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: Aug 9, 2025

A Data-Driven Approach to Quantifying Immune States in Sepsis
07:42

A Data-Driven Approach to Quantifying Immune States in Sepsis

Published on: February 7, 2025

291

Exploring a global interpretation mechanism for deep learning networks when predicting sepsis.

Ethan A T Strickler1, Joshua Thomas2, Johnson P Thomas3

  • 1Physics and Mathematics, East Central University, PO Box 385, Ada, OK, 74820, USA.

Scientific Reports
|February 22, 2023
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

Interpretable Framework for Sleep Monitoring: Applying Statistical Control Charts to Physiological Data Streams.

Sensors (Basel, Switzerland)·2026
Same author

Tumor infiltrating lymphocytes (TILs) as a predictive marker of pathological complete response (pCR) in a diverse patient population with early triple negative breast cancer (TNBC) treated with neoadjuvant real-world KEYNOTE-522 regimen.

Breast cancer research and treatment·2026
Same author

Safety and Feasibility of Transcatheter Aortic Valve Replacement in Patients With Hematologic Neoplasms.

JACC. Case reports·2026
Same author

Laboratories of Harm Reduction and Treatment: Determinants of States' Policy Adoptions to Address the Opioid Epidemic.

Journal of health politics, policy and law·2026
Same author

Form Follows Function: A New Paradigm of Pancreatic Cancer Progression.

bioRxiv : the preprint server for biology·2025
Same author

Drug screen reveals new potent host-targeted antivirals against Mpox virus.

Research square·2025

This study developed a novel interpretation mechanism to identify clinical features for sepsis detection using machine learning models. The mechanism highlighted 17 features, aiding early sepsis detection and clinical decision support.

Area of Science:

  • Biomedical Informatics
  • Artificial Intelligence in Medicine
  • Critical Care Medicine

Background:

  • Sepsis detection remains a challenge, necessitating improved clinical decision support systems.
  • Black-box machine learning models offer potential for sepsis detection but lack interpretability.
  • Identifying key clinical features is crucial for accurate and timely sepsis diagnosis.

Purpose of the Study:

  • To introduce and evaluate a novel mechanism for interpreting black-box machine learning models for sepsis detection.
  • To identify additional clinical features contributing to sepsis detection.
  • To compare model-identified features with expert, clinical, and literature-based features.

Main Methods:

  • Utilized the 2019 PhysioNet Challenge dataset with ~40,000 ICU patients and 40 physiological variables.

More Related Videos

A Reproducible Intensive Care Unit-Oriented Endotoxin Model in Rats
05:56

A Reproducible Intensive Care Unit-Oriented Endotoxin Model in Rats

Published on: February 20, 2021

2.1K
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
09:47

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

1.2K

Related Experiment Videos

Last Updated: Aug 9, 2025

A Data-Driven Approach to Quantifying Immune States in Sepsis
07:42

A Data-Driven Approach to Quantifying Immune States in Sepsis

Published on: February 7, 2025

291
A Reproducible Intensive Care Unit-Oriented Endotoxin Model in Rats
05:56

A Reproducible Intensive Care Unit-Oriented Endotoxin Model in Rats

Published on: February 20, 2021

2.1K
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
09:47

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

1.2K
  • Employed a Long Short-Term Memory (LSTM) network as the representative black-box model.
  • Adapted the Multi-set Classifier for global interpretation and compared results against Random Forest, clinical, academic, and statistical features.

    • Main Results:

    • Identified 17 features used by the LSTM for sepsis classification.
    • Found overlap between LSTM features and Random Forest (11 features), academic (10 features), and clinical (5 features).
    • Highlighted age, chloride ion concentration, pH, and oxygen saturation as potentially important for sepsis development.

    Conclusions:

    • Interpretation mechanisms can enhance the integration of machine learning into clinical decision support for early sepsis detection.
    • The study identified key features and suggests further investigation into novel clinical indicators for sepsis.
    • Further research into interpretation mechanisms and underutilized clinical features is warranted.