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 Videos

A confident decision support system for interpreting electrocardiograms.

H Holst1, M Ohlsson, C Peterson

  • 1Department of Clinical Physiology, Lund University, Sweden.

Clinical Physiology (Oxford, England)
|October 12, 1999
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

The prevalence of chiropractic-related terminology on South African chiropractors' webpages: a cross-sectional study.

Chiropractic & manual therapies·2023
Same author

Insulin facilitates corneal wound healing in the diabetic environment through the RTK-PI3K/Akt/mTOR axis in vitro.

Molecular and cellular endocrinology·2022
Same author

Induction of the integrated stress response in the rat cornea.

Experimental eye research·2021
Same author

Challenges in conducting paediatric trials with off-patent drugs.

Contemporary clinical trials communications·2021
Same author

Hippocampal subfield pathologic burden in Lewy body diseases vs. Alzheimer's disease.

Neuropathology and applied neurobiology·2020
Same author

Haematopoietic stem cells: entropic landscapes of differentiation.

Interface focus·2018

This study developed a method to estimate the error of artificial neural network (ANN) outputs for computer-aided electrocardiogram (ECG) interpretation. ANNs can reliably diagnose myocardial infarction and indicate confidence levels, improving their clinical acceptance.

Area of Science:

  • Cardiology
  • Artificial Intelligence
  • Medical Informatics

Background:

  • Computer-aided interpretation of electrocardiograms (ECGs) is common, but physician hesitancy persists due to a lack of confidence metrics in automated advice.
  • Physicians require reliable decision support systems that provide confidence estimations for computer-generated diagnostic recommendations.

Purpose of the Study:

  • To develop and validate a method for estimating the error of artificial neural network (ANN) outputs in ECG interpretation.
  • To enhance the reliability and clinical acceptance of AI-driven diagnostic tools in cardiology.

Main Methods:

  • Trained an ANN on 1249 ECGs from patients with and without anterior myocardial infarction.
  • Utilized separate datasets for training, error calculation, and performance testing of the ANN.

Related Experiment Videos

  • Employed the area under the receiver operating characteristic (ROC) curve to evaluate network performance and error estimation.

    • Main Results:

    • The ANN achieved a high overall performance (ROC AUC: 0.887).
    • ECGs with the lowest estimated errors showed near-perfect classification accuracy (ROC AUC: 0.995).
    • The developed method effectively quantifies the confidence of ANN-based diagnoses.

    Conclusions:

    • Artificial neural networks can be trained to accurately diagnose myocardial infarction from ECGs.
    • The proposed error estimation method allows ANNs to signal diagnostic confidence, increasing their potential as reliable clinical decision support systems.
    • This approach addresses physician concerns regarding automated ECG interpretation, paving the way for wider adoption of AI in clinical practice.