WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis
IntroductionRepresentation learning allows artificial intelligence (AI) models to learn useful features from large, unlabelled datasets. This can reduce the need for labelled data across a range of downstream tasks. It was hypothesised that wave segmentation would be a useful form of electrocardiogr...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-03-01
|
| Series: | Frontiers in Physiology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fphys.2022.760000/full |
| _version_ | 1828876635185610752 |
|---|---|
| author | Rob Brisk Rob Brisk Raymond R. Bond Dewar Finlay James A. D. McLaughlin Alicja J. Piadlo Alicja J. Piadlo David J. McEneaney David J. McEneaney |
| author_facet | Rob Brisk Rob Brisk Raymond R. Bond Dewar Finlay James A. D. McLaughlin Alicja J. Piadlo Alicja J. Piadlo David J. McEneaney David J. McEneaney |
| author_sort | Rob Brisk |
| collection | DOAJ |
| description | IntroductionRepresentation learning allows artificial intelligence (AI) models to learn useful features from large, unlabelled datasets. This can reduce the need for labelled data across a range of downstream tasks. It was hypothesised that wave segmentation would be a useful form of electrocardiogram (ECG) representation learning. In addition to reducing labelled data requirements, segmentation masks may provide a mechanism for explainable AI. This study details the development and evaluation of a Wave Segmentation Pretraining (WaSP) application.Materials and MethodsPretraining: A non-AI-based ECG signal and image simulator was developed to generate ECGs and wave segmentation masks. U-Net models were trained to segment waves from synthetic ECGs. Dataset: The raw sample files from the PTB-XL dataset were downloaded. Each ECG was also plotted into an image. Fine-tuning and evaluation: A hold-out approach was used with a 60:20:20 training/validation/test set split. The encoder portions of the U-Net models were fine-tuned to classify PTB-XL ECGs for two tasks: sinus rhythm (SR) vs atrial fibrillation (AF), and myocardial infarction (MI) vs normal ECGs. The fine-tuning was repeated without pretraining. Results were compared. Explainable AI: an example pipeline combining AI-derived segmentation masks and a rule-based AF detector was developed and evaluated.ResultsWaSP consistently improved model performance on downstream tasks for both ECG signals and images. The difference between non-pretrained models and models pretrained for wave segmentation was particularly marked for ECG image analysis. A selection of segmentation masks are shown. An AF detection algorithm comprising both AI and rule-based components performed less well than end-to-end AI models but its outputs are proposed to be highly explainable. An example output is shown.ConclusionWaSP using synthetic data and labels allows AI models to learn useful features for downstream ECG analysis with real-world data. Segmentation masks provide an intermediate output that may facilitate confidence calibration in the context of end-to-end AI. It is possible to combine AI-derived segmentation masks and rule-based diagnostic classifiers for explainable ECG analysis. |
| first_indexed | 2024-12-13T08:26:08Z |
| format | Article |
| id | doaj.art-a92145f2a92848feafeb10676157d5f7 |
| institution | Directory Open Access Journal |
| issn | 1664-042X |
| language | English |
| last_indexed | 2024-12-13T08:26:08Z |
| publishDate | 2022-03-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Physiology |
| spelling | doaj.art-a92145f2a92848feafeb10676157d5f72022-12-21T23:53:54ZengFrontiers Media S.A.Frontiers in Physiology1664-042X2022-03-011310.3389/fphys.2022.760000760000WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram AnalysisRob Brisk0Rob Brisk1Raymond R. Bond2Dewar Finlay3James A. D. McLaughlin4Alicja J. Piadlo5Alicja J. Piadlo6David J. McEneaney7David J. McEneaney8Faculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomCardiology Department, Craigavon Area Hospital, Craigavon, United KingdomFaculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomFaculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomFaculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomFaculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomCardiology Department, Craigavon Area Hospital, Craigavon, United KingdomFaculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, United KingdomCardiology Department, Craigavon Area Hospital, Craigavon, United KingdomIntroductionRepresentation learning allows artificial intelligence (AI) models to learn useful features from large, unlabelled datasets. This can reduce the need for labelled data across a range of downstream tasks. It was hypothesised that wave segmentation would be a useful form of electrocardiogram (ECG) representation learning. In addition to reducing labelled data requirements, segmentation masks may provide a mechanism for explainable AI. This study details the development and evaluation of a Wave Segmentation Pretraining (WaSP) application.Materials and MethodsPretraining: A non-AI-based ECG signal and image simulator was developed to generate ECGs and wave segmentation masks. U-Net models were trained to segment waves from synthetic ECGs. Dataset: The raw sample files from the PTB-XL dataset were downloaded. Each ECG was also plotted into an image. Fine-tuning and evaluation: A hold-out approach was used with a 60:20:20 training/validation/test set split. The encoder portions of the U-Net models were fine-tuned to classify PTB-XL ECGs for two tasks: sinus rhythm (SR) vs atrial fibrillation (AF), and myocardial infarction (MI) vs normal ECGs. The fine-tuning was repeated without pretraining. Results were compared. Explainable AI: an example pipeline combining AI-derived segmentation masks and a rule-based AF detector was developed and evaluated.ResultsWaSP consistently improved model performance on downstream tasks for both ECG signals and images. The difference between non-pretrained models and models pretrained for wave segmentation was particularly marked for ECG image analysis. A selection of segmentation masks are shown. An AF detection algorithm comprising both AI and rule-based components performed less well than end-to-end AI models but its outputs are proposed to be highly explainable. An example output is shown.ConclusionWaSP using synthetic data and labels allows AI models to learn useful features for downstream ECG analysis with real-world data. Segmentation masks provide an intermediate output that may facilitate confidence calibration in the context of end-to-end AI. It is possible to combine AI-derived segmentation masks and rule-based diagnostic classifiers for explainable ECG analysis.https://www.frontiersin.org/articles/10.3389/fphys.2022.760000/fullartificial intelligenceelectrocardiogram (ECG)machine learningexplainable AIrepresentation learning |
| spellingShingle | Rob Brisk Rob Brisk Raymond R. Bond Dewar Finlay James A. D. McLaughlin Alicja J. Piadlo Alicja J. Piadlo David J. McEneaney David J. McEneaney WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis Frontiers in Physiology artificial intelligence electrocardiogram (ECG) machine learning explainable AI representation learning |
| title | WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis |
| title_full | WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis |
| title_fullStr | WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis |
| title_full_unstemmed | WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis |
| title_short | WaSP-ECG: A Wave Segmentation Pretraining Toolkit for Electrocardiogram Analysis |
| title_sort | wasp ecg a wave segmentation pretraining toolkit for electrocardiogram analysis |
| topic | artificial intelligence electrocardiogram (ECG) machine learning explainable AI representation learning |
| url | https://www.frontiersin.org/articles/10.3389/fphys.2022.760000/full |
| work_keys_str_mv | AT robbrisk waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT robbrisk waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT raymondrbond waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT dewarfinlay waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT jamesadmclaughlin waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT alicjajpiadlo waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT alicjajpiadlo waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT davidjmceneaney waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis AT davidjmceneaney waspecgawavesegmentationpretrainingtoolkitforelectrocardiogramanalysis |