information (Hermelin & O'Connor, from methods that stress specific periods of one- 1970). to-one teaching in a distraction-free environment The programs reviewed adopt a number of (Lovaas, 1987; Rogers & DiLalla, 1991) to those different strategies to assist the child with autism that capitalize on brief intensive teaching opport- with the inevitable changes in routine and activet- unities that spontaneously occur in the classroom ies that a typical day involves. First, most throughout the day (McGee et al., 1994). General- programs are highly structured and routine. ization of skills to low-support environments also Programs that are more child directed and com- can be achieved in a variety of ways. Fading the plex, such as the Walden Preschool Program, rely level of prompts while gradually increasing the on the assistance of several teachers who systemat- level of environmental complexity is a common ically facilitate each child's use of the environment strategy. For example, prompts may initially and transitions from one activity to another. In involve handover-hand guidance and eventually other programs, visual cues, such as colored move to modeling of the learned behavior. Social shapes, photos, and written words, often are used behaviors initially may be taught in the context of to label and define specific activities and activity adult-child interactions and eventually involve one centers. Specific strategies for minimizing the or more peers. Like other programs, the TEACCH child's distress and confusion during transitions program advocates changing only one feature of from one activity to another are often imple- the environment at a time during the process of mented, including giving ample warning, walking generalization. the child through the transition, providing a transitional object and other visual aids, and carrying out a familiar ritual during the transition. In addition, visual daily schedules, in the form of to teach the child appropriate behaviors to cope pictures or written words, are used in many pro- effectively with the situation, with particular emp- grams. hasis on teaching the child communication skills. In the previous example, the child may be pro- Element Four: A Functional vided with a board that shows two activities for the Approach to Problem Behaviors child to choose from so that his or her choice- Young children with autism often display very making skills are promoted. In addition, the child challenging behavior. The preschool curricula rev- may be taught how to request help through words, iewed previously all address this issue. The goal pictures, gestures, and/or manual sign. Teachers advocated by most programs is to prevent the may have the child practice the new requesting development of problem behaviors. Most pro- skill and then prompt the skill at appropriate times grams achieve this by increasing the children's in the classroom before they expect the child to interest and engagement in classroom activities. In request help on his or her own. the Walden Program, this approach is described as "increasing fun decreases behavior problems" Element Five: Transition (McGee et al., 1994, p. 146). For example, chil- from the Preschool Classroom dren are provided with choices regarding in which As discussed previously, children with autism have activities they want to engage, and attempts are difficulty making changes easily and generalizing made to include many activities that involve highly previously acquired skills to new environments. preferred play materials or topic areas. Another Given these difficulties, it is understandable that strategy for engaging the child in classroom act- the transition from preschool to kindergarten or ivities is to provide a highly structured classroom first grade is challenging for many children with environment, thereby preventing behavior prob- autism. Most early intervention programs reviewed lems by increasing the children's understanding of recognize this transition as a critical point in the the classroom routine and specific activities and by children's education and consequently devote a promoting the children's independence and success great deal of time and effort to make the transition (see, especially, the TEACCH program). as successful as possible. Among the programs, When problem behaviors persist, this typically is there is general agreement regarding how best to handled by carrying out a functional assessment of facilitate a successful transition to placements the problem behavior and teaching alternative beyond preschool. Descriptions of these methods appropriate behaviors that serve the same function follow. for the child. This approach has been described in It is generally agreed that one of the most impor- detail by Donnellan, Mirenda, Mesaros, and Fass- tant factors in preparing children for post- bender (1984) and others. It usually involves three preschool placements is teaching them to function steps. First, a detailed record of the behavior is as independently as possible in the classroom. Be- made, including the situations in which the beha- cause independence