Pre- and Paralinguistic Vocal Production in ASD: Birth Through School Age
Purpose of Review
We review what is known about how pre-linguistic vocal differences in autism spectrum disorder (ASD) unfold across development and consider whether vocalization features can serve as useful diagnostic indicators.
Differences in the frequency and acoustic quality of several vocalization types (e.g., babbles and cries) during the first year of life are associated with later ASD diagnosis. Paralinguistic features (e.g., prosody) measured during early and middle childhood can accurately classify current ASD diagnosis using cross-validated machine learning approaches.
Pre-linguistic vocalization differences in infants are promising behavioral markers of later ASD diagnosis. In older children, paralinguistic features hold promise as diagnostic indicators as well as clinical targets. Future research efforts should focus on (1) bridging the gap between basic research and practical implementations of early vocalization-based risk assessment tools, and (2) demonstrating the clinical impact of targeting atypical vocalization features during social skill interventions for older children.
KeywordsAutism Paralinguistics Prosody Early diagnosis Speech production Acoustic properties
This work supported by the Autism Science Foundation ASF #19-006 (grantee: Yankowitz), and NIDCD R03DC017944, “Infant Vocalizations as Early Markers of Autism Spectrum Disorder” (PI: Parish-Morris).
Compliance with Ethical Standards
Conflict of Interest
The authors declare that they have no conflicts of interest.
Human and Animal Rights and Informed Consent
This article does not contain any studies with human or animal subjects performed by any of the authors.
Papers of particular interest, published recently, have been highlighted as: • Of importance •• Of major importance
- 2.Kanner L. Autistic disturbances of affective contact. Neuro Child. 1943;2:217–50.Google Scholar
- 5.•• Fusaroli R, Lambrechts A, Bang D, Bowler DM, Gaigg SB. Is voice a marker for autism spectrum disorder? A systematic review and meta-analysis. Autism Res. 2017;10:384–407. This systematic review and meta-analysis demonstrates significant differences in pitch and pitch range, and calls for systematic study of multivariate prediction based on high accuracy in studies conducted so far. CrossRefGoogle Scholar
- 9.• Redford MA, Kapatsinski V, Cornell-Fabiano J. Lay Listener Classification and Evaluation of Typical and Atypical Children’s Speech. Lang Speech. 2018;61:277–302. This study demonstrates that lay listeners are sensitive to speech features in ASD, which contribute to perceptions of disorder and likeability. CrossRefGoogle Scholar
- 10.• Patel SP, Kim JH, Larson CR, Losh M. Mechanisms of voice control related to prosody in autism spectrum disorder and first-degree relatives. Autism Res. 2019;12(8):1192–210. This study suggests that atypical audio-vocal integration may be a genetically-based mechanism of prosodic differences in ASD. CrossRefPubMedPubMedCentralGoogle Scholar
- 11.Marchi E, Schuller BW, Baron-Cohen S, Golan O, Bölte S, Arora P, et al. Typicality and emotion in the voice of children with autism spectrum condition: evidence across three languages. INTERSPEECH. 2015.Google Scholar
- 12.Asgari M, Bayestehtashk A, Shafran I. Robust and Accurate Features for Detecting and Diagnosing Autism Spectrum Disorders. Lyon: Proc Interspeech; 2013.Google Scholar
- 14.Santos JF, Brosh N, Falk TH, Zwaigenbaum L, Bryson SE, Roberts W, et al. Very early detection of Autism Spectrum Disorders based on acoustic analysis of pre-verbal vocalizations of 18-month old toddlers. 2013 IEEE Int Conf Acoust Speech Signal Process. 2013. p. 7567–71.Google Scholar
