Early identification of children on the autism spectrum is crucial for early intervention with long-term positive effects on symptoms and skills. The need for improved objective autism detection tools is emphasized by the poor diagnostic power in current tools. Here, we aim to evaluate the classification performance of acoustic features of the voice in children with autism spectrum disorder (ASD) with respect to a heterogeneous control group (composed of neurotypical children, children with Developmental Language Disorder [DLD] and children with sensorineural hearing loss with Cochlear Implant [CI]). This retrospective diagnostic study was conducted at the Child Psychiatry Unit of Tours University Hospital (France). A total of 108 children, including 38 diagnosed with ASD (8.5 ± 0.25 years), 24 typically developing (TD; 8.2 ± 0.32 years) and 46 children with atypical development (DLD and CI; 7.9 ± 0.36 years) were enrolled in our studies. The acoustic properties of speech samples produced by children in the context of a nonword repetition task were measured. We used a Monte Carlo cross-validation with an ROC (Receiving Operator Characteristic) supervised k-Means clustering algorithm to develop a classification model that can differentially classify a child with an unknown disorder. We showed that voice acoustics classified autism diagnosis with an overall accuracy of 91% [CI95%, 90.40%-91.65%] against TD children, and of 85% [CI95%, 84.5%-86.6%] against an heterogenous group of non-autistic children. Accuracy reported here with multivariate analysis combined with Monte Carlo cross-validation is higher than in previous studies. Our findings demonstrate that easy-to-measure voice acoustic parameters could be used as a diagnostic aid tool, specific to ASD.

Purpose Phonological complexity is known to be a good index of developmental language disorder (DLD) in normal-hearing children, who have major difficulties on some complex structures. Some deaf children with cochlear implants (CIs) present a profile that evokes DLD, with persistent linguistic difficulties despite good audiological and environmental conditions. However, teasing apart what is related to auditory deficit or to language disorder remains complex. Method We compared the performance of three groups of school-age children, 33 children with CI, 22 with DLD, and 24 with typical development, on a nonword repetition (NWR) task based on phonological complexity. Children with CI were studied regarding their linguistic profile, categorized in four subgroups ranging from excellent to very poor performance. Influence of syllable length and phonological structures on the results of all the children were explored. Results The NWR task correctly distinguished children with DLD from typically developing children, and also children with CI with the poorest linguistic performance from other children with CI. However, most complex phonological structures did not reliably identify children with CI displaying a profile similar to that of children with DLD because these structures were difficult for all of the children with CI. The simplest phonological structures were better at detecting persistent language difficulties in children with CI, as they were challenging only for the children with the poorest language outcomes. Conclusions The most complex phonological structures are not good indices of language disorder in children with CI. Phonological complexity represents a gradient of difficulty that affects normal-hearing and deaf children differently.

Language-related change-detection processes are often investigated using syllables that are very simple in terms of phonological structure. However, phonological complexity is known to be challenging for young typically developing children and pathological populations. We investigated brain correlates of phonological processing and their age-related changes with a passive change-detection protocol including stimuli of varying phonological complexity, which allowed comparing responses to simple and complex phonological deviancies. Mismatch Negativity (MMN) and Late Discriminative Negativity (LDN) responses were recorded in both school-age children (n = 22) and adults (n = 24). MMN was similar for simple and complex phonological deviancy in both groups, whereas LDN appeared to be modulated by phonological complexity, albeit with different patterns according to age. In response to complex phonological change, children displayed a larger LDN response with a typical fronto-central scalp distribution, while adults showed an additional right-posterior activity but no larger amplitude than for simple change. Thus, LDN appears to be a good electrophysiological index of phonological complexity processing. This study validated the use of the LDN through this protocol for the investigation of phonological complexity processing throughout the development.