Development of Communication
Major advances have been made over the past two decades in delineating and understanding the communication and language difficulties of children with autism. The characterization of communication deficits in the diagnostic criteria for autism has changed dramatically. Until about 1980, peculiar speech patterns were emphasized, such as echolalia, pronoun reversal, and unusual intonation (Baltaxe and Simmons, 1975; American Psychiatric Association, 1980). Now, verbal and nonverbal communication are considered a core deficit in the diagnostic criteria for autistic spectrum disorders (American Psychiatric Association, 1987; 1994). This change highlights the recognition that children with autistic spectrum disorders not only have difficulty in the acquisition of speech and language, but also have difficulty understanding and using nonverbal behavior in communicative interactions.
The level of communicative competence attained by individuals with autism has been found to be an important predictor of outcome (Garfin and Lord, 1986; McEachin et al., 1993). The presence of fluent speech (using multiword combinations spontaneously, communicatively, and regularly) before the age of 5 continues to be a good prognostic indicator of IQ scores, language measures, adaptive skills, and academic achievement in adolescence (Venter et al., 1992). Moreover, the severity of the communicative impairment may be one of the greatest sources of stress for families (Bristol, 1984).
There is much heterogeneity in the speech, language and communication characteristics of children with autistic spectrum disorders. Language impairments in autistic spectrum disorders range from failure to
develop any functional speech to the development of functional but idiosyncratic use of spontaneous speech and language (Lord and Paul, 1997). One-third (Bryson, 1996) to one-half (Lord and Paul, 1997) of children and adults with autism do not use speech functionally. For both verbal and nonverbal individuals, impairments in social or pragmatic aspects of language and related cognitive skills are the most salient (Wetherby et al., 1997).
CORE COMMUNICATION DEFICITS
Research over the past decade has identified core communication deficits in children with autism that fall into two major areas: joint attention and symbol use (Dawson et al., 1990; Kasari et al., 1990; McArthur and Adamson, 1996; Mundy et al., 1990; Sigman and Ruskin, 1999; Stone et al., 1997; Wetherby et al., 1998). Joint attention reflects difficulty coordinating attention between people and objects and is evident by deficits in orienting and attending to a social partner; shifting gaze between people and objects; sharing affect or emotional states with another person; following the gaze and point of another person; and being able to draw another persons’ attention to objects or events for the purpose of sharing experiences.
Symbol use reflects difficulty learning conventional or shared meanings for symbols and is evident in deficits in using conventional gestures; learning conventional meanings for words; and using objects functionally and in symbolic play.
Joint attention has been found to be a significant predictor of language outcome. Mundy et al. (1990) found that measures of gestural joint attention (e.g., showing or pointing to direct attention) at initial testing were a significant predictor of language development 1 year later for preschool children with autism. The failure to acquire gestural joint attention appears to be a critical milestone that impairs language development and an important target for early communication intervention.
Similarly, children with autism do not compensate for their lack of verbal skills with gestures; they show limited gestural use, both in quantity and quality. They predominantly use primitive motoric gestures to communicate (i.e., leading, pulling or manipulating another’s hand). They lack the use of many conventional gestures, such as showing, waving, pointing, nodding the head and symbolic gestures depicting actions (Loveland and Landry, 1986; McHale et al., 1980; Stone and Caro-Martinez, 1990; Stone et al., 1997; Wetherby et al., 1998; Wetherby et al., 1989).
Moreover, in this population, there is much variability in the capacity to use vocal communication which likely contributes to the wide range of verbal skills. Some children with autism have been found to use a limited
consonant inventory and less complex syllabic structure, while others show adequate complexity of vocalizations (McHale et al., 1980; Stone and Caro-Martinez, 1990; Wetherby and Prutting, 1984; Wetherby et al., 1989).
The vast majority of those who do learn to talk go through a period of using echolalia, the imitation of speech of others, which may be immediate or delayed (Prizant et al., 1997). An echolalic utterance is usually equivalent to a single word or a label for a situation or event. Many children learn to use echolalia purposefully in communicative interactions, and eventually are able to break down the echolalic chunks into smaller meaningful units as part of the process of transitioning to a rule-governed, generative language system (Prizant and Rydell, 1993).
Children with autism who progress beyond echolalia usually acquire more advanced aspects of grammar: that is, they develop grammatical skills in the same general progression as typically developing children, but show persisting problems in following the social rules and shifting between speaker and listener roles of conversation (Baltaxe, 1977; Tager-Flusberg, 1996), which are the pragmatic aspects of language.
