Joint Attention and Language Development in Young Children with Autism

Background: Joint attention (JA) occurs when two individuals focus on the same object or event (Baldwin, 1995). There are two major kinds of joint attention: response to joint attention (RJA), which refers to the process by which children follow the attentional focus of their social partners, and initiation of joint attention (IJA), which refers to the children's ability to direct their social partners' attention on an object or event (Corkum & Moore, 1995). Impairment in joint attention is an early sign of autism, and may be a major reason for the delayed language development seen in individuals with autism. Objectives: To investigate the JA of young children with autism across a 2-year time span. We report data from the first two visits (4 months apart). Methods: We included 10 typically developing toddlers (TD: mean age = 20.45 months), and 12 children with autism (ASD: mean age = 32.32 months), who were matched on expressive vocabulary at Visit 1. Children engaged in a 30-minute play session with their parents. Sessions at Visit 1 were coded for (a) duration and number of RJA episodes, (b) duration and number of IJA episodes, (c) duration and number of episodes in which the parent attends the same object as the child, but the child only passively participates (Passive Attention), (d) pointing and eye contact during a JA episode. At Visit 2, children's language during the play session was coded for (a) total number of words (types and tokens), (b) mean length of utterance (MLU), (c) wh-words and questions (types and tokens), (d) verbs (types and tokens). Children were also administered the MacArthur Communicative Development Inventory (CDI), Vineland Adaptive Behavior Scales, and Mullen Scales of Early Learning (Visit 1). Results: Compared to TD children, the ASD group engaged in significantly shorter episodes in which they had responded to JA (ASD M = 0.9 minutes per episode; TD M = 1.2 minutes per episode, p < .05), and fewer episodes of JA which they initiated (ASD M = .16; TD M = 1.0, p < .05). Moreover, children with ASD engaged in more episodes of Passive Attention with their parents (ASD M = 10.33; TD M = 1.5). For children with ASD, the duration and number of RJA episodes positively correlated with their scores on the CDI (Visit 1), Vineland (all subscales), Mullen (all subscales), as well as their total number of utterances and verb type (both types and tokens), and MLU. Their duration of IJA episodes correlated positively with their wh-words and questions (ps < .01). Regression analyses demonstrated that, for the ASD group, higher Mullen Expressive Language scores and longer RJA episodes at Visit 1 predicted both types and tokens of total words, and verb tokens at Visit 2. Conclusions: This study showed that toddlers with ASD engage in less joint attention with their mothers and initiate fewer JA episodes compared to TD children. We conclude that both RJA and IJA seem to be related to later vocabulary development in children with ASD.