- 19.Developmental Disabilities Monitoring Network Surveillance Year 2010 Principal Investigators, Centers for Disease Control and Prevention (CDC). Prevalence of autism spectrum disorder among children aged 8 years - autism and developmental disabilities monitoring network, 11 sites, United States, 2010. Morb Mortal Wkly Rep Surveill Summ Wash DC 2002. 2014;63:1–21.Google Scholar
- 23.• Roche L, Zhang D, Bartl-Pokorny KD, Pokorny FB, Schuller BW, Esposito G, et al. Early Vocal Development in Autism Spectrum Disorder, Rett Syndrome, and Fragile X Syndrome: Insights from Studies Using Retrospective Video Analysis. Ther Adv Neurol Disord. 2018;2:49–61. This article reviews studies of infant vocalizations gathered through retrospective home videos. CrossRefGoogle Scholar
- 29.Sullivan K, Sharda M, Greenson J, Dawson G, Singh NC. A novel method for assessing the development of speech motor function in toddlers with autism spectrum disorders. Front Integr Neurosci. 2013;7.Google Scholar
- 31.Northrup JB, Iverson JM. Vocal Coordination During Early Parent-Infant Interactions Predicts Language Outcome in Infant Siblings of Children with Autism Spectrum Disorder. Infancy Off J Int Soc Infant Stud. 2015;20:523–47.Google Scholar
- 32.•• Pokorny FB, Schuller B, Marschik PB, Brueckner R, Nyström P, Cummins N, et al. Earlier Identification of Children with Autism Spectrum Disorder: An Automatic Vocalisation-Based Approach. ISCA. 2017:309–13. This small study demonstrates that machine learning applied to a standardized acoustic feature set from vocalizations collected at 10 months can accurately predict diagnostic outcome at age 3. Google Scholar
- 39.Chericoni N, de Brito WD, Costanzo V, Diniz-Gonçalves A, Leitgel Gille M, Parlato E, et al. Pre-linguistic Vocal Trajectories at 6–18 Months of Age As Early Markers of Autism. Front Psychol. 2016;7.Google Scholar
- 45.Schoen E, Paul R, Chawarska K. Vocal productions in toddlers with autism spectrum disorders. Speech Sound Disord Child San Diego Plur Publ Inc. 2009:181–204.Google Scholar
- 53.•• McDaniel J, D’Ambrose Slaboch K, Yoder P. A meta-analysis of the association between vocalizations and expressive language in children with autism spectrum disorder. Res Dev Disabil. 2018;72:202–13. This meta-analysis demonstrates that early vocalizations are predictive of expressive language ability in ASD, suggesting they may be both a marker and an intervention target. CrossRefGoogle Scholar
- 54.•• Esposito G, Hiroi N, Scattoni ML. Cry, Baby, Cry: Expression of Distress As a Biomarker and Modulator in Autism Spectrum Disorder. Int J Neuropsychopharmacol. 2017;20:498–503. This article reviews a line of evidence from humans and mouse models suggesting atypical cry in ASD, which elicits altered response from caregivers. CrossRefPubMedPubMedCentralGoogle Scholar
- 66.•• English MS, Tenenbaum EJ, Levine TP, Lester BM, Sheinkopf SJ. Perception of Cry Characteristics in 1-Month-Old Infants Later Diagnosed with Autism Spectrum Disorder. J Autism Dev Disord. 2019;49:834–44. This small study suggests that cry may be an extremely early-emerging bio-behavioral marker of ASD. Cries of 1-month-old infants later diagnosed with ASD were rated as more distressing, atypical, and indicative of pain than cries of TD infants. CrossRefGoogle Scholar
- 73.Lyakso E, Frolova O, Grigorev A. A Comparison of Acoustic Features of Speech of Typically Developing Children and Children with Autism Spectrum Disorders. In: Ronzhin A, Potapova R, Németh G, editors. Speech Comput. Springer International Publishing; 2016. p. 43–50.Google Scholar
- 77.DePape A-MR, Chen A, Hall GB, Trainor LJ. Use of Prosody and Information Structure in High Functioning Adults with Autism in Relation to Language Ability. Front Psychol. 2012;3.Google Scholar