skills are acquired over a vior occurred, time of day, events that preceded the period of years, most programs begin to systemat- behavior, and how others responded. Based on ically teach them as soon as a child enters the patterns apparent in the behavior record, hypot- preschool program. Powers (1992) suggested the heses are then generated regarding the function the following beginning list of "survival" skills that behavior serves for the child. Often this is a children need in order to function independently in communicative function ("I need help," "I don't a general classroom: like this"). For example, the child who throws toys • Complying with adult requests only during free play time may be requesting help • Taking turns from teachers, perhaps because he or she becomes • Listening to directions from afar or near confused during the unstructured free play period • Sitting quietly during activities and has not learned to find and start a new activity • Volunteering on his or her own. The third step is to change the • Raising one's hand to solicit attention environment to support appropriate behavior and • Walking in line • Using toilets in classroom versus in the kindergarten or first grade is increased. In this hallway way, the child's performance in the new placement • Picking up toys after use can be evaluated, and skill impairments can be • Communicating about basic needs identified. Then, the child can work on refining the skills needed for his or her new placement in the Children who possess these skills are much more familiar and safe environment of the preschool likely to function successfully in a general edu- classroom. In this way, the child is given the cation environment and to have access to a wider greatest chance of achieving success in his or her range of post-preschool placements. McClannahan post-preschool placement. and Krantz (1994) suggested several more advan- ced classroom skills, such as following through Element Six: Family Involvement with adult instruction, sustaining engagement in Parents' involvement in the education of their child work activities, using a delayed reinforcement sys- with autism has been viewed as an important factor tem, and moving from one activity to another with for success since the 1980s. On the heels of an minimal assistance. unfortunate era in which "refrigerator parents" Most programs take an active role in assisting were seen as the cause of autism, Schopler and parents and school districts in finding a placement Reichler (1971) were among the first to advocate for each child. Teachers help investigate the pos- parents as "co-therapists" in interventions for chil- sible placement options and assist in finding the dren with autism. In the programs first designed by placement that is the best match for the child based Schopler, who eventually established the on factors such as class size, degree of classroom TEACCH program, parents were taught basic structure, and teaching style. Once a placement is techniques of developmental therapy, participated found, preschool staff often visit the classroom and in therapy with their child and a staff therapist, and make an inventory of the skills the child will need continued therapy at home. Schopler and Reichler in order to function as independently as possible in found that the children made significant progress the new placement. These skills are then incorp- and that they tended to demonstrate higher devel- orated into the child's preschool curriculum goals opmental skills when interacting with their parent and are systematically taught, practiced, prompted, than when interacting with the therapist. Because and refined. McClannahan and Krantz (1994) pro- parents spend so much more time with their chil- vided several examples of such skills, which might dren, it is recognized that they often can achieve a include requesting a bathroom pass, using a cubby, greater understanding of their child's needs and and saying the new teacher's name. As part of their provide unique insight into creating an intervene- efforts in preparing children for post-preschool tion plan. By including parents in the interventions placements, the Princeton program encourages with children, greater maintenance and general- parents to engage their children in integrated ization of skills also can be achieved. Also, in- settings such as religious schools, dance or gym- cluding parents in the interventions with young nastics classes, and after-school recreation prog- children with autism can increase parents' feelings rams. These preliminary settings are used to eval- of relatedness with their child and increase their uate the child's ability to function in an integrated sense of competence as parents, thereby decreasing setting and to work on skills that he or she will emotional stress and facilitating well-being (see need in order to be successful. Most preschool Chapters 12 and 22). programs also assist in training the post-preschool In keeping with Schopler and Reichler's (1971) staff. For instance, staff from the May program advocacy, all of the programs recognize that provide an 8-day