In lieu of conventional means of communicating, children with autism may develop idiosyncratic, unconventional, or inappropriate behaviors to communicate, such as self-injurious behavior, aggression, or tantrums. Despite the fact that at least 50 percent of individuals with autism display some functional speech and language skills (Lord and Paul, 1997), challenging behaviors such as aggression, tantrums, and self-injury are often used to procure attention, to escape from a task or situation, to protest against changes of schedule and routine, or to regulate interactions in a predictable manner. Carr and Durand (1985) reported that aggression, tantrums, and self-injury were more likely to occur in situations with a high level of task difficulty and a low level of adult attention. Challenging behaviors need to be considered relative to the child’s repertoire of verbal and nonverbal communicative behaviors and may reflect limitations in symbolic capacity.
Further evidence of a deficit in the symbolic capacity in autism is the limited ability to develop symbolic or pretend play. Although play is a social-cognitive skill, a lack of varied, spontaneous make-believe play is one of the four possible features of the impairment in communication in the most recent Diagnostic and Statistical Manual (DSM-IV) (American Psychiatric Association, 1994). Children with autism show significant deficits in symbolic or make-believe play (i.e., using pretend actions with objects) and limited abilities in functional play (i.e., using objects functionally) (Dawson and Adams, 1984; Sigman and Ungerer, 1984; Wetherby and Prutting, 1984; Wing et al., 1977). Functional and symbolic play skills have been found to be significantly correlated with receptive and expressive language (Mundy et al., 1987; Sigman and Ruskin, 1999). In contrast
to deficits in functional object use and symbolic play, children with autism often perform at similar or sometimes even higher levels on nonsocial constructive play (e.g., using objects in combination to create a product, such as putting puzzles together) in comparison with typically developing children or children with language delays at the same language stage (Wetherby and Prutting, 1984; Wetherby et al., 1998).
Exploring developmental patterns in communication and symbolic abilities has contributed to better understanding of the nature of these problems in autism. Stone et al. (1997) and Wetherby et al. (1998) compared the developmental profiles of 2- to 4-year-old children with autistic spectrum disorders with that of children with delayed language who were at the same language stage. Using similar strategies for gathering communication samples, both researchers reported a similar profile in children with autistic spectrum disorders, characterized by a distinct constellation of strengths and weaknesses in parameters of communication. Specifically, the children with autistic spectrum disorders showed comparable use of communication to request and protest, but significantly less use of gaze shifts, shared positive affect, conventional gestures, and communication for joint attention. They performed at comparable levels of constructive play but significantly poorer levels of language comprehension and symbolic play. Correlational findings from the Wetherby et al. (1998) study showed that children who displayed a greater capacity to coordinate attention and affect were more likely to communicate for more social reasons, to use a larger repertoire of conventional gestures, to have a higher rate of communicating, and to employ better repair strategies. These findings underscore the importance of addressing these core deficits in interventions for children with autism and have important implications for predicting which children will benefit from specific intervention approaches.
PLANNING FOR INTERVENTION
Researchers and educators have debated the question of how communication goals and objectives for children with autism and related disabilities should be derived. The perspective espoused by traditional behavioral programs has been to establish goals and objectives a priori (e.g., Lovaas, 1981). Behavioral discrete-trial programs begin with general compliance training to get a child to sit in a chair, look at the clinician, and imitate nonverbal behavior in response to verbal commands. Speech is taught as a verbal behavior, and objectives are targeted beginning with verbal imitation, following one-step commands, receptive discrimination of body parts, objects, person names and pictures, and expressive labeling
in response to questions. Later, language objectives include prepositions, pronouns, same/different and yes/no.
More contemporary behavioral approaches have developed goals for outcomes from a functional assessment. Goals and objectives are individualized, based on a child’s repertoire of communicative behaviors, teaching functional equivalents of challenging behavior, and addressing the child’s individual needs. The functional emphasis focuses on goals that affect a child’s access to choices of activities in which to participate, opportunities for social interaction, and community settings (Brown et al., 1979; Horner et al., 1990). Contemporary behavioral programs emphasize teaching communication skills so that greater access is provided to a variety of people, places and events, thereby enhancing the quality of life of children with autistic spectrum disorders.
The perspective espoused by developmentally oriented approaches has been to focus on the communicative meaning of behaviors and to target goals and objectives that enhance a child’s communicative competence by moving the child along a developmental progression (Lahey, 1988). Contemporary developmentalists begin with social-communicative goals, including gaze to regulate interaction, sharing positive affect, communicative functions, and gestural communication. Language goals are mapped onto social communication skills and are guided by a developmental framework (Greenspan and Wieder, 1997; Klinger and Dawson, 1992; Wetherby et al., 1997).
Goal-setting in an augmentative and alternative communication (AAC) intervention is usually guided by a developmental perspective. Beukelman and Mirenda (1998) state that the goals of an AAC intervention are to assist individuals with severe communication disorders to become communicatively competent in the present, with the view toward meeting their future communication needs.
