A Multimodal AI Framework for Early Autism Identification Using CbRNN–Ensemble Modelling and Advanced NCIAPPF‐Based Temporal Encoding
International Journal of Developmental Neuroscience · 2026
DOI: 10.1002/jdn.70117 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Developmental Neuroscience, 2026. Verfasst von Indhu Bonthala und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science and Engineering, Sreenidhi Institute of Science and Technology, Yamnampet, Telangana, India. Hauptthemen laut MEDLINE: Autism Spectrum Disorder, Artificial Intelligence, Autistic Disorder. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 35 Literaturangaben.
Abstract (Originalsprache)
To improve child developmental results and enable prompt interventions, early identification of autism spectrum disorder (ASD) is crucial. Conventional diagnostic instruments, while clinically standardized, remain constrained by subjectivity, limited scalability and dependence on expert evaluation. This study introduces AutiScan , an AI‐driven framework that integrates a Convolution‐based Recurrent Neural Network Ensemble Model (CbRNN‐EM) with the neural context‐integrated adaptive reprocessing and pattern filtering (NCIAPPF) technique for capturing temporal dynamics of neural encoding. The proposed system analyses multimodal data including facial expressions, eye‐gaze patterns, speech prosody and behavioural cues to extract spatial, emotional and sequential features relevant to early ASD markers. NCIAPPF enhances preprocessing by adaptively filtering noise, normalizing multimodal signals and preserving fine‐grained temporal dependencies essential for robust pattern recognition. The fused representations are classified using an ensemble layer, achieving superior accuracy compared with traditional deep learning and machine learning prototypes. The results of the experiments show that AutiScan offers a dependable, scalable and comprehensible early ASD screening tool, underscoring the opportunity for practical clinical and educational implementation.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Computer Science and Engineering, Sreenidhi Institute of Science and Technology, Yamnampet, Telangana, India.
ORCID-Profile: Shaik Khasim Saheb
Zitieren
APA
Bonthala, I., Chadubudhi, S., Linga, S., & Saheb, S. K. (2026). A Multimodal AI Framework for Early Autism Identification Using CbRNN–Ensemble Modelling and Advanced NCIAPPF‐Based Temporal Encoding. International Journal of Developmental Neuroscience. https://doi.org/10.1002/jdn.70117
Vancouver
Bonthala I, Chadubudhi S, Linga S, Saheb SK. A Multimodal AI Framework for Early Autism Identification Using CbRNN–Ensemble Modelling and Advanced NCIAPPF‐Based Temporal Encoding. International Journal of Developmental Neuroscience. 2026. doi:10.1002/jdn.70117
BibTeX
@article{Bonthala2026,
author = {Bonthala, Indhu and Chadubudhi, Sravanthi and Linga, Sreeja and Saheb, Shaik Khasim},
title = {A Multimodal AI Framework for Early Autism Identification Using CbRNN–Ensemble Modelling and Advanced NCIAPPF‐Based Temporal Encoding},
journal = {International Journal of Developmental Neuroscience},
year = {2026},
doi = {10.1002/jdn.70117}
}
EUVIMED