Interventions to Improve Communication

Didactic Approaches

A large body of research has demonstrated that didactic approaches are an effective means of initially developing attention to and understanding of language, as well as initiating speech production in preverbal children with ASD. Discrete Trial Instruction (DTI) entails dividing the chosen skill into components and training each component individually, using highly structured, drill-like procedures. Intensive training utilizes shaping, prompting, prompt fading, and reinforcement strategies. Trials continue until the child produces the target response with minimal prompting; at which point the next step in the hierarchy of behaviors (such as correctly pointing to the named picture from among two pictures) is presented and trained.

Several group studies have shown that structured intervention based on behavioral principles is useful in improving expressive language in children with ASD [28–30]. A relatively large literature based on single case studies has also demonstrated the efficacy of these approaches in eliciting vocal imitation [31], and speech from nonverbal children [32], [33], [30]. However, these approaches rely heavily on teacher direction, prompted responses and contrived forms of reinforcement. An inherent weakness in didactic approaches lies in the fact that they often lead to a passive style of communication, in which children respond to prompts to communicate but do not initiate communication or transfer the behaviors acquired to situations outside the teaching context [34]. These difficulties in generalizing and maintaining behaviors taught through didactic approaches, along with changes in theoretical views of language learning that emphasized the central role of social exchanges in the acquisition of language, led to the introduction of more naturalistic methods of intervention.

Contemporary Applied Behavior Analysis (ABA)

A variety of practitioners of applied behavior analysis [35–37] have confronted the major shortcoming of classic behavioral approaches: the lack of generalization. Hart & Risley [38] were the first to attempt to apply operant principles to more functional communicative situations. Their major insight concerned the importance of making access to the object of the child’s interest contingent on the child’s initiating some communicative exchange about it. These contemporary ABA, or naturalistic methods share some aspects of the didactic approaches from which they are an outgrowth. They are moderately teacher-directed, address goals specified by the adult, and rely on reinforcement, although the reinforcement is more intrinsic (the attainment of the child’s desire, or a social reinforcement) than the tangible rewards used in more traditional DTI methods. The effectiveness of these methods has been amply documented in the literature, at least in single case studies, both for initiating speech in previously nonverbal children [39], as well as for increasing the complexity of spoken language [40]. Direct comparisons of TD and naturalistic approaches have shown some advantage for the more natural techniques, including maintenance and generalization of new behaviors [41], although not all investigators accept this [42].

Several techniques that fall within the category of contemporary ABA approaches to communication intervention include prompt-free training [43], incidental teaching [38, 44], and mand-modeling in which the caregiver requests a response from the child [45]. Milieu teaching is an umbrella term that refers to this range of methods that are integrated into a child’s natural environment [24]. The early language skills of children with a range of developmental disorders, including autism, have been shown to be enhanced through Milieu teaching methods [46], [35], [47, 48]). Milieu teaching approaches include:

• Training in everyday environments

• Activities that take place throughout the day, rather than only at “therapy time.”

• Preferred toys and activities are included so that participation in activities is self-reinforcing.

• Encouragement of spontaneous communication by using “expectant waiting” and refraining from prompting.

• Waiting for the child to initiate teaching episodes by gesturing or indicating interest in a desired object or activity.

• Providing prompts and for expansion of the child's initiation.

• Rewarding child responses with access to a desired object or activity.

In general, the Milieu approaches like these have been shown to be associated with increased ability to initiate communication in children who did not show this ability previously [49], [50]. Nonverbal children have developed speech using these methods [51], and increases in the frequency, spontaneity and elaboration of language have also been documented [41], [24], [52]. Yoder and Stone [48] showed that children with autism who had some communicative behavior prior to intervention were more likely to develop speech in this program than were children with fewer communication behaviors at baseline.

In addition to the specific techniques described here, several comprehensive intervention programs have been designed within the contemporary ABA format. Some examples are listed in Table 1. A few studies [28, 29] have shown these programs to be more effective in improving language and cognitive outcomes, both short-term and over the course of two years, than eclectic community interventions.

