Image-encoded biological and non-biological variables may be used as shortcuts in deep learning models trained on multisite neuroimaging data.

Journal: Journal of the American Medical Informatics Association : JAMIA
Published Date: Nov 17, 2023

Abstract

OBJECTIVE: This work investigates if deep learning (DL) models can classify originating site locations directly from magnetic resonance imaging (MRI) scans with and without correction for intensity differences.

Authors

  • Raissa Souza
    Department of Radiology, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada.
  • Matthias Wilms
    Department of Radiology, University of Calgary, Calgary, AB, Canada.
  • Milton Camacho
    Biomedical Engineering Program, University of Calgary, Canada; Department of Radiology, University of Calgary, Canada. Electronic address: milton.camachocamach@ucalgary.ca.
  • G Bruce Pike
    Departments of Radiology and Clinical Neurosciences, University of Calgary, Calgary, Canada.
  • Richard Camicioli
    Neuroscience and Mental Health Institute and Department of Medicine (Neurology), University of Alberta, Edmonton, Alberta, Canada.
  • Oury Monchi
    Department of Clinical Neurosciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. oury.monchi@ucalgary.ca.
  • Nils D Forkert
    Department of Radiology, University of Calgary, Calgary, Canada. nils.forkert@ucalgary.ca.

Keywords

External Resources

Popular Topics
Recent Journals