Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Apr 22, 2026

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
07:11

Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

Published on: December 8, 2023

2.6K

A universal and efficient method to compute maps from image-based prediction models.

Mert R Sabuncu

    Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
    |October 17, 2014
    PubMed
    Summary
    This summary is machine-generated.

    Related Concept Videos

    You might also read

    Related Articles

    Articles linked to this work by shared authors, journal, and citation graph.

    Sort by
    Same author

    A view-engage-predict framework for enhancing brain-behavior mapping with naturalistic movie-watching fMRI.

    Communications biology·2026
    Same author

    BPD-Neo: An MRI Dataset for Lung-Trachea Segmentation with Clinical Data for Neonatal Bronchopulmonary Dysplasia.

    Scientific data·2026
    Same author

    Generating Novel Brain Morphology by Deforming Learned Templates.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same author

    AtlasMorph: Learning conditional deformable templates for brain MRI.

    Medical image analysis·2026
    Same author

    Author's Reply: Unrecognized Biases and Validation Gaps in TraceOrg for Automated Autosomal Dominant Polycystic Kidney Disease Volumetry.

    Journal of the American Society of Nephrology : JASN·2026
    Same author

    Knockout: A simple way to handle missing inputs.

    Transactions on machine learning research·2026
    Same journal

    LiftReg: Limited Angle 2D/3D Deformable Registration.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same journal

    Inverse Consistency by Construction for Multistep Deep Registration.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same journal

    Can Crowdsourced Annotations Improve AI-based Congestion Scoring For Bedside Lung Ultrasound?

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same journal

    Equivariant Filters for Efficient Tracking in 3D Imaging.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same journal

    Lobar Lung Density Embeddings with a Transformer encoder (LobTe) to predict emphysema progression in COPD.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    Same journal

    uniGradICON: A Foundation Model for Medical Image Registration.

    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention·2026
    See all related articles

    This study introduces a new framework to interpret complex machine learning models in biomedical imaging. It helps identify brain regions associated with cognitive impairment predictions without needing ground truth labels.

    Area of Science:

    • Biomedical Image Computing
    • Machine Learning
    • Neuroimaging Analysis

    Background:

    • Supervised learning algorithms like Support Vector Machines are widely used in biomedical image computing for prediction models.
    • Interpreting complex machine learning models, especially kernel-based methods, remains a significant challenge for biological insights.
    • Existing methods for deriving predictive region maps often rely on strong assumptions about the model or data, or fail to use covariance structures.

    Purpose of the Study:

    • To propose a computationally efficient and universal framework for quantifying associations in black box machine learning models.
    • To demonstrate the informativeness of examining associations with predictions, even without ground truth labels.
    • To apply the framework to neuroimaging data for predicting cognitive impairment.

    More Related Videos

    Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
    14:08

    Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

    Published on: April 13, 2013

    43.7K

    Related Experiment Videos

    Last Updated: Apr 22, 2026

    Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping
    07:11

    Author Spotlight: Insights into Visual Cortex Research Through Wide-View fMRI Mapping

    Published on: December 8, 2023

    2.6K
    Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
    14:08

    Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

    Published on: April 13, 2013

    43.7K

    Main Methods:

    • Developed a novel, computationally efficient framework for analyzing black box machine learning models.
    • Utilized a theoretical perspective to show the value of prediction associations without ground truth labels.
    • Applied the method to structural neuroimaging data and machine learning models predicting cognitive impairment.

    Main Results:

    • The proposed framework effectively quantifies associations captured by complex machine learning models.
    • Analysis without ground truth labels provided valuable insights into model predictions.
    • The method generated biologically meaningful association maps from neuroimaging data.

    Conclusions:

    • The developed framework offers a universal and efficient approach to interpreting machine learning models in biomedical imaging.
    • The findings highlight the utility of analyzing prediction associations for understanding model behavior.
    • The approach successfully identified relevant brain regions associated with cognitive impairment prediction in neuroimaging studies.