Experimental Evaluation of Deep Learning Assisted Autism Spectrum Disorder Prediction Scheme using Convolutional Classification Principle

Abstract

Autism Spectrum Disorder (ASD) is characterized by challenges in social communication and restricted or repetitive behaviors. With rising ASD incidence, early diagnosis and intervention become increasingly crucial for improving developmental trajectories and quality of life. Traditional diagnostic methods are often subjective and limited. This study evaluates a deep learning-based approach for ASD prediction using Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). Leveraging the Autism Brain Imaging Data Exchange (ABIDE) dataset, which includes neuroimaging, genetic, and behavioral data, the model integrates spatial and temporal features to enhance prediction accuracy. The combined feature set yielded the highest accuracy of 93.7%, demonstrating the efficacy of integrating multiple data types. The proposed Convolutional Recurrent Hybrid (CRH) model achieved a superior accuracy of 98%, outperforming traditional models. This research highlights the potential of deep learning techniques in improving ASD diagnostics.