Cargando…

Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis

Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neur...

Descripción completa

Detalles Bibliográficos
Autores principales: Storås, Andrea M., Andersen, Ole Emil, Lockhart, Sam, Thielemann, Roman, Gnesin, Filip, Thambawita, Vajira, Hicks, Steven A., Kanters, Jørgen K., Strümke, Inga, Halvorsen, Pål, Riegler, Michael A.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10378376/
https://www.ncbi.nlm.nih.gov/pubmed/37510089
http://dx.doi.org/10.3390/diagnostics13142345
_version_ 1785079750403817472
author Storås, Andrea M.
Andersen, Ole Emil
Lockhart, Sam
Thielemann, Roman
Gnesin, Filip
Thambawita, Vajira
Hicks, Steven A.
Kanters, Jørgen K.
Strümke, Inga
Halvorsen, Pål
Riegler, Michael A.
author_facet Storås, Andrea M.
Andersen, Ole Emil
Lockhart, Sam
Thielemann, Roman
Gnesin, Filip
Thambawita, Vajira
Hicks, Steven A.
Kanters, Jørgen K.
Strümke, Inga
Halvorsen, Pål
Riegler, Michael A.
author_sort Storås, Andrea M.
collection PubMed
description Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of “black box” models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes.
format Online
Article
Text
id pubmed-10378376
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-103783762023-07-29 Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis Storås, Andrea M. Andersen, Ole Emil Lockhart, Sam Thielemann, Roman Gnesin, Filip Thambawita, Vajira Hicks, Steven A. Kanters, Jørgen K. Strümke, Inga Halvorsen, Pål Riegler, Michael A. Diagnostics (Basel) Article Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of “black box” models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes. MDPI 2023-07-11 /pmc/articles/PMC10378376/ /pubmed/37510089 http://dx.doi.org/10.3390/diagnostics13142345 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Storås, Andrea M.
Andersen, Ole Emil
Lockhart, Sam
Thielemann, Roman
Gnesin, Filip
Thambawita, Vajira
Hicks, Steven A.
Kanters, Jørgen K.
Strümke, Inga
Halvorsen, Pål
Riegler, Michael A.
Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title_full Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title_fullStr Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title_full_unstemmed Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title_short Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
title_sort usefulness of heat map explanations for deep-learning-based electrocardiogram analysis
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10378376/
https://www.ncbi.nlm.nih.gov/pubmed/37510089
http://dx.doi.org/10.3390/diagnostics13142345
work_keys_str_mv AT storasandream usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT andersenoleemil usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT lockhartsam usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT thielemannroman usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT gnesinfilip usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT thambawitavajira usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT hicksstevena usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT kantersjørgenk usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT strumkeinga usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT halvorsenpal usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis
AT rieglermichaela usefulnessofheatmapexplanationsfordeeplearningbasedelectrocardiogramanalysis