workshop to teachers and others parents are a critical component in early inter- who will be working with their graduating chil- vention with children with autism. All devote time dren. Finally, the transition from preschool to the and resources to training parents and including next placement is often accomplished gradually. them in the educational process. This is achieved Children's time in the preschool classroom often is in different ways. Some programs require signi- systematically decreased, whereas their time in ficant time commitments on the part of the family and strongly endorse the parents' use of specific classroom (Strain & Cordisco, 1994). Most pro- behavior strategies to increase the children's skills. grams emphasize the need to help the child with Most programs, however, allow parents to choose autism develop independence, initiative, and how they want to be involved and which skills choice-making skills. they want to work on at home. There is variability across programs regarding whether parent training FUTURE DIRECTIONS is conducted at home, at school, or both. All pro- grams are sensitive to the stresses often encoun- Suggestions for Improving Second-Generation tered by families who have a child with autism and Research on Early Intervention provide parent groups and other types of emotional One of the most promising findings in this review support. was that all eight early intervention programs demonstrated notable gains by children with aut- Intensity of Intervention ism. Because the level of success achieved across and Other Common Elements programs was fairly similar, conclusions regarding We believe it is important to emphasize that all recommended practices for young children with but one of the programs involve at least 20 autism stem from an analysis of the programs' hours a week of school-based intervention (the common features. There are, however, limitations LEAP program is 15 hours per week), and so- regarding the conclusions that can be made at this me involve a substantially greater number of point concerning the effectiveness of early inter- hours. The range of school-based hours per vention in autism. For example, no firm conclu- week is from 15 to 40, with an average of 27 sions can be drawn yet regarding which types of hours. Although it is not possible to conclude interventions work best for different children (e.g., from these data what is the necessary and verbal versus. nonverbal) or how much intervene- sufficient number of hours per week of school- tion is needed in order to be maximally effective. based intervention required for a positive Such issues might have been addressed using the outcome, this information does offer an est- statistical technique of meta-analysis if invest- imate of the number of school-based service igators had utilized similar methods for char- hours needed by young children with autism. acterizing individual differences in children (e.g., Unfortunately, 27 hours is a substantially gre- standardized IQ scores) and for quantifying pro- ater amount of intervention than that often gress. provided by most public school systems during To facilitate future analyses of the effectiveness the preschool period. Moreover, all programs of early intervention in autism, investigators sho- included a parent training component, and, uld consider standardizing measures of cognitive, therefore, the number of hours of intervention language, social, and adaptive abilities and that children are actually receiving is likely measures of rates of developmental change across substantially greater than the number of programs. This would allow for a direct compare- school-based hours provided. ison across different programs and provide infor- In addition to the six elements described that mation regarding the effectiveness of a specific are common to most programs reviewed, several type of intervention for children with different other features were often observed. These include levels of ability. In addition, investigators rarely the use of a range of augmentative communication provided adequate information regarding the type methods, occupational therapy services, and an and amount of support services that children emphasis on the development of trusting, positive received when they graduated to general education social relationships (rather than focusing solely on environments. This made it difficult to evaluate the acquisition of specific social behaviors) (e.g., what was meant when it was reported that the child Rogers & DiLalla, 1991). Furthermore, many pro- was successfully included in a general education grams advocate the use of typically developing classroom. Thus, it is also recommended that in- peers as promoters of social behavior for children vestigators consider standardizing the way that with autism and as positive role models in the post-preschool placements and support services are characterized. Finally, in an era when federal and infant also allows the infant to control the amount state funding of early intervention is in jeopardy, of stimulation experienced by placing the infant in including a