- 78.Lyakso E, Frolova O, Grigorev A. Perception and Acoustic Features of Speech of Children with Autism Spectrum Disorders. In: Karpov A, Potapova R, Mporas I, editors. Speech Comput. Springer International Publishing; 2017. p. 602–12.Google Scholar
- 83.Shriberg L, Kwiatkowski J, Rasmussen CR, Lof GL, Miller J. The Prosody-Voice Screening Profile (PVSP): Psychometric Data and Reference Information For Children Phonology Project Technical Report No. 1. 1997.Google Scholar
- 84.Tanaka H, Sakti S, Neubig G, Toda T, Nakamura S. Linguistic and Acoustic Features for Automatic Identification of Autism Spectrum Disorders in Children’s Narrative. In: Proc Workshop Comput Linguist Clin Psychol Linguist Signal Clin Real. Baltimore: Association for Computational Linguistics; 2014. p. 88–96.CrossRefGoogle Scholar
- 85.Kakihara Y, Takiguchi T, Ariki Y, Nakai Y, Takada S. Investigation of Classification Using Pitch Features for Children with Autism Spectrum Disorders and Typically Developing Children. Am J Signal Process. 2015;5:1–5.Google Scholar
- 86.Schuller B, Steidl S, Batliner A, Vinciarelli A, Scherer K, Ringeval F, et al. The INTERSPEECH 2013 computational paralinguistics challenge: social signals, conflict, emotion, autism. 2013.Google Scholar
- 87.Baird A, Amiriparian S, Cummins N, Alcorn AM, Batliner A, Pugachevskiy S, et al. Automatic Classification of Autistic Child Vocalisations: A Novel Database and Results. Interspeech 2017. ISCA; 2017. p. 849–53.Google Scholar
- 88.Constantino J, Gruber C. The Social Responsiveness Scale Manual, Second Edition (SRS-2). Los Angeles: Western Psychological Services; 2012.Google Scholar
- 89.Ringeval F, Marchi E, Grossard C, Xavier J, Chetouani M, Cohen D, et al. Automatic Analysis of Typical and Atypical Encoding of Spontaneous Emotion in the Voice of Children. Proc INTERSPEECH 2016 17th Annu Conf Int Speech Commun Assoc ISCA. San Francisco, CA, United States; 2016. p. 1210–4.Google Scholar
- 90.American Psychiatric Association, American Psychiatric Association, editor. Diagnostic and statistical manual of mental disorders: DSM-5. 5th ed. Washington: American Psychiatric Association; 2013.Google Scholar
- 91.Deng J, Cummins N, Schmitt M, Qian K, Ringeval F, Schuller B. Speech-based Diagnosis of Autism Spectrum Condition by Generative Adversarial Network Representations. Londres: 7th Int Digit Health Conf; 2017. p. 53–7.Google Scholar
- 92.• Schmitt M, Marchi E, Ringeval F, Schuller B. Towards Cross-lingual Automatic Diagnosis of Autism Spectrum Condition in Children’s Voices. Speech Commun 12 ITG Symp. 2016. p. 1–5. This machine learning classification study finds that across several languages, acoustic features can accurately classify ASD diagnosis when the model is trained on only speakers of the same language. However, model performance dramatically decreases when trained on speakers of a different language. Google Scholar
- 93.Cho S, Liberman M, Ryant N, Cola M, Schultz RT, Parish-Morris J. Automatic detection of Autism Spectrum Disorder in children using acoustic and text features from brief natural conversations. Proc Interspeech. Graz, Austria; 2019.Google Scholar
- 94.• Nakai Y, Takiguchi T, Matsui G, Yamaoka N, Takada S. Detecting Abnormal Word Utterances in Children With Autism Spectrum Disorders: Machine-Learning-Based Voice Analysis Versus Speech Therapists. Percept Mot Skills. 2017;124:961–73. This study represents a first step toward evaluating how automatic speech-based diagnostic classifiers perform compared to professionals. Machine learning performed slightly better than speech language pathologists, who were provided with an extremely limited dataset (single word utterances). CrossRefGoogle Scholar
- 95.Xu D, Yapanel U. Gray S. LENA Found: Reliability of the LENA Language Environment Analysis System in young children’s natural home environment; 2009.Google Scholar