Age-stratified multimodal MRI and machine learning to explore autism-related brain characteristics in youth - Takeaways - MDSpire

Age-stratified multimodal MRI and machine learning to explore autism-related brain characteristics in youth

  • By

  • Garazi Casillas Martinez

  • Anthony Winder

  • Kimberly Amador

  • Eneko Uruñuela

  • Matthias Wilms

  • Sarah J. MacEachern

  • Nils D. Forkert

  • July 2, 2026

  • 0 min

Share

  • 1

    The study investigates machine learning models using multimodal neuroimaging data to identify brain characteristics associated with autism.

  • 2

    A total of 144 participants aged 5 to 18 years provided structural MRI, diffusion MRI, and resting-state functional MRI data for analysis.

  • 3

    Classification accuracies for unimodal models ranged from 66.3% to 77.6%, with diffusion MRI showing the highest performance in younger children.

  • 4

    The multimodal approach improved classification performance, achieving accuracies of 78.9%, 76.7%, and 70.5% across different age cohorts.

  • 5

    The study highlights the importance of age-specific analyses in understanding autism-related brain features and suggests future applications for other neurodevelopmental conditions.

Original Source(s)

Age-stratified multimodal MRI and machine learning to explore autism-related brain characteristics in youth

  • by Garazi Casillas Martinez, Anthony Winder, Kimberly Amador, Eneko Uruñuela, Matthias Wilms, Sarah J. MacEachern, Nils D. Forkert

  • July 2, 2026

Related Content

Electroencephalographic abnormalities and clinical phenotypes in children with autism spectrum disorder: a single center cohort study

Diversity of words and words for diversity

Autism Diagnosis Is Expanding — At a Cost

Autism may be overdiagnosed — and, Liao and Fombonne argue, that children with the greatest needs may be paying the price.