Key result
Federated learning yields ~3-point lower arrhythmia accuracy compared with centralized training.
Why the study?
Training diagnostic ECG models on pooled patient data raises sensitivity concerns, while federated learning on battery-powered wearables forces a compromise among diagnostic accuracy, formal privacy guarantees, and communication cost.
Does privacy-preserving federated learning maintain diagnostic classification accuracy while reducing communication overhead for wearable arrhythmia detection compared to centralized training on the MIT-BIH Arrhythmia Database?
Does privacy-preserving federated learning maintain diagnostic classification accuracy while reducing communication overhead for wearable arrhythmia detection compared to centralized training on the MIT-BIH Arrhythmia Database?
In a simulation of wearable ECG arrhythmia detection, federated learning reduces uplink communication volume by 44.9-fold and tolerates client label skew with modest accuracy loss, but naive per-round unclipped differential privacy severely degrades classification performance.
Answers are grounded in this paper and cite their sources.
Simulation supports federated learning feasibility for arrhythmia detection with reduced communication; leaves open whether differential privacy can preserve utility for clinical deployment.
Background Wearable devices make continuous electrocardiogram (ECG) monitoring routine, but training diagnostic models on pooled patient data sits uneasily with how sensitive that data is. Federated learning avoids pooling, yet on battery-powered wearables it forces a compromise among diagnostic accuracy, formal privacy guarantees and communication cost. This study measures that compromise for a lightweight arrhythmia classifier with 4709 parameters, trained by federated averaging across ten simulated clients on the Massachusetts Institute of Technology–Beth Israel Hospital (MIT-BIH) Arrhythmia Database. Three factors were varied: communication volume relative to centralized training, the severity of non-independent and identically distributed (non-IID) label skew across clients, and the differential privacy (DP) budget ε. Results The centralized baseline reached 98.3% accuracy. Federated training without privacy noise reached 95.0% ± 1.1% (mean ± standard deviation over five seeds) while sending 44.9 times fewer uplink bytes than a single transfer of the raw training data at 64-bit precision. Accuracy fell from 95.8% to 89.9% as the Dirichlet concentration α dropped from 1.0 to 0.1, a loss of 5.9 percentage points under extreme label skew. Differential privacy behaved very differently: mean accuracy ranged from 58.4% to 79.2% across ε values from 0.1 to 10, with standard deviations of up to 27.5 percentage points. Accuracy peaked at ε = 2 and did not rise at ε = 5 or ε = 10, and no setting came close to the 95.0% obtained without noise. Conclusions For a model of this size, communication cost and client heterogeneity are manageable, whereas the privacy mechanism is the limiting factor. The non-monotonic, high-variance privacy results point to the per-round, unclipped noise injection used here and not to a fixed ceiling on private federated arrhythmia detection. A clipped, budget-accounted mechanism should be tested before any clinical deployment.
No takes yet. Share an insight, caveat, or question.
Dhruv Shah (2026) studied Cardiac arrhythmia (n=48). Federated learning arrhythmia classifier (4,709-parameter multilayer perceptron with Federated Averaging and differential privacy) vs. Centralized training baseline was evaluated on Arrhythmia classification accuracy (MD -3.3 percentage points). Federated learning without privacy noise achieved 95.0% arrhythmia accuracy versus 98.3% centralized with 44.9-fold less uplink data, while differential privacy dropped accuracy to 58.4%-79.2%.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: