Every analysis step you run — from a button in the Dataset inspector, Bulk Pipelines, or Cohort comparison — maps to one of these pipelines. This reference lists what each one takes as input and produces as output, so you know what to run and in what order.
ECG
| Pipeline | Input | Outputs |
|---|
preprocess_ecg | WFDB, CSV, or EDF waveform | R-peaks, RR intervals, SQI, artifact rate, waveform preview |
compute_hrv | Preprocess run, RR artifact, or RR CSV upload | Time/frequency/nonlinear HRV plus advanced metrics (SDANN, DFA, entropy, fragmentation, PRSA AC/DC) and selectable PSD method (Welch/Lomb/multitaper/Burg) |
full_ecg_analysis | WFDB, CSV, or EDF waveform | Chained preprocess + HRV in one run |
delineate_ecg | ECG waveform (+ optional preprocess run) | P/Q/R/S/T fiducials; PR, QRS, QT/QTc, ST, T-wave morphology |
compute_arrhythmia_burden | RR intervals from preprocess | Irregularity %, pause burden, beat flags |
windowed_hrv | Preprocess run, RR artifact, or RR CSV | Sliding-window HRV, quality/missingness, trends, feature matrix |
Wearable modalities and motion
| Pipeline | Input | Outputs |
|---|
analyze_eda | EDA CSV | Tonic (SCL) / phasic decomposition, skin conductance responses (SCRs) |
analyze_acc | ACC CSV (time + x/y/z or magnitude) | ENMO activity intensity, motion burden, per-window activity |
motion_quality | ECG preprocess run + ACC run | Motion burden %, % of beats during motion, high-motion intervals |
RR-only
| Pipeline | Input | Outputs |
|---|
compute_hrv | RR CSV (format=rr_csv) | HRV metrics without waveform preprocess |
PPG
| Pipeline | Input | Outputs |
|---|
preprocess_ppg | PPG waveform | Pulse peaks, IBI intervals, quality, preview |
compute_prv | PPG preprocess run | Pulse rate variability metrics |
analyze_ppg_morphology | PPG waveform (+ optional preprocess run) | Per-pulse amplitude, rise/decay time, width, area, dicrotic notch |
compute_ptt | ECG preprocess run + PPG run | Pulse arrival/transit time (R-peak → PPG foot/peak), per-beat PAT/PTT |
Blood pressure
| Pipeline | Input | Outputs |
|---|
analyze_bp | BP CSV | SBP/DBP/MAP means, variability, dipping classification |
analyze_abp | ABP waveform CSV (signal_type=abp) | Beat-level SBP/DBP/MAP, pulse pressure, instantaneous HR, max upstroke dP/dt |
Respiration and coupling
| Pipeline | Input | Outputs |
|---|
analyze_respiration | Respiration channel | Rate, breath timing, quality |
compute_rsa | ECG preprocess + respiration run | Respiratory sinus arrhythmia metrics |
compute_baroreflex | HRV/RR run + BP run | Baroreflex sensitivity — sequence method + cross-spectral transfer function (TF gain, coherence, phase, latency) |
Cohort
| Pipeline | Input | Outputs |
|---|
cohort_summary | Study subjects with completed HRV | Group means, SDs, p-values, plot series |
cohort_windowed_summary | Subjects with completed windowed HRV | Group comparison on per-subject medians and across-window CV |
compute_risk_scores | Cohort with outcomes metadata | Research-only signal-derived stratification |
Analyze → Bulk Pipelines runs any of these across every eligible dataset in the study in one action, instead of one dataset at a time.
Interpretation and reporting
| Pipeline | Input | Outputs |
|---|
interpret_results | Run ID(s) or cohort context | Descriptive or mechanistic narrative JSON — powers Interpret everywhere it appears |
A study-level report bundle with an auto-generated methods section (generate_report) is not yet available from the UI.
Job lifecycle
Every run moves through the same states, shown as a status badge on the run detail page and in the Jobs panel:
pending → running → completed | failed
Versioning
Every run records a pipeline version, visible on the run detail page's Summary tab. Include run IDs and versions in methods sections for reproducibility.