A Longitudinal Study of the Spontaneous Initiation of Vocal and Gestural Communication in Infants at Heightened Risk for Autism

Background: Autism is characterized, among other things, by impairments in the spontaneous initiation of behavior, especially joint attention, symbolic play, and social communication.
Objectives: Because autism is generally not diagnosed before age two, our goal was to examine the spontaneous initiation of communicative behavior in infancy to assess its possible value for earlier identification. To maximize the likelihood that the sample would contain children eventually receiving an autism diagnosis and to evaluate the initiation of communicative behavior in infants at heightened risk for autism, we observed younger siblings of children already diagnosed with autism (High Risk Group, HR) and compared them to later-born infants with no such family history (Low-Risk Group, LR).
Methods: Fifteen HR and 15 LR infants were videotaped for 45 minutes at home in naturalistic interaction and semi-structured play with the caregiver at 13 and 18 months of age. At 36 months, HR infants were administered the ADOS-G. At that time, two children received a confirmed autism diagnosis. Videotapes were coded for spontaneous (i.e., production initiated by the infant rather than in direct response to adult communication) Vocalizations, Deictic Gestures (Request, Give, Show, Point), and Gesture/Vocalization combinations. Vocalizations were further divided into Words, Communicative Non-Word Vocalizations (occurring with eye contact or gesture to an interlocutor), and Non-Communicative Non-Word Vocalizations (no interlocutor present or no concomitant eye contact/gesture). 
Results: Rates (per 10 minutes) of spontaneous initiation of communication involving Words (p<.001), Communicative Non-Word Vocalizations (p=.04), Gesture/Vocalization combinations (p<.001), and Gives (p<.001) increased significantly from 13 to 18 months for both groups.  At both ages and overall, rates of spontaneous communication by HR infants were significantly lower than those for LR infants for Words (p=.02), Communicative Non-Word Vocalizations (p=.03), Shows (p=.001), and Gesture/Vocalization combinations (p=.001).  Finally, a significant Group effect for Points (p=.05) was qualified by a significant Group by Age interaction (p=.028) such that a small and non-significant group difference in rate of spontaneous Points at 13 months (LR = .186; HR = .148) had ballooned to a large and significant difference by 18 months (LR = .572; HR = .266).  No other main or interaction effects were significant. Finally, even by 13 months the two HR infants later receiving an autism diagnosis stood out.  They produced no Words; but more tellingly, they also produced few spontaneous Communicative Non-Word Vocalizations, only a single spontaneous Deictic Gesture, and no Gesture/Vocalization Combinations. At 18 months they remained at or near the bottom of the distribution on all initiated communication variables. 
Conclusions: Results are consistent with a growing literature indicating that the younger siblings of children with an autism diagnosis are generally at heightened risk for communicative delay. In addition, although more research is needed, results also suggest that even by 13 months, when it is generally too early to expect high rates of spontaneous Word or Point production, greatly reduced rates of initiation of Communicative Non-Word Vocalizations and Shows, both of which are characteristic of typically developing 13 month olds, may index an eventual autism diagnosis.