One major purpose of communication assessment is to document change as an outcome measure of treatment. However, most formal or standardized language assessment measures focus primarily on language form and rely on elicited responses. Because language impairments associated with autism are most apparent in social-communicative or pragmatic aspects of language, formal assessment instruments can provide information about only a limited number of aspects of communication for children with autism (Schuler et al., 1997; Prizant et al., 1997; Wetherby and Prizant, 1999). Formal language measures are especially imprecise in measuring nonverbal aspects of communication and therefore are not sufficient, particularly for low-functioning children with autism. In many
situations, the tests used for pre- and post-assessment are different, due to the child’s increasing age, making interpretation of results difficult.
Another major purpose of assessment is to provide information for educational planning that can be directly translated into goals, strategies, and outcome measures for communication enhancement. Several communication abilities have been identified as important to assess for children with autism: use of eye gaze and facial expression for social referencing and to regulate interaction, range of communicative functions expressed, rate of communicating, use of gestures and vocal/verbalizations, use of repair strategies, understanding of conventional meanings, and ability to engage in conversation (Schuler et al., 1997; Wetherby et al., 2000). Wetherby et al. (1997) point out that communicative abilities of children with autism should be documented in natural communicative exchanges, with a child’s symbolic abilities serving as a developmental frame of reference. To supplement formal measures, the systematic use of informal procedures to assess language and communication is needed. In order to gather an accurate picture of the communication and symbolic abilities of children with autism, a combination of assessment strategies has been recommended, including interviewing significant others (i.e., parents, teachers) and observing in everyday situations to find out how a child communicates in the home, classroom, and other daily settings (Wetherby and Prizant, 1999).
Although there is consensus on the importance of enhancing communication abilities for children with autism, intervention approaches vary greatly, and some even appear to be diametrically opposed. The methodological rigor in communication intervention studies in terms of internal and external validity and measures of generalization has been stronger than in many other areas of autism intervention studies. Nevertheless, there have been relatively few prospective studies with controls for maturation, expectancy, or experimenter artifacts. The strongest studies in terms of internal validity have been multiple baseline, ABAB, or similar designs that have included controls for blindness of evaluations (see Figure 1–1 in Chapter 1). There have been almost no studies with random assignment, although about 70 percent of the studies included well-defined cohorts of adequate sample size or replication across three or more subjects in single subject designs (see Figure 1–2 in Chapter 1). A substantial proportion of communication interventions have also included some assessment of generalization, though most often not in a natural setting (see Figure 1–3 in Chapter 1).
In order to examine the critical elements of treatment programs that affect the speech, language, and communication skills of children with
autism, it is useful to characterize the active ingredients of treatment approaches along a continuum—from traditional, discrete trial approaches to more contemporary behavioral approaches that used naturalistic language teaching techniques to developmentally oriented approaches (Prizant and Wetherby, 1998; Anderson and Romanczyk, 1999; Prizant and Rubin, 1999). The earliest research efforts at teaching speech and language to children with autism used massed discrete trial methods to teach verbal behavior by building labeling vocabulary and simple sentences. Lovaas (1977, 1981) provided the most detailed account of the procedures for language training using discrete trial approaches. Outcomes of discrete trial approaches have included improvements in IQ scores, which are correlated with language skills, and improvements in communication domains of broader measures, such as the Vineland Adaptive Behavior Scales (McEachin et al., 1993). A limitation of a discrete trial approach in language acquisition is the lack of spontaneity and generalization. Lovaas (1977) stated that “the training regime…its use of ‘unnatural’ reinforcers, and the like may have been responsible for producing the very situation-specific, restricted verbal output which we observed in many of our children” (p. 170). In a review of research on discrete trial approaches, Koegel (1995) noted that “not only did language fail to be exhibited or generalize to other environments, but most behaviors taught in this highly controlled environment also failed to generalize” (p. 23).
There is now a large body of empirical support for more contemporary behavioral approaches using naturalistic teaching methods that demonstrate efficacy for teaching not only speech and language, but also communication. These approaches include natural language paradigms (Koegel et al., 1987), incidental teaching (Hart, 1985; McGee et al., 1985; McGee et al., 1999), time delay and milieu intervention (Charlop et al., 1985; Charlop and Trasowech, 1991; Hwang and Hughes, 2000; Kaiser, 1993; Kaiser et al., 1992), and pivotal response training (Koegel, 1995; Koegel et al., 1998). These approaches use systematic teaching trials that have several common active ingredients: they are initiated by the child and focus on the child’s interest; they are interspersed and embedded in the natural environment; and they use natural reinforcers that follow what the child is trying to communicate. Only a few studies, all using single-subject designs, have compared traditional discrete trial with naturalistic behavioral approaches. These studies have reported that naturalistic approaches are more effective at leading to generalization of language gains to natural contexts (Koegel et al., 1998; Koegel et al., 1992; McGee et al., 1985).
