Skip to main content
Scientific Reports logoLink to Scientific Reports
. 2025 Sep 29;15:33322. doi: 10.1038/s41598-025-19777-y

Correction: Transformer-based deep learning ensemble framework predicts autism spectrum disorder using health administrative and birth registry data

Kevin Dick 1,2,3,✉, Emily Kaczmarek 3, Robin Ducharme 4, Alexa C Bowie 4, Alysha L J Dingwall-Harvey 1,3,4, Heather Howley 1,2,3, Steven Hawken 3,4,5,6,7, Mark C Walker 1,3,4,5,6,8,9,10, Christine M Armour 1,2,3,10,11,✉
PMCID: PMC12480944  PMID: 41023127

Correction to: Scientific Reports 10.1038/s41598-025-90216-8, published online 07 April 2025

The original version of this Article contained an error in Figure 5, where the “Labour Type” variable was attributed to the dataset NSO instead of the BORN BIS dataset. The original Figure 5 and accompanying legend appear below.

Fig. 5.

Fig. 5

Summary of the top-ranking risk factors determined using SHAP analysis across the three independent datasets. The right-most SHAP summary plot depicts the violin plot distribution for each factor. The possible feature values for each variable are tabulated in Supplementary Table S3; individual KDE plots comparing ASD and controls are illustrated in Supplementary Figure S2. Both the C2 and C16 screen for acetylcarnitine analytes.

The original Article has been corrected.

Contributor Information

Kevin Dick, Email: kdick@bornontario.ca.

Christine M. Armour, Email: carmour@cheo.on.ca


Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES