Research on Speech Development Trajectories and Predictive Models in Children With Autism Spectrum Disorder

Overview

Recent studies indicate that children with ASD have a significantly higher risk of co-occurring speech sound disorders than typically developing children. Early atypical speech development may be a critical yet overlooked bottleneck hindering their language improvement. Given the unique phonetic features of Mandarin, it is essential to investigate speech development in Mandarin-speaking children with ASD. This study aims to construct developmental trajectories and establish early identification and prognosis prediction models for this population.

1. Procedures. This study plans to enroll 60 children with autism spectrum disorder (ASD group) and 60 typically developing children (TD group), aged 18-60 months, matched by age and sex. At baseline, both groups will undergo assessments of development, language, and parent-child interaction, as well as fNIRS; the ASD group will additionally receive clinical diagnostic evaluations. Subsequently, a one-year follow-up will be conducted, with both groups completing language and parent-child interaction assessments at 3, 6, 9, and 12 months, and undergoing repeated developmental and fNIRS evaluations at the final 12-month visit. 2. Demographic questionnaire and clinical data. The demographic questionnaire is completed by the child's primary caregiver, detailing child's name, gender, date of birth, height, weight, heart rate. Clinical data will be ascertained from the medical records, including information about DSM-5 diagnosis and comorbid conditions. 3. Sample size. This study ultimately determined a sample size of 120 participants (60 in the ASD group and 60 in the TD group). The estimation was based on the minimum sample size required to analyze the "group × time" interaction effect using a mixed-effects model (88 participants, achieving 85% power), and fully considered the following factors: ① Meeting the analytical requirements for both the main effects of group and time; ② The need to increase the sample size to 110 participants if multiple comparison corrections (Bonferroni correction, α=0.0083) are applied to the six speech indicators; ③ Considering practical operational challenges such as the long follow-up period, multiple repeated measurements, and anticipated dropout rates, the sample was ultimately expanded to 120 participants to ensure the robustness of the study. 4. Statistical analysis. SPSS 25.0 software (SPSS Inc) was used for statistical analysis for the scale data, The speech development trajectories of the two groups can be analyzed using longitudinal methods such as mixed-effects models. Repeated measures ANOVA will compare speech characteristics between the ASD and TD groups at five time points (baseline, 3, 6, 9, and 12 months). Correlation analysis (Pearson or Spearman, depending on normality) will assess the relationship between speech development and core ASD symptoms. ANOVA/MANOVA will examine group differences in speech perception tasks, and exploratory factor analysis will identify underlying factors influencing speech perception.All statistical analyses were conducted with a statistical threshold P-value of \<0.05. The data collected through fNIRS will be performed via NirSpark analysis software. 5. Ethical matters and data protection. The children participated in the study will sign the informed consent (obtained from the guardian). And this study was approved by the local ethics committee. children's name will be abbreviated and the research data will be assigned a code then to provide to the researcher. The authorization from parents on the children's health information remains valid until the study is completed. After that, researchers will delete private information from the study record.