Key result
Protocol details matched-seed ECG ensemble controls and transport diagnostics across excluded and longer SPH recordings.
Why the study?
This deposit specifies new analyses for matched-seed ECG ensemble controls and external transport diagnostics before their execution.
This prospective extension protocol specifies pre-execution analytical methods for evaluating matched-seed ECG ensemble controls and external transport diagnostics using ResNet and JEPA models.
Answers are grounded in this paper and cite their sources.
Pre-registered protocols for deep learning electrocardiogram ensembles may reduce reporting bias; leaves open external performance until validation completes.
Prospective extension protocol for matched-seed ECG ensemble controls and external transport diagnostics. The deposited bundle fixes five new training seeds per family (101, 202, 303, 404, 505), equal-budget ResNet–ResNet, JEPA–JEPA and ResNet–JEPA probability blends, scoring of 4,587 previously excluded SPH recordings, retrospective threshold and calibration diagnostics, paired patient uncertainty, and fixed-window full-waveform inference for 5,442 longer recordings. Historical SPH outcomes were already known when this protocol was prepared. This deposit specifies new analyses before their execution; it does not retrospectively preregister the original study or establish an untouched confirmatory cohort. No new outcomes from these extensions have been computed. Bundle SHA-256: 2eb3a319832673cf9aff9a758b9c9002e58c220dd025d7bfcbd4e0707b12e13c. The ZIP contains the protocol, analysis registry, data-free analysis code, synthetic tests and source recipe hashes. No ECG records, patient-level predictions or credentials are included.
No takes yet. Share an insight, caveat, or question.
Muhammad Adnan (2026) studied this question. Matched-seed ECG ensemble controls (ResNet–JEPA probability blends) and external transport diagnostics vs. Equal-budget ResNet–ResNet and JEPA–JEPA probability blends was evaluated. Matched-seed ECG ensemble controls and transport diagnostics are specified in this protocol for 4,587 excluded and 5,442 longer SPH recordings, with no new outcomes computed yet.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: