Book Title: Multidisciplinary Research Nexus: Exploring Intersections of Knowledge
Editor: Prof. Amos R
Chapter: 13
DOI: https://doi.org/10.59646/559/13
Authors: Ambujakshi T C, Amos R
Abstract
Early identification of Autism Spectrum Disorder (ASD) enables timely intervention that can improve developmental outcomes, yet clinical diagnosis remains resource-intensive and often delayed. This paper proposes an intelligent, multimodal framework for early ASD screening that combines caregiver questionnaires, automated video and audio analysis, and optional biological signals (e.g., microbiome / wearable sensors) with machine learning models to produce a privacy-aware risk score and explainable decision support for clinicians. We review recent ML approaches to ASD screening, describe recommended datasets and pre-processing pipelines, detail the proposed system architecture (data fusion, model selection, and explain ability), and present an evaluation protocol. We also discuss ethical, regulatory, and deployment considerations. We ground our recommendations in recent literature and public datasets, and provide 12 up-to-date references.