Deep Learning for Neurodevelopmental Disorder Diagnosis: Leveraging Retinal Images with Self-supervised Learning

Abstract

Neurodevelopmental disorders like autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) have posed substantial challenges to the cognitive, social, and emotional development of children. Early and precise diagnosis of ASD and ADHD holds significant importance in facilitating timely interventions and support, ultimately enhancing outcomes for individuals affected by these conditions. Conventional diagnostic approaches for these disorders have depended on behavioral assessments, clinical observations, and structured interviews, which are labor-intensive and susceptible to subjective judgments made by healthcare professionals. In this research paper, we proposed a deep learning-based system for diagnosing neurodevelopmental disorders. The proposed system utilizes retinal images as input and generates disorder predictions. We introduce a novel self-supervised learning approach to attain superior accuracy outcomes. We demonstrated that the proposed system achieves state-of-the-art accuracy on large-scale retinal image datasets with extensive experimental results.