ABSTRACT
Managing joint attention with communicative partners is essential for successful linguistic communication. For children who use visual systems such as AAC, various factors including partner strategies, environmental constraints and expectations, and children's speed and ease of communication may all influence the effectiveness with which children using AAC learn joint attention control within conversations. This paper compares early joint attention behaviors (e.g. showing an object) in 1- to 2-year-old children with physical impairments based on expectations from typically developing children.  Children at risk for being nonspeaking tended to produce fewer joint attention behaviors than expected for their language stages, regardless of their relative language skill. Implications for clinical intervention with young children relying on AAC are discussed.

BACKGROUND
As typically developing children develop beyond their second year, a greater percentage of their conversation revolves around sharing ideas, rather than early emphasis on requesting and rejecting objects of activities.  Sharing ideas within conversation requires skills at controlling joint attention, or paying attention to the same thing at the same time, usually initiated by an action of either the child or the parent. Joint attention skills are critical for children at early ages to learn the correct association between words and their referents. Successful language learners acquire very early skills at checking for the directing adult attention (Baldwin, 1995). Responding to joint attention cues from adults (e.g. "Look!!" with a pointing cue) is a good predictor of later symbolic language development in children with disabilities (Mundy, Sigman, & Kasari, 1990).

In classrooms, attention to teacher activity is tacitly expected but rarely taught.  This is not unique to classrooms with students relying on augmentative and alternative communication (AAC), but it presents a more complicated issue than for children who are primarily vocal communicators.  If children demonstrate difficulty in shifting attention from their AAC system to the teacher's behavior, this will influence their learning and use of new language skills for classroom communication. Many children, particularly children with pervasive developmental disabilities, have difficulty maintaining or effectively responding to joint attention signals from others.  In fact, joint visual regard may be actively avoided by children with sensory processing problems.

When introducing new AAC language concepts, particularly new picture symbols, partners often redirect the child's attention to that symbol rather than waiting for them to discover it and what it might mean. However, children learn new vocabulary more effectively when the labeling follow their line of attention than when we try to redirect their attention (Tomasello & Farrar, 1986). For children with visually represented AAC strategies, the ability to successfully shift and control other interactants' attention may influence their skill at communicating with multimodal strategies and shifting from their own communication system to that used by their communication partners. If children demonstrate early difficulty with joint attention behaviors, this could reduce their effectiveness at developing many later communicative skills such as socially referenced conversations, information sharing, or topic maintenance and control.

RESEARCH QUESTIONS:
1. Do children with physical impairments demonstrate joint attention behaviors with equivalent frequency as typically developing children at similar language stages, relative to other communicative behaviors such as behavior regulation and social interaction?

2.  What clinical implications for AAC intervention are suggested by young children's patterns of joint attention development and use during nonspeaking communicative exchanges?  

METHOD
Subjects: The research results reported here are part of a 5-year longitudinal study of communicative development in children with physical impairments who are at risk for being nonspeaking.  Children & parents receive 2-4 hour visits in their homes to participate in a variety of cognitive and communicative measures over a period of 18 months each (6 visits every 3 months). The following data are reported on 18 of 30 children, mean corrected age 18.4 mo. (range 12-24). The samples of communicative behaviors and test scores are taken from their first visit only. The children have primary diagnoses of physical impairments and demonstrate risk for vocal expressive development (McDonald, 1980). Children may have secondary diagnoses such as cognitive or sensory impairments (with some useable hearing or vision).  The children's average scores on the Battelle Developmental Inventory (age equivalence in months) were: overall: 8.5 mo., receptive communication: 12.4 mo, expressive communication 9.8 mo.

