{
  "id": 7840017,
  "title": "Automatic Quality Control and Error Correction in MRI linear registration via a Residual Parameter Prediction Network for T1w MRI",
  "url": "https://urgent.news/2026/09/16/automatic-quality-control-and-error-correction-in-mri-linear",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-16T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.10.750693v1?rss=1"
  },
  "original_language": "en",
  "account": "Errors in linear registration of MRI scans can have a ripple effect, impacting downstream tasks such as nonlinear registration, volumetric estimations, deformation-based morphometry, and voxel-based morphometry analyses. Due to their subtle nature and difficulty in detection, these errors are particularly challenging to identify and correct. To address this issue, researchers have developed the Residual Affine COefficient Optimization Network (RACOON), a novel framework aimed at detecting and correcting linear registration errors specifically in T1-weighted MRI scans registered to the MNI-ICBM152 standard space.\n\nRACOON employs a dual-module approach, with the correction module and the classification module. The correction module focuses on minimizing residual misalignment, quantified by the Root Mean Square Error (RMSE) of 0.778 mm on a synthetic dataset. This achievement is particularly noteworthy as it is comparable to the variability observed in repeated quality control (QC)-passed registrations using the same pipeline. The classification module, on the other hand, employs a residual parameter prediction network to identify potential registration errors with impressive accuracy. It achieved a balanced accuracy of 76.8% and a precision of 74.4%, surpassing existing state-of-the-art methods.\n\nThe RACOON framework is designed to be open source and publicly accessible, allowing researchers and practitioners to leverage its capabilities in their own work. The implementation and source code can be found at https://github.com/ZhaojinChen/RACOON, facilitating further research and development in the field of MRI linear registration and error correction.",
  "summary": "Errors in linear registration can propagate to downstream nonlinear registration and bias volumetric estimations, deformation-based morphometry (DBM) and voxel-based morphometry (VBM) analyses. Subtle linear registration errors are particularly challenging as they are difficult to detect and may not result in obvious failures in nonlinear registration but still affect downstream results.…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "Nature",
        "title": "Quality control of glycogen through direct ubiquitylation by RNF213",
        "url": "https://urgent.news/2026/09/15/quality-control-of-glycogen-through-direct-ubiquitylation-by-rnf213",
        "published": "2026-09-15T00:00:00.000Z"
      }
    ]
  },
  "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."
}