"When will my child talk?" is a common question that speech-language pathologists hear from parents of young children with autism spectrum disorders (ASD). This question signals the anxiety of parents coming to terms with an unexpected vision of their child's development. SLPs want to respond in a credible way that is sensitive to parents' hopes for their child. No one fully understands why some children with ASD develop speech and others do not. In this article, we discuss recent research on predictors of language outcomes for children with ASD, and applications to clinical practice.
We include the views of four leading autism researchers about the implications of their research for SLPs working with children with ASD—Catherine Lord, professor of psychology and psychiatry, University of Michigan; Pamela Wolfberg, associate professor of special education, San Francisco State University, and founder of the Autism Institute on Peer Relations and Play; Connie Kasari, professor of education at University of California at Los Angeles; and Amy Wetherby, L.L. Schendel Professor of Communication Disorders, Florida State University.
Predictors of Language Outcomes
Individual child characteristics (e.g., nonverbal cognitive level) are one type of predictor associated with outcomes across several areas of development, including language. Other factors more external to the child, such as interventions and parent interaction strategies, also affect language development. Recent research is providing clues about child characteristics and external factors that influence language outcomes for children with ASD.
The specific autism spectrum diagnosis of a young child is associated with varying language outcomes. The literature we reviewed for this article focuses on children with diagnoses of autistic disorder and pervasive developmental disorder-not otherwise specified (PDD-NOS). Children with autistic disorder generally achieve lower levels of language compared with those with PDD-NOS (Anderson et al., 2007). This finding is unsurprising, as children with PDD-NOS by definition show fewer ASD symptoms (see sidebar on p. 10 for ASHA's documents on ASD).
Early Language and Nonverbal Cognitive Skills
Early language and verbal ability are strong predictors of later language among children with ASD (Smith, Mirenda, & Zaidman-Zait, 2007). Variability in expressive language for children with ASD is considerable and puzzling. As Catherine Lord, University of Michigan, explains, "It looks to us that the trajectory of language development is much less linear for children with autism. That is, they don't just go steadily from one word to 20 words...And that means that you have to be much more careful when you are drawing inferences about a child's language status at 2 years and what that will mean at age 10 or 14." Some young children with ASD are nonverbal throughout the first few years of life, while others may engage in verbal imitation and echolalia, and/or use words or phrases independently to communicate. Early language comprehension also is a strong predictor of later verbal and nonverbal development for toddlers with ASD (Wetherby, Watt, Morgan, & Shumway, 2007).
In the research of Lord and colleagues, the nonverbal cognitive ability of children with ASD at age 2 also was a strong predictor of later language as measured at age 9 (Anderson et al., 2007). This finding highlights the need for interdisciplinary assessment and intervention teams, to help identify the full range of abilities and goals for children with ASD that may maximize social-communicative outcomes.
The fact that early language and cognition predict later language for children with ASD may seem obvious, yet these findings reinforce the importance of starting interventions with children with ASD as early as possible. Beyond this, however, research has identified specific foundational skills that predict language outcomes in ASD. Social and joint attention, imitation, and play have been especially salient variables, which are discussed below.
Social and Joint Attention
Most infants demonstrate a keen interest in people even as newborns. Infants preferentially attend to certain stimuli that may assist them in language learning. For instance, they prefer "child-directed speech," characterized by positive affect, higher and more variable pitch, and simpler, more repetitive structure and content, to "adult-directed speech." Tuning in to caregivers' talk and other communicative behaviors (e.g., gaze, facial expression, gestures) likely promotes early language development. By contrast, children with ASD often do not prefer child-directed speech, and may even prefer non-speech stimuli (Klin, 1991; Kuhl, Coffey-Corina, Padden, & Dawson, 2005). Early failure to tune in to others may be one reason for the severe language delays of many children with ASD. Our own research suggests that the extent to which young children with ASD attend to child-directed speech is highly related to their concurrent and later language (Watson, Roberts, Baranek, Perryman, & Mandulak, 2007).
Joint attention—the ability to coordinate attention between another person and some object or event in the environment—is an important foundation for language development (Carpenter & Tomasello, 2000). Typically the ability to respond to and initiate joint attention bids begins as early as 3-6 months and is well-established by 18 months of age, with refinements up to at least 3 years of age.
Children with ASD, conversely, have severe joint attention impairments. Multiple studies have demonstrated that children with ASD are able to coordinate attention to make requests better than for purely social purposes. Parents and SLPs are aware that children with ASD often make requests by taking the adult's hand and leading the adult to a place. However, these children rarely look at us just to "check in" or show or give us a toy just to engage in a social-communicative exchange. Such deficits in joint attention are among the earliest markers of ASD and are important predictors of later language outcomes of children with ASD (Anderson et al., 2007; Wetherby et al., 2007).
Almost 30 years ago, Elizabeth Bates and her colleagues demonstrated that imitation skills predict language ability in typically developing children, who are able to imitate at birth and begin to imitate actions on objects as early as 9 months (Bates, Benigni, Bretherton, Camaioni, & Volterra, 1979).
Children with ASD demonstrate significant impairments in imitation, including imitation of face and body movements and imitation of actions on objects. The relationship between imitation and language development in children with ASD has been the focus of several studies, which found that the earlier imitation abilities of children with ASD are predictive of later language skills (Charman et al., 2003; Stone & Yoder, 2001; Toth, Munson, Meltzoff, & Dawson, 2006).
Jean Piaget's writings in the early to mid-20th century sparked considerable research into the developmental relations between play and language in young children. Although debate continues over how and why play and language development are linked, considerable research has demonstrated they are related to each other, especially symbolic play and language (Casby, 1997; Spencer, 1996).
Marked quantitative and qualitative differences exist in the play of children with ASD compared with typically developing peers. Overall, toy play of children with ASD is more repetitive and less flexible and elaborate (Williams, Reddy, & Costall, 2001). In addition, play skills do not progress in a predictable developmental pattern (i.e., moving from exploratory to relational to functional to symbolic play). Exploratory play continues to dominate past the time at which typically developing children incorporate more sophisticated play strategies (Jordan & Libby, 1997). Children with ASD spend less time than their typically developing peers in functional or symbolic pretend play, especially when their play is not structured or supported by other people (Jarrold, Boucher, & Smith, 1996). Early pretend play skills are another predictor of later receptive and expressive language skills for children with ASD (Charman, et al., 2003; Toth, et al., 2006).
Although clinicians use a variety of play-based interventions, Pamela Wolfberg (personal communication, 2007) identifies five main themes reflected in current research-based efforts:
- Recognizing the importance of play for children's learning, development, and socio-cultural participation
- Adapting to each child's unique interests, developmental levels, and learning style
- Identifying and responding to what is intrinsically motivating for the child
- Supporting opportunities that include typically developing peers as play partners
- Being open to blending "best practices" to meet each child's full potential for play
For children with typical development or disabilities, high levels of parental responses contingent on the child's focus of attention are related to better language and developmental outcomes (Trivette, 2003). As a group, parents of children with ASD are essentially no different from parents of typically developing children in terms of following their child's focus of attention and adding language to their child's play (Watson, 1998).
Given the link between these prelinguistic abilities and language outcomes for children with ASD, a natural question is whether language outcomes for children with ASD can be improved by focusing on joint attention, imitation, symbolic play, and caregiver contingent interaction strategies in early intervention. Research is beginning to support the value of interventions addressing some of these key predictors of language outcomes.
Several recent studies have targeted improvements in joint attention. Kasari's research group examined the relative benefits of symbolic play or joint attention intervention for preschoolers with autism who were all enrolled in an applied behavior analysis (ABA) program (Kasari, Freeman, & Paparella, 2006). "We got what we taught," Connie Kasari summarized. "If we taught joint attention, we got better joint attention, and if we taught symbolic play, we got better symbolic play. If you don't focus on either, you will likely not get change. It seems clear that the children need direct instruction." More importantly, children who participated in joint attention or symbolic play interventions showed better expressive language skills than children who only participated in ABA after a one-year period (Kasari, Paparella, & Freeman, 2008).
Another relevant finding was that children who had five or fewer expressive words prior to intervention benefited more from the joint attention intervention than from the symbolic play intervention. In fact, the children with low language levels in joint attention intervention showed an impressive 13-month gain in expressive language in a one-year period. Although additional research is warranted, this finding implies that joint attention intervention should be a priority for children with very limited expressive language.
