Modeling the phenotypic architecture of autism symptoms from time of diagnosis to age 6.

The team fed every ADI-R item into a latent-class model. They wanted to see if autism symptoms stay the same from diagnosis until age six.

Kids were already diagnosed with ASD. No new treatment was tested. The math simply looked for hidden groups that share symptom paths.

Two big symptom factors stayed steady for almost everyone. A tiny group—about one in eleven—broke the pattern.

By age six that small group showed a clear drop in core symptoms. The model flags who may take a milder road.

Moss et al. (2008) saw the same age window and also found symptom scores can dip. Their work set the stage; Stelios added a model that spots the improvers up front.

Bitsika et al. (2018) argued autism is just one severity line. Stelios agrees most kids stay on that line, yet the model shows a real off-ramp for a few.

YWilson et al. (2023) tracked adaptive skills far past age six and still found turning points around kindergarten. Stelios gives an earlier signal you can watch.

You can run the same factor model after your intake ADI-R. If a child lands in the “decreasing” class, plan brief re-evaluations and lean into naturalistic teaching. Share the finding with families so they know early scores are not set in stone.