There is one significant problem with all the naturalistic ABA methods, however. Because of their reliance on allowing children to initiate, requiring interventionists to make more on-line decisions, they involve a significant level of training for interventionists, and rely to some degree on clinician sensitivity and skill. Although it has been demonstrated that both parents and peers can be taught to deliver naturalistic interventions successfully [32], [53], [54], these methods, nevertheless, require more moment-to-moment decision-making than didactic programs in which adult actions are clearly specified, unambiguous, and easily trained. Moreover, unlike didactic approaches, there are few materials available that outline comprehensive curricula; nor are there training opportunities that make it possible to master the approach from independent study of published manuals or through inservice training. Currently, then, it is difficult to disseminate these methods faithfully and efficiently to interventionists beyond those who created the approaches and their direct trainees.

Developmentally-Based, Social-Pragmatic Strategies

This approach is based on the assumption that children with autism develop language in the same way and following the same sequence as do typically speaking children. This means that children who do not speak will first be encouraged to use other means of communicating to get intentions across, in order to discover the value of communication in the ability to regulate others’ behavior and control interactions. Proponents of this approach, then, advocate the use of a variety of nonverbal forms of communication being used as a “stepping stone” to speech. The central tenets of this group of approaches include:

• Following the child’s attentional lead; allowing the child to choose the course of the interaction and use of materials;

• using the normal sequence of communicative development provides the best guidelines for determining intervention goals;

• providing intensified opportunities for children with ASD to engage in activities that are similar to those in which typically developing peers engage, in the belief that these are the most effective contexts for learning social and communication skills;

• learning opportunities (“teachable moments”) that naturally arise in the course of interactions should be exploited, rather than relying on a predetermined curriculum. Intervention is carried out in the context of natural daily routines, such as bathing, dressing, and feeding.

• targeting functional goals for intervention; that is, targeted behaviors should be applicable to daily, meaningful activities in a variety of contexts apart from the one in which they were originally learned;

• assuming the “prerequisite” role of nonverbal communication, including gestures, gaze, vocalization, and other nonvocal means, in the development of language.

Augmentative and Alternative Communication Strategies

For children with ASD with significantly delayed development of speech, alternative methods of communication are often introduced. Originally developed to assist individuals with severe sensory or motoric impediments to speech production, these strategies, commonly referred to as AAC, provide nonvocal options for communication. The main tenet of this approach is that all children have a need to communicate, and if speech is not present, communication need not wait for it to develop. Instead, the child is taught other ways to express his intentions. Below are some examples of AAC methods that have been used with children with ASD.

Summary

Developmental-pragmatic approaches are widely used and advocated by many prominent communication specialists. One problem with these methods is that, to an even greater extent than naturalistic approaches, they require a high degree of sensitivity, creativity, and on-line decision-making on the part of interventionists. Programs like Hanen© have demonstrated that parents, teachers, paraprofessionals and others can learn these methods, but they require extensive training, practice, and ongoing support. Moreover, demonstration of child outcomes in response to adult training is weak. To some degree, then, the success of developmental-pragmatic programs may hinge on the careful training, follow-up, support, and talent of those who deliver the intervention.

At this time, developmental-pragmatic approaches have a narrower base in empirical research than the other two approaches to communication intervention in ASD. Some studies have shown them to be effective for eliciting speech in children with other developmental delays (e.g., [61], [75]). In addition, studies of child-centered methods have demonstrated their ability to increase imitation, gaze, turn-taking, and joint attention [76], [77] in preverbal children with autism. Kasari [78] and Whalen, Shreibman, & Ingersoll [79] report on small sample studies showing improvement in communication following training focused specifically on eliciting joint attention. Yoder and McDuffie [47] review research that suggests training children in play and nonverbal communication skills may improve the ability to acquire speech.