There are numerous intervention approaches based on a developmental framework (e.g., Greenspan and Wieder, 1997; Klinger and Dawson, 1992; Wetherby et al., 1997; Prizant and Wetherby, 1998). While
there are many different developmental programs, a common feature of developmental approaches is that they are child-directed. The environment is arranged to provide opportunities for communication, the child initiates the interaction or teaching episode, and the teacher or communicative partner follows the child’s lead by being responsive to the child’s communicative intentions, and imitating or expanding the child’s behavior. Although the empirical support for developmental approaches is more limited than for behavioral approaches, there are several treatment studies that provide empirical support for language outcomes using specific strategies built on a developmental approach (Lewy and Dawson, 1992; Hwang and Hughes, 2000; Rogers and DiLalla, 1991; Rogers and Lewis, 1989) and many case studies, with Greenspan and Wieder (1997) providing the largest case review. Developmental approaches share many common active ingredients with contemporary naturalistic behavioral approaches and are compatible along most dimensions (Prizant and Wetherby, 1998).
Teaching Speech and Language
Gains in speech and language outcomes for children with autism have been documented using a variety of behavioral and developmental intervention approaches. Numerous studies have investigated methods of teaching specific receptive and expressive language skills. Most of these studies have used a behavioral method ranging from discrete-trial to naturalistic. Studies have reported good outcomes for teaching specific language content, such as single-word vocabulary, describing objects and pictures, responding to questions, and increasing speech intelligibility (see Goldstein, 1999; Koegel et al., 1998; Krantz et al., 1981). Very positive outcomes have been reported by McGee and colleagues (1999) through natural reinforcers of vocalization, speech shaping, and incidental teaching. They reported that 36 percent of the toddlers studied used verbalizations at program entry with a mean age of 2 years 5 months, and 82 percent were verbalizing meaningful words 1 year later. Most other programs have reported about children entering at 3 years of age or later, and therefore, the impressive treatment outcomes may be related to the young age at entry of treatment.
Research that has documented changes in the communication skills of children with autism falls into three major categories organized by the goal of the intervention: functional communication training to replace challenging behavior, increases in initiation of verbal and nonverbal communication, and increases in the core communication skills.
There is strong empirical support for the efficacy of functional communication training to replace challenging behaviors. This approach includes a functional assessment of the particular behavior to determine its function for a child (e.g., desire for tangible or sensory item, attention, or to escape a situation or demand) and teaching communication skills that serve efficiently and effectively as functional equivalents to challenging behaviors, a method that has been documented to be the most effective for reductions in challenging behavior (Horner et al., 1990; see Horner et al., 2000).
There are also some findings concerning the use of augmentative communication strategies. In a literature review of the functional communication training research, Mirenda (1997) found that eight children with autism, three of whom were under 8 years of age, were able to learn to use AAC to replace challenging behaviors. Their problem behaviors included self-injurious behavior, aggression, crying, screaming, property destruction, tantrums, non-compliance, and self-stimulatory behaviors. These children were systematically taught to use AAC with messages congruent with the function of the behavior, such as “Look at me” (attention); “I want__” (tangible); “I need a break” (escape). This intervention resulted in a substantial and immediate decrease in the problem behaviors, and the use of AAC for functional communication training was maintained over the course of a year. Naturalistic behavioral language interventions leading to improved communicative skills have also been associated with reductions in disruptive behavior (Koegel et al., 1992) and provide further evidence supporting the relationship between communication and behavior.
There is a growing body of research on increasing the initiation of communication in children with autism. Initiation of communication has been described as a pivotal behavior: the more often a child initiates communication, the more often it will trigger responses from others, which will in turn enhance and expedite the improvement of other communication and language skills (Koegel, 1995). Two important findings have been reported (Koegel et al., 1999). First, children who show more spontaneous, self-initiated communication at the beginning of treatment show more favorable language treatment outcomes. Second, in specific contexts, self-initiated communication can be taught to children with autism who show few or no spontaneous communication and has been associated with favorable treatment outcomes (Charlop et al., 1985; Charlop and Trasowech, 1991). In general, truly spontaneous, self-initiated, socially directed behaviors are much more difficult, though not impossible, to teach (Watson et al., 1989) and require a combination of developmental and naturalistic teaching methods.
In spite of the large number of studies documenting the core communication deficits associated with autism (i.e., joint attention and symbolic
capacity), there are only a few studies that have documented intervention effects on these core deficits. Most of the comprehensive programs do not present data targeting improvement in these skills. Exceptions are Rogers and Lewis (1989), who documented improvements in symbolic play as a result of a structured, developmentally based program (see Chapter 6) and studies of symbolic play with pivotal response treatment (Thorp and Schreibman, 1995; Stahmer, 1995). Other studies that have documented improvement in these core communication deficits have demonstrated increases in gaze to regulate interaction, shared positive affect, use of conventional gestures, and joint attention. Lewy and Dawson (1992) compared the effects of a child-directed teaching strategy in which the adult imitated the child’s behavior with an adult-directed teaching strategy in a group comparison study. They demonstrated that the imitation strategy improved gaze, turn-taking, object use, and joint attention in children with autism, while the adult-directed strategy did not lead to these communicative gains. More recent studies have used single-subject designs to provide systematic evidence of naturalistic language teaching techniques that improve joint attention skills in children with autism (Buffington et al., 1998; Hwang and Hughes, 2000; Pierce and Schriebman, 1995). Thus, naturalistic behavioral or structured developmental methods appear to be an effective way to address the core communication deficits of autism.