Yoder and Stone (2006a, b) recently compared the effects of a Responsive Teaching/Prelinguistic Milieu Therapy (RT/PMT) approach to intervention using the Picture Exchange Communication System (PECS) for preschool children with ASD. All children entered intervention with very limited or no expressive language. Children in both intervention groups showed strong gains in language. In terms of gains in spoken language, the PECS intervention was more effective overall, but Yoder and Stone provide important information to help match interventions to child characteristics.
The effects of the PECS intervention were maintained best in children who began the intervention with more varied interests in different objects. Children with low interest in objects at the beginning of the intervention showed better long-terms gains in language if they participated in RT/PMT intervention, which included a focus on toy play (2006a). Stone and Yoder (2006b) also examined the effects of intervention on communication functions; in these analyses, they found that the RT/PMT intervention resulted in more improvements in joint attention, but only for children who began the intervention with at least some joint attention skills. This research serves to remind us that the most effective intervention may vary based on the intervention goal(s) as well as the individual characteristics of the children.
Imitation is a component of many intervention programs for young children with ASD, and is often suggested as a "pivotal skill" that can be used to acquire other abilities (e.g., play skills). Ingersoll and Schreibman (2006) examined changes in various domains following an intervention program using naturalistic strategies to teach children with ASD to imitate actions on objects. Children showed gains in spontaneous imitation, joint attention, pretend play, and language. The results lend support to the value of working on imitation skills as a component of an early intervention program.
Parent education also has been shown to be both effective and important to parents. An extensive review of intervention research in ASD identified parental involvement as one of the key elements of effective programs (National Research Council, 2001). Several parent programs that emphasize caregiver responsiveness—Responsive Teaching (Mahoney & MacDonald, 2007), Relationship Development Intervention (Gutstein, Burgess, & Montfort, 2007), and Floortime (Wieder & Greenspan, 2003)—have reported promising outcomes for children with ASD, including gains in language, cognitive, and social-emotional skills, although more rigorous research is needed. ASHA's guidelines on ASD provide a summary of evidence related to intervention approaches for ASD (see sidebar).
Implications for Practitioners
Despite the growing evidence related to predicting language outcomes in young children with ASD, research cannot provide precise answers to parents who ask, "When will my child talk?" Particularly in working with toddlers, we recommend that the findings we have reviewed be used not to try to predict whether or when a child will talk, but rather to focus assessments and interventions on aspects of preverbal behavior that will increase the likelihood of good language outcomes ( Figure 1 [PDF]). The following assessment and intervention recommendations address important aspects of comprehensive clinical programs.
Given the contributions of nonverbal cognitive skills, joint attention, play, and imitation to social-communication development, the need to evaluate these core abilities in children with ASD is clear. Assessing the child's interest in toys, play skills, imitation, and joint attention abilities can help SLPs choose a communication program. In addition, SLPs should assess what children can do with and without supports (i.e., spontaneous vs. prompted behaviors). "Standardized tests put the child in a respondent role," Amy Wetherby explained, "and you've got to go beyond that and see what children do when they need to initiate as well as respond in an ongoing interaction." To get a complete picture of a child, SLPs need to collaborate with parents and other professionals.
Involving parents in assessment is essential to determine such influences as home routines, how the child functions within those routines, parent-child interactions and parents' level of satisfaction with those interactions, parent priorities, and intervention approaches that best fit with parent values and styles as well as with child needs. Comprehensive assessments provide a fuller understanding of the child and family, increasing the likelihood that we will give parents information, strategies, and resources.
Early intervention targeting joint attention, imitation, and play abilities appears to have cascading benefits for the language and social development of children with ASD. Thus, SLPs and other service providers should target these abilities, and use assessment information about specific child characteristics and family needs to select and administer effective intervention approaches. Cathy Lord summarized some key insights gleaned from early intervention research in ASD. "Keep the children engaged," Lord said. "Help them understand how their communication can have an effect, and increase the sheer volume of practice in terms of how many times a day they get to communicate when someone responds to their communication in a positive way."
Understanding the content and strategies of a variety of communication intervention approaches will help interventionists determine the best match for the child and family. SLPs' expertise in social-communication skills makes us uniquely suited for intervention with children with ASD.
"Parents may not be aware of the importance of prelinguistic skills," noted Connie Kasari, "but speech-language pathologists understand these precursors [joint attention, play skills]."
"That's why SLPs are the perfect people to provide education for teachers and families. A core deficit in autism is language, and we need to get these kids talking by the time they are 6 or 7. Who better than SLPs?"