careful cost–benefit analysis would be the role of initiator (Field, 1977, 1979). Studies useful (see Lovaas, 1987). have shown that children with autism exhibit inc- reased amounts of social attention and respon- Intervention Methods siveness when they are imitated (Dawson & for Infants with Autism Adams, 1984; Dawson & Galpert, 1990; Tieger- As the ability to identify infants with autism man & Primavera, 1981, 1984). improves, so will the need to develop methods for Imitation by parents has been a component of intervening with young infants with autism and other early intervention programs as well. their parents. Although it is likely that such inter- Mahoney and Powell (1988) developed a parenting ventions will incorporate most, if not all, of the program for children with developmental disa- elements described previously, they will also have bilities from birth to 3 years old (Transactional features that are specifically tailored to the needs Intervention Program). The goal of the program of a young infant and his or her parents. was to increase parental use of turn-taking skills by This section briefly describes the early interven- imitating the child's behavior and following the tion research conducted by Dawson and her collea- child's lead. Children of parents who were high in gues. Elsewhere, Dawson and her co-workers implementation of the program made greater deve- (Dawson & Galpert, 1990; Dawson & Lewy, 1989; lopmental gains than children of parents who were Klinger & Dawson, 1992) have discussed in depth low in implementation. These gains were mediated one approach to very early intervention with young by affective characteristics of the parents, with children with autism. Because these strategies are children of high-affect parents displaying the based on typical patterns of parent-infant inter- greatest gains. action, they are particularly well suited for use Elaborating on the basic technique of imitation, with infants with autism and their parents. Dawson Klinger and Dawson (1992) developed a series of (1991; Dawson & Lewy, 1989) has suggested that, social interactive strategies designed to promote like typically developing infants, young children early social skills in young children with autism, with autism have a narrow range of optimal stim- including contingency, social gaze, turn taking, ulation. Their ability to make sense of and relate to imitation, shared attention, and communicative in- their environments depends in large part on the tent. These strategies are based on the following amount of regularity and familiarity provided. five principles: Based on these assumptions, Dawson developed a 1. They are modeled from naturally occurring series of intervention strategies designed to allow patterns of early social interaction. As in typ- children with autism to maximize optimal levels of ical development, social skills are facilitated stimulation and to increase their attention to and through play rather than explicitly taught. comprehension of social information. 2. They are based on knowledge of typical devel- The first strategy involves closely following the opmental sequences, progressing from very child's lead, usually by imitating the child's beh- simple interactions to increasingly complex so- avior. Imitation of an infant's behavior is one of cial interactive skills. the most common forms of communication bet- 3. They incorporate the principle of scaffolding, ween a parent and infant during the first year of whereby parents provide stimulation that is life. Parental imitation serves several social func- close to or slightly above the child's current tions, including increasing infant attentiveness to developmental level (Bruner, 1982). This is social interaction (Piaget, 1962), promoting turn- done by creating an augmented scaffold in taking behavior (Stern, 1985), facilitating a sense which social experiences are not only geared of self as related to others (Uzgiris, 1981), enhan- toward the child's developmental level but also cing imitation of others, and shaping emotional are exaggerated and simplified so that the rele- expression and empathy (Malatesta & Haviland, vant aspects of social interaction are distilled, 1982; Meltzoff & Gopnik, 1993). Imitating an become highly salient, and are more easily comprehended. If we were to implement what we now know 4. They are designed to be sensitive to the child's how to do and to cease all efforts at the develop- narrow range of optimal stimulation. ment of new methods, it is very likely that approx- 5. They place the child in the role of initiator, imately half of all children with autism would be thus allowing the child to regulate the amount able to function in a general education program by of stimulation received and to avoid a passive the time they reached elementary school age. If role in the interaction. intervention methods were initiated at even earlier ages, which should be possible given the early Based on case studies using these methods detection methods being developed, the outcome (Klinger & Dawson, 1992), it can be concluded for children with autism by 5 years of