Augmentative and Alternative Communication and Assistive Technology
For children with autism who do not acquire functional speech or have difficulty processing and comprehending spoken language, augmentative and alternative communication (AAC) and assistive technology (AT) can be useful components of an educational program. There is disagreement about whether to use AAC to train speech and language for young children with autistic spectrum disorders. There is relatively little rigorous, systematic research to elucidate characteristics of children and the components of AAC and AT that may interact to produce effective (or ineffective) intervention. However, available findings are summarized in some detail here to provide a snapshot of this emerging area.
AAC is defined as “an area of clinical practice that attempts to compensate (either temporarily or permanently) for the impairment and disability patterns of individuals with severe expressive communication disorders” (American Speech-Language-Hearing Association, 1989:7). AAC may involve supporting existing speech or developing independent use of a nonspeech symbol system, such as sign language, visual symbols (pictures and words) displayed on communication boards, and voice output devices with synthesized and digitized speech. AT is any commer-
cial, hand-made, or customized device or service used to support or enhance the functional capabilities of individuals with disabilities. AT includes computer-assisted instruction, mobility devices, high and low technology adaptations and AAC.
The methods and tools of AAC interventions, properly applied, are tailored to unique strengths and needs of individuals with autism. AAC includes the use of visual language systems, such as visual icons or words representing specific communicative units, which capitalize on strong visual processing of many children with autism. The visual information is static and predictable, and enables the child with autism to rely on recognition rather than recall memory to receive language input or generate language output. AAC provides a motorically simple way to communicate needs and may preempt the development of challenging coping behaviors. Low-technology AAC tools, such as picture systems, can be relatively simple and inexpensive to implement (Hodgdon, 1995).
Relationship Between AAC and Speech and Language Development
There is empirical evidence that systematic teaching of speech using a naturalistic behavioral approach is efficacious for many children, particularly if treatment can begin by 2-1/2 years of age (McGee et al., 1999). However, a substantial proportion of children fail to make meaningful gains in speech (with failure rates ranging from about 20 percent to 40 percent). For those children who do acquire speech, the degree of spontaneity and complexity of language is not clearly reported in most research studies. There is now a body of research on AAC and speech and language acquisition in children with autism that is important to consider, particularly for those children who make slow or minimal gains in other programs.
There is a dearth of research on communication assessment strategies for children with autism using AAC. In one case study of a child with autism (Light et al., 1998), the AAC assessment principles of the Communication Participation Model (Beukelman and Mirenda, 1998) were implemented to gather information needed for an effective AAC intervention. Based on this participation plan and a variety of informal assessments, a comprehensive multimodal AAC intervention (speech, pointing, a communication book, a laptop computer with synthesized speech) was implemented, increasing the level of communication and participation for this child.
There have been numerous experimental studies of the efficacy of teaching sign language to children with autism (see Goldstein, 1999). These studies have demonstrated that total communication (speech plus sign language) training resulted in faster and more complete receptive
and expressive vocabulary acquisition than speech training alone for many children with autism (Barrera et al., 1980; Barrera and Sulzer-Azaroff, 1983; Carr and Dores, 1981; Layton, 1988; McIlvane et al., 1984; Yoder and Layton, 1988). These findings support several conclusions for children with autism:
There is no evidence that use of AAC systems as collaterals to language instruction results in delays in the acquisition of speech, though specifying the advantages and disadvantages of using AAC in support of the development of speech in different populations remains a research question.
There is evidence that sign language enhances the use of speech for some children.
There is no evidence to suggest that sign language interferes with the development of speech.
Children with good verbal imitation skills demonstrate better speech production than those with poor verbal imitation skills, with or without AAC.
Children with poor verbal imitation skills are the best candidates for an AAC system, such as sign language, because they are likely to make poor progress in speech acquisition without AAC.
Seal and Bonvillian (1997) analyzed sign language formation of 14 low-functioning students with autism and found that the size of the sign vocabulary and accuracy of sign formation were highly correlated with measures of fine motor abilities and tests of apraxia, which is a neurogenic impairment of planning, executing, and sequencing movements (LaPointe and Katz, 1998). These findings support the role of a motor impairment in the level of competence attained in sign language and speech acquisition for children with autism, in addition to their social-communication and symbolic deficits. It is important to note that simple signs may be a support for children learning to speak or an additional mode of communication for children who have no speech or limited speech. However, it is very rare to find a child with autism who learns to sign fluently (in sentences) and flexibly. Signing is not generally an entry point into a complex, flexible system.
