Answers are grounded in this paper and cite their sources.
Pupillometry data hold promises for assessing the severity of autism spectrum disorder, which could help inform more effective interventions for affected children. This study proposes extracting the correlation dimension as a nonlinear feature from pupillometry data to regress the severity. We employed a publicly available dataset of 29 affected children, with severity levels labeled according to the Childhood Autism Rating Scale score. We trained machine learning models to predict the labeled scores using the correlation dimension of pupil diameter as the sole predictive feature. The best model (support vector machine) achievedQ2= 0.59 in nested leave-one-out cross-validation, explaining 59% of the variance in held-out subjects. These results support the correlation dimension of pupil diameter as a promising marker for the severity of autism spectrum disorder, with nonlinear models outperforming linear ones in this dataset.
No takes yet. Share an insight, caveat, or question.
Ma et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: