Days Modeled
79
2026-01-04 to 2026-03-23
Latent States
5
KF + UKF estimation
Alert Events
22
1 in last 7 days
Post-Drug Days
8
Since 2026-03-16
Log-Likelihood
-313.43
EM-learned parameters
Latent State Trajectories
The Kalman smoother estimates 5 latent physiological states from noisy, intermittent sensor data. Solid lines show the smoothed posterior mean; shaded bands show 95% credible intervals. Dotted lines overlay the UKF estimates for comparison.
Observations vs Model Estimates
Raw sensor observations (dots) overlaid with the model's filtered estimates (lines). Good model fit is indicated by the line tracking the dots closely.
Prediction Performance
One-Step-Ahead Prediction Quality
- mean_rmssd: R-sq = 0.353, RMSE = 2.08
- mean_hr: R-sq = 0.159, RMSE = 5.97
- spo2_average: R-sq = 0.027, RMSE = 0.54
- temperature_deviation: R-sq = -0.191, RMSE = 0.27
- sleep_efficiency: R-sq = 0.020, RMSE = 3.33
Innovation Alerts (22 total, 1 in last 7 days)
2026-03-23: HRV (RMSSD) deviation = 3.3 sigma
Ruxolitinib Drug Response (started 2026-03-16)
| State | Pre-drug Mean | Post-drug Mean | Shift (SD) | Direction | p-value | Time Constant |
|---|
| Autonomic Tone | -0.083 | 0.2678 | +0.474 | improved | 0.2438 | 3.17 days |
| Cardiac Reserve | -0.4354 | 0.5381 | +1.071 | improved | 0.0398 | 5.37 days |
| Circadian Phase | 0.9591 | 0.0688 | -1.659 | stable | 0.0 | 4.77 days |
| Inflammation Level | -0.5617 | 0.0919 | +0.673 | worsened | 0.0077 | 4.89 days |
| Sleep Quality | -0.135 | 0.3732 | +0.725 | improved | 0.0 | 0.28 days |
KF vs UKF Comparison
The Unscented Kalman Filter uses nonlinear circadian dynamics and exponential autonomic decay with sigma-point propagation (no Jacobian needed). Large differences from the linear KF suggest significant nonlinear dynamics.
Multi-Modal Sensor Fusion Quality
Sensor Contribution to State Estimation
- mean_rmssd: 32.7%
- mean_hr: 22.5%
- sleep_efficiency: 16.6%
- temperature_deviation: 14.1%
- spo2_average: 14.0%
Residual Diagnostics
White noise residuals (Ljung-Box p > 0.05) indicate the model captures temporal structure well. Significant autocorrelation suggests model misspecification for that sensor.
Disclaimer: This is a computational model for research and self-monitoring purposes only. It is NOT a medical device and should NOT be used for clinical decision-making. The Oura Ring is a consumer wearable; its measurements have known limitations in accuracy. All state estimates are model-dependent and should be interpreted with appropriate uncertainty. Consult qualified healthcare professionals for medical decisions.