The use of visual symbol systems has received attention recently because of the limited outcomes with signs and the visual strengths of many children with autistic spectrum disorders. Picture Communication Symbols are the most commonly used line drawings for augmenting spoken language. Other visual symbols used include tangible or real objects, photographs, rebus symbols and several commercial symbol-to-word computer programs, such as Picture-It, Pix-writer, and Writing with Symbols 2000. Visual symbols have been used successfully with children with
autism to increase compliance, enhance communicative initiations and responses, and decrease verbal prompt dependence (Mirenda and Santogrossi, 1985; Steibel, 1999). Communication partner training in using visual symbols, with parents, practitioners, and peers has been shown to be relatively simple (Steibel, 1999; Garrison-Harrell et al., 1997; Cafiero, 1995).
Because children with autism have difficulty pointing and show strengths in using contact gestures, they may benefit from using a giving gesture to make choices or indicate a selection from an array of objects or visual symbols. The most widely used exchange system, the Picture Exchange Communication System (PECS) (Frost and Bondy, 1994), is a structured program that teaches the exchange of symbols for communication. PECS is a systematic behavioral program that teaches a child to initiate communicative requests by approaching the communication partner and exchanging the symbol for the desired object. It includes protocols for expanding communication from single to multiple words and for increasing communicative function from requesting to labeling and commenting. Bondy and Frost (1994) reported a case review of a group of preschoolers with autism who were taught PECS. Of 19 children who used PECS for less than 1 year, only two (10%) acquired independent speech, while five used speech with PECS, and 12 children used PECS as their sole communication. Of 66 children using PECS for 2 years, 39 (59%) developed independent speech, 20 developed speech as they used PECS, and 7 used only PECS. Thus, for most preschoolers introduced to PECS, it took more than 1 year after initiating PECS to observe independent speech, and many continued to have very limited spontaneous use of language. Speech tended to develop once the children were able to use 30–100 symbols to communicate (Frost and Bondy, 1994). Furthermore, the overall communication development of the children was strongly related to their overall level of intellectual functioning.
The only other published study using PECS was reported by Schwartz et al., (1998) on 11 children with autistic spectrum disorders who attended an integrated preschool. These children required an average of 11 months to exchange “I want+symbol” sentence strips with adults and 14 months with peers. In this study, 6 (55%) of the 11 children developed functional and complex speech, and the 5 who did not were able to use PECS effectively to communicate. The authors state, however, that their study did not control for maturation or the effects of other components of their school program. Whether comparable outcomes with PECS and the concomitant development of speech would be expected without the specific intervention or with older children is not known.
There is even less research on the effectiveness of other AAC systems used by children with autism. A voice output communication aid (VOCA) is a portable AAC device that produces synthesized or digitized (re-
corded) speech. Particular messages can be accessed through visualgraphic symbols, words, or letters on the VOCA display. A VOCA can range from a single switch device that delivers a limited number of voice messages to a more complex VOCA that delivers a series of communicative units or messages, often related to a specific theme or activity, and has the capability for thousands of messages. Dynamic display devices are VOCAs in which a child points to a particular generic symbol, such as “lunch”, and a new “lunch specific” board instantly opens up with the vocabulary needed for the child to make requests, interact, comment, and question within the lunch context. Some highly sophisticated VOCAs have large capacities for storing complex spoken and written text and can operate with personal computers.
A possible advantage of a VOCA over a low-technology symbol board is the ability to facilitate more normalized, natural interactions and provide verbal models for speech development due to the voice output. Four preschool children with autism with little or no functional speech were taught to use VOCAs with line drawing displays to make requests. Using a naturalistic behavioral teaching method, all four of the children successfully learned to use their VOCAs to request, make social comments, and respond to questions in a contextually appropriate and spontaneous manner in 1 to 3 months (Schepis et al., 1998). These findings are preliminary but suggest the potential value of VOCAs to support communicative interactions of children with autism.
In addition to using visual symbols for communicative output, AAC interventions have also used visual symbols to augment communicative input from others. Recently, this equally important “input” aspect of AAC is being recognized, and several case studies demonstrate the effectiveness of augmented input for young children with autism (Hodgdon, 1995; Quill, 1997; Peterson et al., 1995). One of the most widely used AAC input techniques is the use of visual schedules. The visual schedule enables a child with autism to understand the sequence of an activity through the visual input. The TEACCH (Treatment and Education of Autistic and Related Communication Handicapped Children) program, developed and implemented in North Carolina for almost 30 years, has utilized visual schedules and protocols to promote independence, self-management, and task completion (Schopler et al., 1983; Marcus et al., 2000; see Chapter 12). A few studies have demonstrated independent task engagement and completion through pictorial representation of the task components, called “within-task” schedules (Hall et al., 1995; Mirenda et al., in press; Pierce and Schreibman, 1994). Schedules that provide predictability as students transition from one activity or environment to another are called “between-task” schedules and are also being implemented with young children with autism; however, evidence of success is only in case study format.
