What happened

The study enrolled children aged 17 to 36 months at four primary-care clinics. Tablet activities supplied behavioral measurements for a model evaluated with nested cross-validation.

Why it matters

Quantitative observations could complement caregiver questionnaires during routine visits.

How AI contributed

Computer vision measures behavior during tablet tasks; XGBoost combines digital phenotypes.

Where the evidence stands

Current stage: Research. Research published or a model demonstrated.

  1. 01ResearchResearch published or a model demonstrated.Current stage
  2. 02ValidationIndependent or real-world validation.Not documented
  3. 03Human / Field TrialClinical trial or real-world field testing.Not documented
  4. 04DeploymentClinical, industrial or public deployment.Not documented
  5. 05Measurable ImpactA measurable real-world outcome.Not documented

Research findings can change with further testing. Practical use and measurable benefit require their own evidence.

What the evidence tells us

The model was evaluated within the study cohort, not in an independent deployment. Not all children considered neurotypical received comprehensive diagnostic evaluation, and sample selection may affect performance. Screening results cannot establish a diagnosis or demonstrate improved outcomes.

Progress over time

  1. The study enrolled children aged 17 to 36 months at four primary-care clinics. Tablet activities supplied behavioral measurements for a model evaluated with nested cross-validation.

Sources & context

Nature Medicine