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

STNAGNN: Data-driven Spatio-temporal Brain Connectivity beyond FC.

Proceedings of machine learning research·2026
Same author

HYPERBOLIC MODEL AGGREGATION FOR FEDERATED LEARNING IN FMRI.

Proceedings. IEEE International Symposium on Biomedical Imaging·2026
Same author

Integrated transcriptomic and metabolomic analyses elucidate the molecular mechanisms underlying quality formation during Rosa acicularis fruit ripening.

Food research international (Ottawa, Ont.)·2026
Same author

Geometry-Guided Local Alignment for Multi-View Visual Language Pre-Training in Mammography.

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

Quantifying Gas-Phase Crosstalk and SiO/Gr Reactivity Competition Governing Thermal Runaway in Composite-Anode Batteries.

Advanced materials (Deerfield Beach, Fla.)·2026
Same author

Dose-aware diffusion model for 3D PET image denoising: Multi-institutional validation with reader study and real low-dose data.

Medical image analysis·2026

Related Experiment Video

Updated: May 17, 2025

Probing the Brain in Autism Using fMRI and Diffusion Tensor Imaging
12:21

Probing the Brain in Autism Using fMRI and Diffusion Tensor Imaging

Published on: September 12, 2011

25.0K

Self-Supervised Pre-training Tasks for an fMRI Time-series Transformer in Autism Detection.

Yinchi Zhou1, Peiyu Duan1, Yuexi Du1

  • 1Department of Biomedical Engineering, Yale University, New Haven, CT, USA.

Machine Learning in Clinical Neuroimaging : 7Th International Workshop, MLCN 2024, Held in Conjunction with MICCAI 2024, Marrakesh, Morocco, October 10, 2024, Proceedings. MLCN (Workshop) (7Th : 2024 : Marrakesh, Morocco)
|March 31, 2025
PubMed
Summary

This study introduces a novel transformer-based framework for Autism Spectrum Disorder (ASD) detection using resting-state fMRI data. Self-supervised pre-training with ROI masking significantly improved classification accuracy, outperforming models trained from scratch.

Keywords:
AutismSelf-supervised LearningTransformerfMRI

More Related Videos

Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time
07:12

Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time

Published on: July 1, 2014

12.2K
Application of Granger Causality Analysis of the Directed Functional Connection in Alzheimer's Disease and Mild Cognitive Impairment
08:43

Application of Granger Causality Analysis of the Directed Functional Connection in Alzheimer's Disease and Mild Cognitive Impairment

Published on: August 7, 2017

7.8K

Related Experiment Videos

Last Updated: May 17, 2025

Probing the Brain in Autism Using fMRI and Diffusion Tensor Imaging
12:21

Probing the Brain in Autism Using fMRI and Diffusion Tensor Imaging

Published on: September 12, 2011

25.0K
Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time
07:12

Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time

Published on: July 1, 2014

12.2K
Application of Granger Causality Analysis of the Directed Functional Connection in Alzheimer's Disease and Mild Cognitive Impairment
08:43

Application of Granger Causality Analysis of the Directed Functional Connection in Alzheimer's Disease and Mild Cognitive Impairment

Published on: August 7, 2017

7.8K

Area of Science:

  • Neuroscience
  • Machine Learning
  • Medical Imaging

Background:

  • Autism Spectrum Disorder (ASD) presents diagnostic challenges due to symptom variability.
  • Resting-state functional magnetic resonance imaging (rs-fMRI) is a key tool for studying brain activity in ASD.
  • Transformer models show promise for analyzing complex sequential data like fMRI.

Purpose of the Study:

  • To develop and evaluate a transformer-based self-supervised learning framework for ASD classification using rs-fMRI data.
  • To investigate the efficacy of different masking strategies in self-supervised pre-training for enhancing model performance.
  • To address overfitting issues common in small neuroimaging datasets.

Main Methods:

  • A transformer-based self-supervised framework was developed to directly analyze rs-fMRI time-series data.
  • Self-supervised pre-training involved reconstructing randomly masked fMRI data, with strategies including masking entire ROIs or time points.
  • The pre-trained model was fine-tuned for ASD classification and validated on two public datasets using cross-validation.

Main Results:

  • Masking entire ROIs during self-supervised pre-training yielded superior performance compared to masking time points.
  • This ROI masking strategy resulted in an average improvement of 10.8% in AUC and 9.3% in subject accuracy.
  • The proposed framework demonstrated improved performance across varying amounts of training data compared to training from scratch.

Conclusions:

  • Transformer-based self-supervised learning, particularly with ROI masking, is effective for ASD classification from rs-fMRI.
  • The developed framework offers a promising approach to improve diagnostic accuracy for ASD.
  • The study provides a valuable, openly available resource for the research community.