There is some preliminary evidence for AAC systems for both generating and receiving communication that have demonstrated increases in language and social participation for children with autism. Aided language stimulation is receptive language training in natural environments in which the communication partner highlights or touches pictures while speaking the corresponding words (Elder and Goossens, 1994; Goossens et al., 1995). Aided language stimulation is an interactive, generative use of visual symbols, using a developmental, rather than a behavioral approach. The natural aided language approach, an analog of aided language stimulation, uses visual language as a second language in the child’s environment (Cafiero, 1995; 2000). In natural aided language, every environment has a corresponding language board with the vocabulary needed to provide receptive language stimulation and opportunities for communicative interaction and expressive language. Although there are no published investigations of these AAC approaches used by children with autism to date, there have been two unpublished doctoral dissertations that have demonstrated significant increases in verbal and picture communicative initiations and responses and increases in utterance length (Cafiero, 1995; Dexter, 1998).
The System for Augmenting Language (SAL) (Romski and Sevcik, 1996) is another AAC system that provides augmented input and output with a VOCA. The VOCA has a communication board overlay of visualgraphic symbols; communication partners augment their verbal input with the VOCA as they interact with their nonverbal communication partner. Romski and Sevcik (1996) conducted a 2-year longitudinal study of SAL with 13 students with moderate or severe intellectual disabilities, including a 7-year-old participant with autism. They reported that all of the students used referential and social-regulatory symbols and that seven of the children, including the child with autism, produced messages with multiple symbols, recognized some printed words paired with their corresponding symbols, and increased the proportion of intelligible spoken words.
Facilitated Communication (FC) is a method for providing support to individuals with severe communication problems as they convey typed messages. Supports consist of emotional (encouragement); physical (stable physical context, supporting the forearm or wrist, pulling back the communicator’s hand, helping isolate the index finger); and communicative (ignoring stereotypic behaviors and utterances, using structured questions) components to stimulate communication (Biklen, 1993). FC differs in critical ways from typical AAC interventions. In traditional AAC, practitioners may guide or systematically prompt a communicator. Only
when a communicator is independently accessing the word board, picture board, or keyboard is the communication considered under the authorship of that individual. Because FC involves continued support, within the AAC paradigm FC is considered to be a teaching strategy and motor access mode that is intended to be faded. The essential issue in FC authorship is whether the communication is under the authorship of the child with autism, the facilitator, or the communicator, or is it a collaboration (see Calculator et al., 1995; Shane, 1994).
There are over 50 research studies of FC with 143 communicators. Based on these research studies, the American Speech-Language-Hearing Association (1994) has stated that there is a lack of scientific evidence validating FC skills and a preponderance of evidence of facilitator influence on messages attributed to communicators (ASHA Technical Report, 1994). Thus, there is now a large body of research indicating that FC does not have scientific validity. Therefore, any significant message communicated by a child through FC should be validated through qualitative and experimental analysis.
While quantitative studies reveal no validation for FC, there are several qualitative studies indicating that some children with autism have developed independent communication skills through training in FC. Beukelman and Mirenda (1998) state that there are a small number of individuals with autism around the world who were communicating through FC and are now independent typists. In these cases, it is quite clear that they are the authors of their messages.
The lack of validation of FC with most individuals with autism and the growing body of research supporting the use of AAC with children with autism suggest that FC should only be considered in relation to broader AAC practices by a team that evaluates a child’s progress in achieving independence in communication. The goal of any AAC system for children with autism is independent functional communication without physical support from a communication partner. The development of keyboarding skills, not simply for literacy learning, but for communicative output, is providing considerable promise in the field of autism. Past research that invalidates FC should not preempt research and practice in keyboarding, literacy learning, and AAC as a communication modality for children with autistic spectrum disorders. However, it draws attention to the need for continued evaluation of independence and functional value in using new techniques.
Assistive Technology, Literacy, and Communication
In AAC/AT and autism research, a link is emerging between literacy learning and functional communication, due to the visual nature of reading and writing and the strong visual-spatial strengths characteristic of
the cognitive and processing styles of children with autistic spectrum disorders. Reading skills have been successfully used as the mode to teach spontaneous verbal communication skills. Systematic instruction in the use of written scripts that focus on commenting and questioning revealed that when scripts were faded, children spontaneously and appropriately verbalized those scripts (Krantz and McClannahan, 1998).
