Vocal Behavior in First Year of Life Is Associated with Autistic Symptoms in Infants at Risk for ASD

Background: Prelinguistic vocal behavior is known to be related to speech development in typical children (Oller, 1999; McCune & Vihman, 2001) and has been shown to be atypical in preschoolers with ASD (Sheinkopf et al., 2000). Expressive language delay is a nearly universal feature of the development of children with ASD (Tager-Flusberg et al., 2005). Objectives: The present study examined prelinguistic vocal behavior in infants at high risk (HR) for developing ASD as a result of having an older sibling diagnosed with ASD, and compared these to vocalizations of low risk (LR) infants who did not have a diagnosed sibling. Our aim was to determine whether HR infants show differences from typical development in terms of their production of speech sounds, syllable structures, and prosody in the prelinguistic period, and whether such differences were related to outcomes in the second year of life. Methods: Subjects were drawn from those participating in a longitudinal study of behavior and development of infant siblings of children with a diagnosis of ASD. As part of participation in the longitudinal study, infants are seen several times during the first year of life. At each visit, standard assessments of behavior and development, as well as data from experimental tasks were collected. The present report presents cross-sectional analysis of vocal productions collected at the 6, 9, and 12 month visits. Vocal samples were audiorecorded during a five-minute free-play mother-child interaction, in which mothers were presented with quiet toys, asked to play with their baby and to attempt to encourage the baby to vocalize. Analyses included: Consonant inventory (Shriberg & Kwiatkowki, 1994) % Canonical syllable production (Oller, 1998) Atypical prosodic productions (Sheinkopf et al., 2000). Analyses were completed by two trained raters with experience in phonetic transcription who were blind to participants' age and risk status. Point-to-point reliability for ratings was over 80%. Results: No significant differences were seen vocal behavior at 6 months. At 9 months HR infants produced significantly fewer consonants and less mature syllables than LR peers. At 12 months the HR group produces less speech and more non-speech vocal behavior than LR. HR participants who took part in the 24 month visit (n=25) were subdivided into two groups: 1) Those in whom clinicians observed some autistic symptoms (not all met full criteria for ASD), n=14; 2) Those in whom autistic symptoms were not observed, n=11. Discriminant function analysis suggested that the number of consonants produced by HR infants at each visit predicted membership in the groups based on symptom presentation at 24 months. At 6 mo., prediction is correct 74% of the time At 9 mo., prediction is correct 77% of the time At 12 mo. , prediction is correct 65% of the time Conclusions: These findings suggest that infants at high risk for ASD do show differences from low risk peers in vocal production in the first year of life. These differences are related to recognition of autistic symptoms in the second year and may be helpful in providing early identification and intervention.