Journal of Cancer Research
ISSN: 2578-3726 | DOI: 10.64030/2578-3726
Suhan Nie, Dongmei Hou, Yaxuan Wang and Lidao Bao
Objective: To investigate the multifactorial mechanisms of auditory hyper-reactivity in children with Autism Spectrum Disorder (ASD) and evaluate a personalized intervention.
Methods: Thirty children with ASD and thirty typically developing (TD) children were enrolled. Using a mixed-methods design, we examined main and interactive effects of sound parameters (frequency, intensity, speed) via auditory assessments, physiological recordings, and machine learning. An 8-week personalized intervention was implemented.
Results: Auditory hyper-reactivity incidence was 28.3% in the ASD group, significantly higher than in the TD group (p<0.001). Significant main effects were found for frequency (F=25.67, p<0.001, η²=0.42), intensity (F=18.92, p<0.001, η²=0.38), and speed (F=14.56, p<0.001, η²=0.31), with significant interactions (p<0.001). Machine learning identified spectral centroid frequency (importance=15.23) and sound pressure level variance (importance=12.67) as key predictors. Post-intervention, hyper-reactivity decreased to 14.5% (p<0.001) and physiological indicators improved (P300 amplitude increased, p<0.001). The regression model explained 67% of variance (R²=0.67).
Conclusion: Auditory hyper-reactivity in ASD children is influenced by multi-parameter interactions. Machine learning-based personalized intervention can effectively improve symptoms, supporting precision rehabilitation.