Conclusions: Early Communication Intervention

Programs to elicit initial communication behaviors and first words range from highly structured behaviorist to open-ended, child-directed methods, with a range of naturalistic approaches in between. Didactic and naturalistic ABA methods are aimed specifically at eliciting speech from preverbal children with ASDs and have established efficacy for doing so in single case studies, although controlled experiments with random assignment to treatments, the “gold standard” of scientific evidence, are still relatively few as of this writing. Developmental-pragmatic approaches are aimed more broadly at the improvement of social communication and interaction, and have a less well-established, though emerging, empirical track record for eliciting first words in children with ASDs. They have been shown, in small sample and case studies, to increase preverbal behaviors such as imitation and joint attention. Developmental approaches that incorporate AAC methods such as Signs and pictures have been shown to be compatible with the development of speech, although again, their efficiency relative to straightforward speech treatment has not yet been established. Importantly, too, little research exists on the relationship between child characteristics and intervention efficacy, so that we do not know at this time which approach will be the ‘best match’ for a particular child. What does appear to be clear is that communicative deficits in preverbal children with autism are amenable to treatment, and that a range of treatments have been shown useful in enhancing communicative behavior. Core Communicative Deficits in Speakers with ASD

Once children with autism begin using words, the course of language development is similar to that seen in other children with disabilities [80]. Even, however, when children with ASD acquire basic language skills, significant challenges to the development of communicative competence remain. Tager-Flusberg, Paul, and Lord [81] point to two commonly observed characteristics of the communication of speakers with ASD:

• Echolalia, the imitation of what has been heard, either directly after it is spoken, or as a delayed echo at a later time.

• pronoun reversal, primarily the tendency to use ‘you’ instead of ‘I.’

Although this was at one time thought to reflect difficulties in ego formation, it is now seen as another instance of echolalia, in which the child refers to himself as ‘you,’ as he has heard others refer to him [82].

Both of these characteristics tend to decrease as language skill increases, and both are seen as transitional behaviors in typical development, as well [81]. Few programs have addressed pronoun reversal directly, since it tends to abate naturally, but a variety of case studies, have reported using behavioral methods successfully to reduce echolalia and replace it with simple request and labeling behaviors [83–85].

Deficits in prosody or musical aspect of speech, including its rate, loudness, pitch, voice quality, and use of stress are also frequently seen in speakers with ASD [86]. Sheinkopf, Mundy, Oller and Steffens [17], as well as Dawson, Osterling, Melztoff and Kuhl [87] showed that even prelinguistic vocalizations in young children with autism contained a significantly higher proportion of atypical vocal characteristics than did those of normal children. Paul, Shriberg, McSweeney, Klin, and Volkmar [88], found that approximately half of high functioning speakers with autism were rated as atypical in elements of prosodic production. When these differences are present they tend to be persistent and to show little change over time, even when other aspects of language improve [89], [90], [91]. Paul, et al. [88, 92] showed that ratings of social and communicative competence were related to ratings of prosody in speakers with ASD, suggesting that difficulties in prosodic production affect listeners’ attributions of competence to these individuals.

To date, there has been little research on the treatment of prosodic deficits in autism. Several commercially available programs have been developed for addressing prosodic difficulties in non-autistic speakers (e.g., [93], [94]), but no research is available on their use with children with autism. Addressing prosodic deficits remains a largely unmet need in this population.

Pragmatics, or appropriate use of language in social situations, is, of course, the most prominent aspect of communicative deficit in this population [95]. Children with autism are less likely than those with typical development to initiate communication, particularly with peers [96]. Overall rates of communication are very low, even in children who speak [97]. These children show reduced interest in language spoken to them; they are less likely to respond in a reciprocal fashion to the communicative bids of peers and more likely to produce self-directed, noncommunicative speech [95]. Paul et al. [98] reported that the following areas of pragmatics were most consistently impaired in speakers with ASD:

• use of irrelevant detail

• inappropriate topic shifts;

• topic preoccupation/perseveration;

• unresponsiveness to partner cues;

• lack of reciprocal exchange;

• Inadequate clarification;

• Vague references;

• Scripted, stereotyped discourse;

• Excessively formal style (for speakers with Asperger syndrome only)

Thus, for children with autism who develop speech, the conversational skills involved in managing turns and topics in discourse, flexibly adopting an appropriate style of speech keyed to the characteristics and setting and conversational partners, and inferring what information will be relevant and interesting to others constitute the greatest difficulties. Unfortunately, there are very few comprehensive, sequenced curricula designed to address these pragmatic deficits. The following discussion reviews a range of programs available for addressing language issues in speakers with ASD, using the same tripartite scheme as was used above.

