Cardiac Arrhythmia Classification From Lead I ECG Recorded in a Free-Living Environment
Bibliographic record
Abstract
OBJECTIVE: Cardiac diseases are a leading cause of global mortality. Electrocardiograms (ECGs) are essential for detecting abnormal cardiac rhythms. Smartwatches can record ECGs, similar to lead I ECGs recorded by a patient vitals monitor in a hospital, potentially helping clinicians in early diagnosis and improved management of cardiovascular diseases. While AI models have classified arrhythmias with human-level accuracy, their potential for broad screening remains underutilized. METHODS: We propose a deep learning based framework for diagnosing various cardiac arrhythmias using 10-second lead I ECG recordings, demonstrating lead I's utility in remote monitoring. Robustness was tested by introducing noise to simulate real-world conditions. Additionally, a novel data similarity assessment metric was developed to enhance transfer learning and external dataset validation. RESULTS: Using over 60,000 ECGs from the PhysioNet Challenge 2021, the trained model classified clean lead I ECGs in one dataset with a test-fold area under receiver operating characteristic curve (AUC), sensitivity, and specificity of 0.915, 0.867 and 0.858 respectively. For signals with 0 dB signal-to-noise ratio from the same dataset, the respective performance metrics dropped slightly to 0.899, 0.862 and 0.818. External validation across three separate datasets showed a minimum AUC of 0.807. The data similarity metric outperformed an existing method in improving classification, particularly with limited target dataset samples, i.e. 50. CONCLUSION: The proposed Cardiac Arrhythmia Risk Evaluation from Lead-I ECG (CARE-I) framework enables accurate arrhythmia detection across diverse populations in real-world noisy environments, thus enhancing model generalisation and early diagnosis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".