{
  "id": 6803083,
  "title": "Quantifying Uncertainty in Brain Age Predictions via Conformal Prediction",
  "url": "https://urgent.news/2026/09/11/quantifying-uncertainty-in-brain-age-predictions-via-conformal",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-11T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.04.749442v1?rss=1"
  },
  "original_language": "en",
  "account": "Youth brain development can manifest as either delayed or accelerated neurodevelopment, discernible through brain magnetic resonance imaging (MRI). Brain age prediction aims to estimate an individual's brain age through machine learning models trained on healthy individuals. The brain age gap (BAG) represents the discrepancy between an individual's brain age and chronological age, and it has been investigated as a potential biomarker for various conditions. However, BAG faces limitations such as reliance on chronological age, regression to the mean, and interpretability issues.\n\nTo address these limitations, we propose brain age intervals (BAIs) as an alternative framework. BAIs represent a normative interval surrounding an individual's chronological age, derived from MRI measures. We employ a statistical framework called conformal prediction, which generates prediction intervals with guaranteed coverage. To validate the effectiveness of BAIs, we trained multiple brain age prediction models on structural and functional MRI scans from the Reproducible Brain Charts dataset, which includes individuals aged 6 to 22. We estimated BAIs for each participant across 100 repeated train-test splits.\n\nThe resulting BAIs exhibited stable empirical coverage closely aligned with the nominal 90% level, with a median coverage of 91%, an interval width of 7.36 years, and a root mean square error (RMSE) of 2.20 years. Furthermore, we explored potential associations between interval coverage and clinical measures, such as the p-factor (a composite score reflecting cognitive abilities) and parental education. However, we found only weak associations, with a modest directional signal in parental education. These findings underscore the viability of uncertainty-aware brain age modeling in youth populations while emphasizing the need for larger, more diverse samples to uncover significant clinical and environmental correlations.",
  "summary": "Youth can exhibit signs of delayed or accelerated neurodevelopment, measurable via brain magnetic resonance imaging (MRI). Brain age prediction seeks to estimate brain age at an individual level using machine learning models fitted on healthy individuals. The brain age gap (BAG): the difference between brain age and chronological age, has been studied as a potential biomarker. However, BAG…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}