Didactic Approaches

ABA programs have been developed to address the expansion of early symbols into more elaborated forms of expression. They focus primarily on improvement of language form; i. e., vocabulary and sentence structures. One example is Teach Me Language [99], a comprehensive language program that provides a step-by-step guide with in-depth detail on intervention activities targeting language areas such as grammar, syntax, concepts and advanced narrative skills. Teach Me Language methodology is behavioral-based. Children are expected to follow a teacher’s lead and regular repetition of drills is a key feature in the program.

The Verbal Behavior program takes a Skinnerian approach to language learning, based on the research of Michael, Sundberg, and Partington [100], [101]. The program, like Teach Me Language, provides a carefully sequenced curriculum for teaching language to children just emerging into symbolic communication, and carries the curriculum through to those learning more advanced language forms. It uses a highly structured behavioral approach, incorporating techniques such as errorless teaching, specific quick-transfer (prompting and fading) procedures, and the use of discrete trial training during both intensive teaching sessions and in more naturalistic contexts. Language goals are structured in terms of Skinnerian categories of verbal behaviors, which include (hierarchically ordered):

• Echoes: practice in imitating verbal behavior;

• Mands: verbal behaviors that produce an immediate benefit for the speaker (e.g., requests);

• Tacts: labels;

• Reception by Feature, Function, and Class (RFFC): responding to commonly used verbal stimuli (words);

• Intraverbals: verbal, nonechoic responses to the speech of others.

Just as in the case of the didactic approaches for children at prelinguistic communication levels, these highly behavioral language teaching programs also have the potential weakness of leading to passive styles of communication and limited levels of generalization. A search of reference databases found no reports of empirical evidence to support the specific use of the Teach Me Language program. However, Partington, Sundberg, and colleagues have published data indicating increases in verbal production using their method (e.g., Partington [101], [102], [103]. Still, little research is available on the functional effects of these programs on real-world communication, or on their consequences for adaptive communication and independence. While they may result in the achievement of target behaviors within the behavioral framework, as most ABA programs do, there are as yet few data on their general, long-term effects on the child’s functioning or independence.

A slightly less traditionally behaviorist approach, the PECS program, also provides a curriculum for children who have acquired basic symbolic communication skills. Once single word exchanges have been mastered, sentence structure is targeted. The aim is for the child to combine an “I want” picture with a picture of a desired item or activity. A “sentence strip” is used to attach the pictures and this is passed to the communication partner. The sentences are further extended by the addition of other words such as adjectives, such as, “I want the red ball.” Sentence strips are often color coded, so that, for example, there may be a green area for noun pictures that are framed in green, a blue area for verb pictures framed in blue, and a red area for adjective symbols framed in red, and so on. The next phase of the PECS moves from using the picture exchanges for requesting to encouraging commenting. This is done by introducing picture cards representing phrases such as, “I see”, “I smell” and “I hear”. Again, research on the efficacy of this stage of the program, or its generalization to functional communication, has not been reported.

Naturalistic Approaches

There is much less research in the effectiveness of naturalistic ABA programs in improving communication for speakers with ASD than there is for prelinguistic children. McClannahan and Kranz [104] provide an approach with empirical support that employs using scripts, either pictured or written, to guide children with autism on what to say in given social situations. The scripts are developed by teachers with the students to address particular social settings. As the children continue to rehearse the scripts with various adults and peers, pieces of script are removed, or faded, so the child is required to provide longer and longer portions of the language on his own. Krantz and McClannahan [105] have presented data suggesting this method leads to improvement in conversation for children with ASD.