Computer-assisted instruction (CAI) includes the use of computer delivered prompts, systematic learning programs, technology based curricular adaptations, writing programs with word prediction, and virtual reality. In a study of four children with autism, Chen and Bernard-Opitz (1993) found more motivation and fewer aberrant behaviors during CAI than during human instruction, though there was variance in the comparative efficacy of these techniques across children. Heimann et al. (1995) conducted an investigation of CAI using an interactive multimedia reading and language software program with 11 children with autism. They documented significant gains in reading, phonological awareness, verbal behavior, and motivation over 5 months. In an investigation of CAI with synthesized speech, Parsons and LaSorte (1993) demonstrated substantial increases in spontaneous utterances when the speech was turned on, compared with when it was turned off and when there was no computer used.
Computer software, such as Boardmaker, enables practitioners to create child-centered, environmentally specific visual language tools for language boards or VOCA displays. Other software programs, such as Picture-It, Pix-Writer, and Writing with Symbols 2000 provide iconic representations for phrases and sentences and can be used to create social stories and adapted curricular materials to augment ordinary auditory and textual information input. To date, there are no published studies on the efficacy of these tools and strategies, although they are gaining popularity among practitioners and parents trained in AAC. In addition, there are no systematic evaluations of computer software that targets children with autism.
FROM RESEARCH TO PRACTICE
Advances in the understanding of autism indicate that the core deficits in communication and language abilities involve joint attention and symbolic capacity. The effectiveness of communication and language intervention programs needs to be documented relative to these core deficits and relative to the target goal of communicative competence in natural language learning environments, with the emphasis on acquisition of functional skills that support successful communicative interactions. The efficacy of communication intervention should be determined by meaningful outcome measures in social communicative parameters, not just
the acquisition of verbal behaviors. Intervention research is needed that helps predict which specific intervention programs or approaches work best with which children. Such research will help families and educators to determine what goals are important and to implement specific teaching strategies designed to best meet those goals.
There is empirical support demonstrating the effectiveness of a range of approaches for enhancing communication skills of children with autism, along a continuum from behavioral to developmental, that differ in both underlying philosophy and specific teaching strategies (Dawson and Osterling, 1997; Rogers, 1996; Prizant and Wetherby, 1998). Single-subject design studies have found that naturalistic behavioral approaches are effective at leading to generalization of language gains to natural environments; generalization has been more limited for traditional discrete trial approaches (Koegel, et al., 1998; Koegel et al., 1992; McGee, et al., 1985). However, there are no group design studies directly comparing the effectiveness of two or more different approaches using randomly assigned, matched control samples (Dawson and Osterling, 1997; Sheinkopf and Siegel, 1998).
Intervention research is not yet available to predict which specific intervention approaches or strategies work best with which children. No one approach is equally effective for all children, and not all children in outcome studies have benefited to the same degree (see Dawson and Osterling, 1997; Rogers, 1996). The most positive outcomes that have been reported have been for 58 percent and 47 percent of the children (Greenspan and Wieder, 1997; McEachin et al., 1993), which means that a large minority of the children did not benefit to this extent. Educators and clinicians could provide extremely useful data by documenting the effectiveness of intervention programs on a child-by-child basis. Based on the available research with this population, progress on language and communication goals should be evident within 2 to 3 months, or different teaching approaches should be considered. In order to determine whether an individual child is benefiting from a particular educational program, measurement of that child’s progress using methods of single-subject design research are helpful.
Shonkoff et al. (1988) propose going beyond traditional measures of language skills to include “ecologically compelling child characteristics” that include more meaningful measures such as a child’s use of core communication skills in natural environments. Since learning in natural environments is the most desirable approach to working with children with autistic spectrum disorders, and spontaneous, initiated language and communicative behavior is of greater value than cue-dependent responding, spontaneity and generalization are particularly important research issues.
The application of functional communication training to the manage-
ment of behavioral problems, and the integration of this approach into overall communication programming remains an area in which continued research can produce value for practice. As work with younger and younger children is undertaken, research that targets goals and documents progress for the core communication skills becomes even more essential, because these skills provide the underpinnings for later social and linguistic competence (Wetherby et al., 2000).
More rigorous research in developmental interventions and interventions that combine or compare naturalistic teaching, focused behavioral and developmental approaches for different aspects of communication and language would contribute valuable perspectives and could contribute ideas for innovative educational techniques. For example, Greenspan and Wieder (1997) suggested that the capacity for complex gestural interaction with shared positive affect was an important predictor to success in their intervention. Future research examining the predictive value of a child’s capacity for joint attention and symbol use could help refine decision-making in treatment and contribute to better understanding of the role of motor functioning in communication and language outcomes.
Studies in autism have focused primarily on child variables and child outcomes. Family variables, considered to be critical to general early intervention research (such as socioeconomic level, stress, supports available, and parents’ involvement in a child’s development), have not been addressed in outcome studies of children with autism (Gresham and MacMillan, 1997). Seminal research on efficacy of early intervention for children with a range of disabilities (Shonkoff et al., 1992) demonstrated that family variables were strong predictors of outcome. Studies of the relationships between family factors and the development and use of communication and language, and the ways in which those factors interact with interventions, would help address this significant gap in understanding.