age would that the developmental goals of young children probably be even more promising. Furthermore, with autism are similar to those of typically devel- our predictions for long-term prognosis in autism oping infants. These include, among others, the and for the need for adult services would likely be development of strategies for self-regulation of substantially revised. Perhaps, then, one of the arousal; the establishment of a trusting, loving rel- most important goals of investigators in this field ationship between the infant and his or her primary is to become more effective communicators of our caregivers that incorporates shared affect and knowledge to the general public in order to garner shared attention; the promotion of social play via their support. In the meantime, it is clear that ritualized games; and the acquisition of commun- significant progress has been made in developing icative intent. Among the programs reviewed for methods of early intervention for children with this chapter, that developed by Rogers and her autism and that our efforts at changing the course colleagues (Rogers & DiLalla, 1991; Rogers & of the lives of individuals with autism through Lewis, 1989; Rogers et al., 1986) comes closest to early intervention have been very worthwhile. sharing these early goals, especially in terms of its emphasis on the development of positive rela- REFERENCES tionships with adults and the use of positive affect Ameli, R., Courchesne, E., Lincoln, A., Kaufman, to increase the child's social attention and moti- A.S., & Grillon, C. (1988). Visual memory vation. processes in high-functioning individuals with autism. Journal of Autism and Developmental The Gap Between What Is Disabilities, 18, 601-615. Known and What Is Implemented American Psychiatric Association. (1994). Diag- As investigators continue to refine methods and nostic and statistical manual of mental develop new ones for helping young children with disorders (4th ed.). Washington, DC: Author. autism and as research studies further investigate Anderson, S.R., Avery, D.L., Dipietro, E.K., the effectiveness of early intervention in autism it Edwards. G.L.. & Christian, WP. (1987). becomes all too apparent that there is a serious gap Intensive home-based early intervention with between the state of our knowledge of early inter- autistic children. Education and Treatment of vention methods for children with autism and the Children, 10, 352-366. methods actually used in most public school Anderson, S.R., Campbell, S., & Cannon, B.O. systems. Despite careful cost–benefit analyses (1994). The May Center for early childhood (Lovaas, 1987), legislators and voting citizens education. In S. Harris & J. Handleman (Eds.). unfortunately often fail to support the level of fun- Preschool education programs for children ding needed to implement the programs described with autism (pp. 15–36). Austin, TX: PRO-ED. in this chapter. For example, it is not uncommon Boucher, J., & Warrington, E. (1976). Memory for a preschool-age child with autism to receive deficits in early infantile autism: Some similar- less than 27 hours of classroom intervention. ities to the amnesic syndrome. British Journal Moreover, the staff and training required to imp- of Psychology, 67, 73–87. lement the programs described in this chapter often Bruner, J. (1982). The organization of action and are lacking. the nature of adult–infant transaction. In E. Tronick (Ed.), Social interchange in infancy Disabilities, 5, 49–58. (pp. 23–25). Baltimore: University Park Press. Ferrara, C., & Hill, S. (1980). The responsiveness Courchesne, E., Townsend, J., Akshoomoff, N.A., of autistic children to the predictability of so- Saitoh, O., Yeung-Courchesne, R., Lincoln, A. cial and nonsocial toys. Journal of Autism and J. James, H.E., Haas, R.H.. Schreibman, L., & Developmental Disorders, 10, 51–57. Lau, L. (1994). Impairment in shifting atten- Field. T. (1977). Effect of early separation, inter- tion in autistic and cerebellar patients. Behav- active deficits and experimental manipulations ioral Neuroscience, 108, 848–865. on infant–mother face-to-face interaction. Curcio, F. (1978). Sensorimotor functioning and Child Development, 48, 731–771. communication in mute autistic children. Field, T. (1979). Visual and cardiac responses to Journal of Autism and Childhood animate and inanimate faces by young and Schizophrenia, 8, 281-292. preterm infants. Child Development, 50, 188– Dawson, G. (1991). A psychobiological perspec- 194. tive on the early socioemotional development Fischer, K., Bullock, D., Rotenberg, E., & Raya, P. of children with autism. In S. Toth & D. Cic- (1993). The dynamics of competence: How chetti (Eds.), Rochester symposium on con-text contributes directly to skill. In R.H. developmental psychopathology (Vol. 3, pp. Wozniak & K. Fischer (Eds.), Development in 208–234). Rochester, NY: University of context: Acting and thinking in specific Rochester Press. environments (pp. 93–117). Hillsdale, NJ: Dawson, G., & Adams, A. (1984). Imitation and Lawrence Erlbaum Associates. social responsiveness in autistic children. Jour- Guralnick, M