Video modeling [106] is a similar procedure. Here, short videos are constructed that illustrate the language used in particular social situations, such as asking for a book from a librarian. The student watches the video with the instructor, discusses and verbally rehearses the scene, then practices it by role-playing with an instructor and eventually trying it in a real setting. Charlop-Christy et al. showed that this procedure is more effective than simply modeling the desired behaviors in vivo for children with ASD.

Brinton et al. [107] developed a method for improving conversational turn-taking and topic maintenance. In a case study, they reported on work, over a two-year period, with a boy with social disability who was taught a set of explicit rules for finding topics of common interest with peers and engaging in conversation about them. Brinton et al. stress that change was slow, but did report improvement in social acceptance as conversational skills improved.

Focused stimulation is a naturalistic means of increasing receptive language skills in children with a variety of disabilities. This technique involves providing an interesting set of play materials, and using simple, repetitive language to talk about the ongoing action in concrete, here-and-now terms, using many examples of forms the child needs to acquire. One single case study [108] provides preliminary support for this strategy in children with ASD.

In addition, many of the comprehensive naturalistic ABA programs outlined in Table 1 provide examples of intervention methods for children who begin speaking.

Developmental-Pragmatic Approaches

There are few examples of detailed programs that take a child-centered approach to expanding basic language in speakers with autism. Prizant and Wetherby [109] advocate using the Social Communication Emotional Regulation Transactional Support (SCERTS) model, a comprehensive approach that allows the incorporation of an eclectic range of treatment methods and focuses on overarching goals that include improving social communication, encouraging behavioral self-regulation, and providing transactional supports to children with autism. This program has not yet been subjected to empirical study.

Quill [110] has presented another comprehensive curriculum for developing social and communicative skills in young children with autism at various levels of functioning. The curriculum is also eclectic and advocates the full range of highly structured, naturalistic approaches and child-centered methods. It suggests focusing on the child’s responsiveness to typical peers, rather than promoting initiations (p. 89). Intervention guidelines include organization of the environment to facilitate participation and cooperation, careful selection of materials, and activities structured to foster the target child’s participation. Again, empirical support has not appeared for this program.

Child-centered methods discussed at the prelinguistic level, including “Floor Time” [111] and RDI [58] also contain components that can be used at higher language levels. Sussman [60] has developed a companion volume to More than Words, called Talkability [112], which is aimed at children who speak. No empirical validation of these methods for improving language skills in children with ASD is available as of this writing.

Increasing social communication

However, as we have seen, a major difficulty for children with ASD who speak, and perhaps THE central problem for students with high-functioning autism and Asperger syndrome concerns not the forms of language, but its use in social contexts, particularly with peers. Apart from elaborating the form and increasing the frequency of language use, then, an additional important goal of communication intervention for children with ASD is to provide supports that allow these children to engage in peer interactions, including pretend play, games, and conversations. Some of the naturalistic approaches discussed already, such as script-fading and video-modeling, go some distance to addressing this issue. McClannanhan and Krantz [104] provide what comes closest to a comprehensive, sequential program for teaching conversational skills to this population.

Much of the literature addressing the more comprehensive problem of increasing social communication opportunities and skills in speakers with ASD is discussed under the rubric of “social skills training,” a topic addressed in detail elsewhere in this issue. The main point to be stated here, however, is that for children with ASD who speak, speaking is not, in itself sufficient. These children will need supports aimed not only at increasing basic vocabulary and sentence structure, but at the pragmatic aspects of language use in the context of social interactions, to address deficits like those identified by Paul et al. [98]. Studies of the effects of a variety of approaches to social skills intervention (See [113] for review) generally concur that trained peers are more effective agents of this intervention that are adults [114], and that the interventions are more effective when they take place in a child’s natural environment, such as the classroom, than in a clinical setting [115].