About this Research Topic
In recent years, there has been a surge of research focused on the application of artificial intelligence (AI) to electrocardiography (ECG). AI methods, particularly deep neural networks (DNN), have demonstrated their potential in facilitating ECG interpretation and improving patient care decision-making.
Multiple studies have explored the use of DNN for the detection of cardiac rhythm abnormalities, myocardial infarction and cardiac structural changes, demonstrating equal or superior accuracy and sensitivity compared to average cardiologist interpretations. AI algorithms trained on large labelled datasets are also expanding the dictionary of human-defined rules, enabling the identification of traits with unclear diagnostic criteria.
In this Research Topic, we welcome contributions in the form of original research, review, mini review, hypothesis and theory, perspective, that cover, but are not limited to, the following aspects:
- Database use/selection in AI models, with focus on real-world ECG recordings, data augmentation, etc.
- The optimal input format for training DNN models: signal-based ECGs, 2D ECG images, or a combination of both.
- Trust and interpretability of AI-assisted diagnoses: techniques that can provide visual insights into AI-assisted diagnoses and uncover features imperceptible to the human eye.
- The comparative utility of 12-lead ECGs, 3-lead experimental or computationally derived vectorcardiograms (VCGs), and single-lead ECG from wearables and smartwatches, as well as that of several pre-processing techniques.
- The optimal AI architecture for training and inference (i.e. Convolutional Neural Networks vs Generative Pretrained Transformers and Vision Transformers).
- The combination of AI with wearable devices to enable timely detection of ECG abnormalities during daily activities.
Ongoing research, addressing the unresolved aspects and considerations outlined above, will pave the way for the widespread adoption of AI-enhanced cardiac electrophysiology. By providing more accurate diagnoses, predicting outcomes, identifying and monitoring individuals at higher risk of cardiovascular events, and assisting therapeutic interventions, AI holds great promise for transforming the field of cardiology.
Dr. Remi Dubois declares the consulting fees for scientific expertise provided to Microport CRM. The other Topic Editors declare no conflict of interest.
Multiple studies have explored the use of DNN for the detection of cardiac rhythm abnormalities, myocardial infarction and cardiac structural changes, demonstrating equal or superior accuracy and sensitivity compared to average cardiologist interpretations. AI algorithms trained on large labelled datasets are also expanding the dictionary of human-defined rules, enabling the identification of traits with unclear diagnostic criteria.
In this Research Topic, we welcome contributions in the form of original research, review, mini review, hypothesis and theory, perspective, that cover, but are not limited to, the following aspects:
- Database use/selection in AI models, with focus on real-world ECG recordings, data augmentation, etc.
- The optimal input format for training DNN models: signal-based ECGs, 2D ECG images, or a combination of both.
- Trust and interpretability of AI-assisted diagnoses: techniques that can provide visual insights into AI-assisted diagnoses and uncover features imperceptible to the human eye.
- The comparative utility of 12-lead ECGs, 3-lead experimental or computationally derived vectorcardiograms (VCGs), and single-lead ECG from wearables and smartwatches, as well as that of several pre-processing techniques.
- The optimal AI architecture for training and inference (i.e. Convolutional Neural Networks vs Generative Pretrained Transformers and Vision Transformers).
- The combination of AI with wearable devices to enable timely detection of ECG abnormalities during daily activities.
Ongoing research, addressing the unresolved aspects and considerations outlined above, will pave the way for the widespread adoption of AI-enhanced cardiac electrophysiology. By providing more accurate diagnoses, predicting outcomes, identifying and monitoring individuals at higher risk of cardiovascular events, and assisting therapeutic interventions, AI holds great promise for transforming the field of cardiology.
Dr. Remi Dubois declares the consulting fees for scientific expertise provided to Microport CRM. The other Topic Editors declare no conflict of interest.
Keywords: 2D ECG images, vectorcardiograms, single-lead ECG, wearables, Convolutional Neural Networks, ECG abnormalities, artificial intelligence, electrocardiography, patient care, ECG, deep neural networks
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