J. (1993). Second generation re- nal of Abnormal Child Psychology, 12, 209– search on the effectiveness of early interven- 226. tion. Early Education and Development, 4, Dawson, G., & Galpert, L. (1990). Mothers' use of 366-378. imitative play for facilitating social responsive- Handleman, J., & Harris, S. (1994). The Douglass ness and toy play in young autistic children. Developmental Disabilities Center. In S. Harris Development and Psychopathology, 2, 151– & J. Handleman (Eds.), Preschool education 162. pro-grams for children with autism (pp. 71– Dawson, G., & Lewy, A. (1989). Arousal, 86). Austin, TX: PRO-ED. attention, and the socioemotional impairments Harris, S., Handleman, J.S., Kristoff, B., Bass, L., of individuals with autism. In G. Dawson & Gordon, R. (1990). Changes in language de- (Ed.), Autism: Nature, diagnosis, and velopment among autistic and peer children in treatment (pp. 49–74). New York: Guilford segregated and integrated preschool settings. Press. Journal of Autism and Developmental Dawson, G., Meltzoff, A., & Osterling, J. (1995). Disorders, 20, 23–31. Children with autism fail to orient to naturally Hart, B.M., & Risley, T.R. (1975). Incidental occurring social stimuli. Poster presented at teaching of language in the preschool. Journal the 1995 Meeting of the Society for Research of Applied Behavior Analysis, 8, 411–420. in Child Development, Indianapolis, IN. Hermelin, B., & O'Connor, N. (1970). Donnellan, A.M., Mirenda, P.L., Mesaros, R.A., & Psychological experiments with autistic Fassbender, L.L. (1984). Analyzing the children. Oxford, England: Pergamon Press. communicative functions of aberrant behavior. Hoyson, M., Jamieson, B., & Strain, P. (1984, Journal of The Association for Persons with Summer). Individualized group instruction of Severe Handicaps, 9, 201–212. normally developing and autistic-like children: Fenske, E.C., Zalenski, S., Krantz, PJ., & The LEAP curriculum model. Journal of the McClannahan, L.E. (1985). Age at intervention Division of Early Childhood, 157–171. and treatment outcome for autistic children in a Jarrod, C., Boucher, J., & Smith, P. (1993). Sym- comprehensive intervention program. Analysis bolic play in autism: A review. Journal of and Intervention in Developmental Autism and Developmental Disorders, 23, 281- 307. programs for children with autism (pp. 87— Kaiser, A.P., Yoder, P.J., & Keetz, A. (1992). 106). Austin, TX: PRO-ED. Evaluating milieu teaching. In S.F. Warren & Lord, C., Storoschuk, S., Rutter, M., & Pickles, A. J. Reichle (Eds.), Communication and (1993). Using the ADI—R to diagnose autism language intervention series: Vol. 1. Causes in preschool children. Infant Mental Health and effects in communication and language Journal, 14, 234—252. intervention (pp. 9–48). Baltimore: Paul H. Lovaas, O.I. (1987). Behavioral treatment and nor- Brookes Publishing Co. mal educational and intellectual functioning in Kanner, L. (1943). Autistic disturbance of affec- young autistic children. Journal of Consulting tive contact. Nervous Child, 2, 217–250. and Clinical Psychology, 55(1), 3—9. Klinger, L.G., & Dawson, G. (1992). Facilitating Lovaas, O.I., Smith, T., & McEachin, J.J. (1989). early social and communicative development Clarifying comments on the young autism in children with autism. In S.F. Warren & J. study: Reply to Schopler. Short and Mesibov. Reichle (Eds.), Communication and language Journal of Consulting and Clinical intervention series: Vol. 1. Causes and effects Psychology, 57, 163—167. in communication and language intervention Mahoney, G., & Powell, A. (1988). Modifying (pp. 157–186). Baltimore: Paul H. Brookes parent—child interaction: Enhancing the Publishing Co. development of handicapped children. Journal Klinger, L., & Dawson, G. (1995). A fresh look at of Special Education, 22, 82—96. categorization abilities in persons with autism. Malatesta, C., & Haviland, J. (1982). Learning dis- In E. Schopler & G.B. Mesibov (Eds.), play rules: The socialization of emotion Learning and cognition in autism (pp. 119– expression in infancy. Child Development, 53, 135). New York: Plenum. 991—1003. Koegel, R., & Johnson, J. (1989). Motivating lan- McClannahan, L., & Krantz, P. (1994). The guage use in autistic children. In G. Dawson Prince-ton Child Development Institute. In S. (Ed.),. Autism: Nature, diagnosis, and Harris & J. Handleman (Eds.), Preschool treatment (pp. 310-325). New York: Guilford education programs for children with autism Press. (pp. 107—126). Austin, TX: PRO-ED. Koegel, R., & Koegel, L.K. (1988). Generalized McEachin, J.J., Smith Tristram, T., & Lovaas, O.I. responsivity and pivotal behavior. In R.H. (1993). Long-term outcome for children with Horner, G. Dunlap, & Koegel (Eds.), Gen- autism who received early intensive behavioral eralization and maintenance: Lifestyle changes treatment. American Journal on Mental in applied settings (pp. 41–66). Baltimore: Retardation, 97, 359—372. Paul H. Brookes Publishing Co. McGee, G., Daly, T., & Jacobs, H.A. (1994). The Koegel. R., & Mentis, M. (1985). Motivation in Walden Preschool. In S. Harris & J. childhood autism: Can't they or won't they? Handleman (Eds.), Preschool education journal of Child Psychology and Psychiatry, programs for children with autism (pp. 127— 26, 185—191. 162). Austin, TX: PRO-ED. Koegel, R., & Schreibman, L. (1977). Teaching Meltzoff, A.N., & Gopnik, A. (1993). The role of autistic children to respond to simultaneous imitation in understanding persons and devel- multiple cues. Journal of Experimental Child oping theories of mind. In S. Baron-Cohen, H. Psychology, 6, 147-156. Tager-Flusberg, & D. Cohen (Eds.), Lewis, V, & Boucher, J. (1988). Spontaneous, in- Understanding other minds: Perspectives from structed and elicited play in relatively able autism (pp. 335—366). Oxford, England: autistic children. British Journal of Oxford University Press. Developmental Psychology, 6, 325-339. Minshew, NJ. & Goldstein, G. (1993). Is autism an Lord, C., & Schopler, E. (1994). TEACCH ser- amnesic disorder? Evidence from the vices for preschool children. In S. Harris & J. California Verbal Learning Test, Handleman (Eds.), Preschool education Neuropsychology, 209-216. Mundy, P., Sigman, M., Ungerer, J., & Sherman. Schopler, E., Reichler, R.J., DeVellis, R.F., & T. (1986). Defining the social deficits of Daly, K. (1980). Toward objective autism: The contribution of non-verbal classification of childhood autism: Childhood communication measures. Journal of Child Autism Rating Scale (CARS). Journal of Psychology and Psychiatry, 27, 657-669. Autism and Developmental Disorders, 10, Osterling, J., & Dawson, G. (1994). Early, 91—103. recognition of children with autism: A study of Siegel, B., Pliner, C., Eschler, J., & Elliott, G.R. first birthday home video tapes. Journal of (1988). How children with autism are diag- Autism and Developmental Disorders, 24, nosed: Difficulties in identification of children 247—257. with multiple developmental delays. Piaget, J. (1962). Play, dreams, and imitation. Developmental and Behavioral Pediatrics, 9, New York: W.W. Norton. 199-204. Powers, M. (1992). Early intervention for children Stern. D.N. (1985). The interpersonal world of the with autism. In D. Berkell (Ed.), Autism: infant: A view from psychoanalysis and Identification, education, and treatment (pp. developmental psychology. New York: Basic 223—251). Hillsdale, NJ: Lawrence Erlbaum Books. Associates. Strain, P.S.. & Cordisco, L.K. (1994). LEAP Prizant, B., & Wetherby, A. (1989). Enhancing Preschool. In S. Harris & J. Handleman (Eds.), language and communication in autism: From Preschool education programs for children theory to practice. In G. Dawson (Ed.), Autism: with autism (pp. 225-252). Austin, TX: PRO- Nature, diagnosis, and treatment (pp. 282— ED. 309). New York: Guilford Press. Tager-Flusberg, H. (1989). A psycholinguistic per- Riguet, C.B., Taylor, N.D., Benarova, S., & Klein, spective on language development in the autis- L.S. (1981). Symbolic play in autistic, Downs, tic child. In G. Dawson (Ed.), Autism: Nature, and normal children of equivalent mental age. diagnosis, and treatment (pp. 92-115). New Journal of Autism and Developmental York: Guilford Press. Disabilities, 11, 439—448. Tager-Flusberg, H. (1993). What language reveals Rogers, S J., & DiLalla, D.L. (1991). A about the understanding of mind in children comparative study of the effects of a with autism. In S. Baron-Cohen; H. Tager- developmentally based instructional model on Flusberg, & D. Cohen (Eds.), Understanding young children with autism and young children other minds: Perspectives from autism (pp. with other disorders of behavior and 138-157). Oxford, England: Oxford University development. Topics in Early Childhood Press. Special Education, 11(2), 29—47. Terman. L., & Merrill, M. (1960). Stanford-Binet Rogers, S J, Herbison, J., Lewis, H., Patone, J., & Intelligence Scale. Boston: Houghton Mifflin. Reis, K. (1986). An approach for enhancing the Tiegerman, E., & Primavera, L. (1981). Object symbolic, communicative, and interpersonal manipulation: An interactional strategy with functioning of young children with autism and autistic children. Journal of Autism and severe emotional handicaps. Journal of the Develop-mental Disorders, 11, 427-438. Division for Early Childhood, 10, 135—148. Tiegerman, E., & Primavera, L. (1984). Imitating Rogers, S J., & Lewis, H. (1989). An effective day the autistic child: Facilitating communicative treatment model for young children with per- gaze behavior. Journal of Autism and vasive developmental disorders. Journal of the Developmental Disorders, 14, 27-38. American Academy of Child and Adolescent Ungerer, J., & Sigman, M. (1981). Symbolic play Psychiatry, 28, 207—214. and language comprehension in autistic chil- Schopler, E., & Reichler, R J. (1971). Parents as dren. Journal of the American Academy of cotherapists in the treatment of psychotic chil- Child and Adolescent Psychiatry, 20, 318-337. dren. Journal of Autism and Childhood Uzgiris, I. (1981). Experience in the social con- Schizophrenia, 1, 87—102. text: Imitation and play. In R. Schiefelbusch & D. Bricker (Eds.), Early language: Acquisition and intervention (pp. 139-168). Baltimore: University Park Press.