Measurement:  The Communicative and Symbolic Behavior Scales (CSBS) (Wetherby & Prizant, 1993) was administered to all children who demonstrated intentional communication, which at the first session included only 18 of the 30 children.  The CSBS provides nine communicative "temptations" that prompt children to request and/or comment on play activities, such as presenting a bag of toys and helping the child pull each one out to show adults. Children's communication may be gestural and/or vocal, but must be directed towards an adult to be considered a communicative behavior (e.g. by looking or reaching towards the adult). Communicative behaviors may represent one of three communicative functions: behavior regulation, joint attention, or social interaction. Eighteen children's scores on the CSBS were analyzed, with one behavior sample per child.  Children's raw scores were converted to scaled scores from 1-5, based on the CSBS norm derived from typically developing children at similar language stages (scaled scores are presumed equivalent across functions and can be compared though not added or subtracted to each other). A Wilcoxon Signed Rank Test compared children's CSBS scaled scores for behavior regulation, social interaction, and joint attention.

RESULTS
Joint attention scaled scores on the CSBS were significantly lower than behavior regulation scores (Z Value: -2.260, p = .02). No other comparisons in communicative behaviors were statistically significant. Therefore, children with severe physical and vocal impairments tended to produce fewer joint attention acts than expected relative to their production of behavior regulation acts, across all skill levels included in the analysis.  Children who rely on AAC therefore seem to be consistently slower than typically children at similar language stages to use joint attention behaviors to label or show objects.  Scanning the data suggests that children monitor parent or adult attention within expectations for their communicative rate (using gaze shifts) but are less likely to control parent joint attention.

DISCUSSION
Theoretical Implications: Young children with physical impairments who are nonspeaking were less likely to direct someone else's attention to an object or activity than typically developing children, and tended to have a greater proportion of their communication represented by behavior regulation. Both joint attention and behavior regulation behaviors were equally physically accessible to the children (e.g. showing for joint attention and giving for requesting require similar physical coordination). The relative limitation in joint attention production may occur for reasons similar to those that explain differences in imitation (Cress, Reynolds & Andrews, 1998), i.e. that both types of behavior represent communication as its own end as a "just because" behavior.  Young children with physical impairments may limit their interactive behaviors with motoric "cost" in circumstances without direct functional or communicative benefits.  Since joint attention accomplishes little tangible purpose other than sharing activities or information, this may not provide enough of an observable response for children who have considerable difficulty initiating communicative behaviors successfully. As children grow older, and more of their language- or age-matched peers conduct conversations around shared ideas rather than requests within shared activities, a difficulty in expressing joint attention functions may cause difficulties in conversation in addition to those introduced by the augmented communication modality itself.  

Clinical Implications:  If children are at risk for "just because" behaviors like joint attention as a corollary of their physical impairments, then particular intervention strategies need to be introduced at very early ages to promote joint attention development and reduce this source of communicative risk.  In addition, the communicative and classroom environment needs to reduce as many barriers to joint attention as possible, and provide practice in sharing information that is relevant to children's own actions.  The following suggestions are provided for early AAC intervention:
 

• Provide early emphasis on responsive interaction feedback, in which partners comment on children's behaviors or activity while it is occurring, and help the child shift their attention between objects or actions of focus and the partner's face.  Social routines around objects may provide a natural opportunity for child attention to both objects and people within familiar contexts.
  
• Follow children's attention and label or comment on what they are paying attention to at the time, using the child's own AAC system strategies.  We may also use these episodes as teaching opportunities to introduce new symbols or signs, but reduce our verbal redirection of children's attention as much as possible in favor of situationally prompting attention and then providing meaningful labels.
  
• Include questions like "what's that?" early in children's behavior and/or symbol repertoires, and model these phrases frequently in communicative routines.
  
• Include words like descriptors, adverbs, and question words earlier on topic boards, and in children's communication systems so that children can make creative and relevant joint attention comments when they may only be able to access a few items at a time.
  
• Imitating a child's own behaviors is an early strategy for fostering turn taking and new joint attention behaviors.  This can be particularly effective for children with PDD (Prizant & Wetherby, 1987).  However, children with physical impairment are less likely than language-age peers to imitate behaviors without a functional context (Cress et al, 1998).