Gradient Biotech
U

Menu

Research Areas

Computational BiologySingle-cell, bulk RNA-seq, spatial, and biomarkers
CardiologyCardiovascular signals, cohorts, and interpretationImmunologyImmune cell states, signaling networks, and repertoire dynamicsNeurologyBrain connectivity, functional imaging, and electrophysiologyOncologyTumor microenvironment, immune response, mutations, and survivalPathologyWhole-slide imaging, tissue quantification, and spatial pathology

Application

Documentation

Documentation

Back to studies

Cohort endpoints

Compare physiological metrics across treatment groups, identify outliers, and explore research-only signal-derived risk stratification tied to outcomes metadata.

Research question

Do HRV, blood pressure, or other cardiovascular metrics differ significantly between groups, and do signal-derived indices associate with clinical outcomes in this cohort?

Who this is for

  • Hospital research programs comparing treatment arms or diagnostic groups
  • Pharma and clinical trial teams evaluating digital physiological endpoints
  • Translational teams attaching MACE or follow-up outcomes to autonomic metrics

Data requirements

DataRequiredPurpose
Datasets with completed HRV (or BP) runsYesMetric values for group comparison
Subjects with group labelsYesTreatment group, diagnosis, sex, age
Outcomes metadataNoRisk scoring and outcome correlation
Multiple subjects per groupRecommendedMeaningful statistical comparison

Workflow

Import subjects with outcomes → Upload and link datasets → Batch HRV → Compare cohorts → Risk scores → Interpret

Step 1 — Import cohort metadata

Upload a subjects CSV including group labels and optional outcomes columns:

external_id,age,sex,treatment_group,diagnosis,mace_event,followup_days
P001,58,M,control,hypertension,0,365
P002,62,F,treated,hypertension,0,365
P003,55,M,treated,hypertension,1,180
P004,60,F,control,hypertension,0,365

Extra outcome columns are stored as flexible metadata for filtering and risk scoring.

Step 2 — Complete metric runs

Ensure HRV (or chosen metric pipeline) has completed for all compared datasets. Use Batch HRV from the cohort page if runs are pending.

Step 3 — Compare groups

Navigate to Studies → Cohort (/experiments/{id}/cohort):

  1. Choose group by — treatment_group, sex, diagnosis, or another metadata field
  2. Select HRV metric — RMSSD, SDNN, LF/HF, or other computed measure
  3. Pick statistical test
  4. Click Compare cohorts

Review:

  • Box/violin chart with per-subject data points
  • Summary table with group means, SDs, and p-values
  • Outlier panel — subjects >2 SD from group mean

Export charts as PNG or PDF for reports.

Step 4 — Research risk stratification (optional)

Run Compute risk scores at cohort level when outcomes metadata is present. Signal-derived stratification indices appear on the cohort workspace and persist on subject hub cards.

This is research exploration only — not clinical decision support. A research disclaimer is shown in the UI.

Step 5 — Drill down and interpret

  • Click outliers to open the subject hub, waveform explorer, or individual HRV run
  • Use Interpret on the cohort page for a narrative citing group statistics
  • Export methods text and run manifest from completed pipeline runs

Expected outputs

  • Group comparison chart with statistical test results
  • Per-subject metric table across the cohort
  • Outlier identification with drill-down links
  • Research risk index per subject (when risk scoring run)
  • AI cohort narrative with cited statistics
  • Exportable figures and run provenance for methods sections

Typical analyses

AnalysisComparisonQuestion
Treatment responsetreated vs. controlDid RMSSD increase after beta-blocker therapy?
Nocturnal autonomic balanceLF/HF by groupIs sympathetic dominance greater in the heart failure subgroup?
Outcome associationMACE event vs. no eventDo subjects with events show lower HRV at baseline?
BP dippingdipper vs. non-dipperDoes dipping classification track with treatment arm?

Related guides