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: Mar 8, 2026

A Magnetic Resonance Imaging Protocol for Stroke Onset Time Estimation in Permanent Cerebral Ischemia
09:59

A Magnetic Resonance Imaging Protocol for Stroke Onset Time Estimation in Permanent Cerebral Ischemia

Published on: September 16, 2017

14.7K

Mixture-of-Skip-Connection Deep Learning Model to Classify Stroke Severity from Diffusion Weighted Imaging Based on

Seunghun Oh1, Hyunsu Jeong1, Yoonjae Cho1

  • 1Medical Science and Engineering, Graduate School of Artificial Intelligence (GSAI), Department of Electrical Engineering, Convergence IT Engineering, Mechanical Engineering and Medical Device Innovation Center, Pohang University of Science and Technology (POSTECH), 77 Cheongam-ro, Nam-gu, Pohang, 37673, Republic of Korea.

Journal of Imaging Informatics in Medicine
|March 6, 2026
PubMed
Summary

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

Effectiveness of Nurse-Led atrial Fibrillation Detection for Patients With Embolic Stroke of Unknown Source in Stroke Units: A Retrospective Study.

Nursing in critical care·2026
Same author

Multiscale photoacoustic imaging of stroke in preclinical models and future directions toward clinical translation [Invited].

Biomedical optics express·2026
Same author

Prognostic Implications of Poststroke Aphasia: A Multicenter Cohort Study in Tertiary Hospitals.

Journal of the American Heart Association·2026
Same author

Introduction to the feature issue Photoacoustic Imaging and Sensing: Beyond Fundamentals to Translation.

Biomedical optics express·2026
Same author

Fully integrated photoacoustic microscopy for multi-scale and longitudinal imaging in translational biomedical applications.

Biomedical optics express·2026
Same author

Cerebral vasodilation and lesion patterns in stroke-like episodes of MELAS: a quantitative angiographic investigation.

Quantitative imaging in medicine and surgery·2026

This study introduces TRACT-NET, an AI model that accurately predicts stroke severity using diffusion-weighted imaging (DWI) and National Institutes of Health Stroke Scale (NIHSS) data. TRACT-NET offers a faster, more reliable alternative to manual stroke assessments.

Area of Science:

  • Neurology
  • Artificial Intelligence
  • Medical Imaging

Background:

  • The National Institutes of Health Stroke Scale (NIHSS) is crucial for stroke management but is labor-intensive and prone to variability.
  • Accurate and rapid stroke severity assessment is vital for timely and effective treatment decisions.

Purpose of the Study:

  • To develop and validate the TRansformer And Coordinate-aTtention NETwork (TRACT-NET) for predicting stroke severity using diffusion-weighted imaging (DWI) and NIHSS scores.
  • To assess TRACT-NET's performance against existing classification models in distinguishing between minor and non-minor strokes.

Main Methods:

  • TRACT-NET integrates a mixture-residual block with 3D coordinate-attentive and self-attentive modules, enhanced by Mamba in the bottleneck stage.
  • The model was trained and validated using DWI and NIHSS data from 273 patients (AMC dataset) and externally validated on 1106 patients (SOOP dataset).
Keywords:
Attention mechanismClassificationDeep learningNIHSSStroke

More Related Videos

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.5K

Related Experiment Videos

Last Updated: Mar 8, 2026

A Magnetic Resonance Imaging Protocol for Stroke Onset Time Estimation in Permanent Cerebral Ischemia
09:59

A Magnetic Resonance Imaging Protocol for Stroke Onset Time Estimation in Permanent Cerebral Ischemia

Published on: September 16, 2017

14.7K
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.5K
  • Performance was evaluated using metrics like sensitivity, specificity, accuracy, and AUC, with gradient-weighted class activation mapping (Grad-CAM) for region analysis.

    • Main Results:

    • TRACT-NET demonstrated superior performance in binary stroke severity prediction compared to other models on both datasets.
    • On the AMC dataset, TRACT-NET achieved an accuracy of 0.8137 and an AUC of 0.8137.
    • On the SOOP dataset, TRACT-NET achieved an accuracy of 0.6896 and an AUC of 0.7094, with Grad-CAM highlighting clinically relevant DWI regions.

    Conclusions:

    • TRACT-NET shows significant potential as an automated tool for stroke severity assessment in emergency settings.
    • The model's ability to analyze DWI scans and NIHSS data can aid clinicians in making faster, more informed treatment decisions.
    • The integration of advanced AI techniques like transformers, attention mechanisms, and state-space models offers a promising direction for neuroimaging analysis.