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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

MicroKAN: Mapping human brain microstructure using diffusion MRI and adaptive nonlinear modeling.

NeuroImage·2026
Same author

I2I-Mamba: Multi-modal medical image synthesis via selective state space modeling.

IEEE transactions on bio-medical engineering·2026
Same author

Meta-Entity Driven Triplet Mining for Aligning Medical Vision-Language Models.

IEEE journal of biomedical and health informatics·2026
Same author

Semi-supervision for clinical contrast-weighted image synthesis from magnetic resonance fingerprinting.

Magma (New York, N.Y.)·2026
Same author

Learning Fourier-Constrained Diffusion Bridges for MRI Reconstruction.

IEEE transactions on medical imaging·2026
Same author

Editorial AI Reviewer (AIR) Trial for Responsible, Secure, and Efficient Peer Review.

IEEE transactions on medical imaging·2026

Related Experiment Video

Updated: Jul 29, 2025

Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging
11:28

Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging

Published on: June 30, 2018

11.7K

BolT: Fused window transformers for fMRI time series analysis.

Hasan A Bedel1, Irmak Sivgin1, Onat Dalmaz1

  • 1Department of Electrical and Electronics Engineering, Bilkent University, Ankara 06800, Turkey; National Magnetic Resonance Research Center (UMRAM), Bilkent University, Ankara 06800, Turkey.

Medical Image Analysis
|May 24, 2023
PubMed
Summary

BolT, a novel blood-oxygen-level-dependent transformer model, enhances functional MRI (fMRI) analysis by capturing diverse time scales. This deep-learning approach improves sensitivity and provides neuroscientifically relevant insights.

Keywords:
ClassificationConnectivityDeep learningExplainabilityFunctional MRITime seriesTransformer

More Related Videos

Optogenetic Functional MRI
06:06

Optogenetic Functional MRI

Published on: April 19, 2016

14.9K
Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation
09:36

Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation

Published on: May 12, 2014

13.9K

Related Experiment Videos

Last Updated: Jul 29, 2025

Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging
11:28

Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging

Published on: June 30, 2018

11.7K
Optogenetic Functional MRI
06:06

Optogenetic Functional MRI

Published on: April 19, 2016

14.9K
Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation
09:36

Extracting Visual Evoked Potentials from EEG Data Recorded During fMRI-guided Transcranial Magnetic Stimulation

Published on: May 12, 2014

13.9K

Area of Science:

  • Neuroscience
  • Machine Learning
  • Biomedical Imaging

Background:

  • Deep learning significantly advanced functional MRI (fMRI) data analysis.
  • Existing methods struggle with sensitivity for contextual representations across varied time scales in fMRI data.

Purpose of the Study:

  • Introduce BolT, a blood-oxygen-level-dependent transformer model, for analyzing multivariate fMRI time series.
  • Enhance the sensitivity of fMRI analysis for contextual representations across diverse time scales.

Main Methods:

  • BolT utilizes a cascade of transformer encoders with a fused window attention mechanism.
  • Temporally-overlapped windows capture local representations, while cross-window attention integrates information across time scales.
  • Progressively increasing window overlap and employing cross-window regularization transition from local to global representations and align classification features.

Main Results:

  • BolT demonstrated superior performance compared to state-of-the-art methods on large-scale public fMRI datasets.
  • Explanatory analyses identified key time points and regions crucial for model decisions.
  • Results align with established neuroscientific findings.

Conclusions:

  • BolT offers a significant advancement in analyzing fMRI time series, particularly for capturing multi-scale temporal dynamics.
  • The model's ability to identify salient neural activity patterns validates its potential for neuroscientific research.
  • BolT provides a powerful tool for understanding brain function through advanced deep learning techniques.