Table 3.

Summary of features extracted from the ECG.

Feature notation and publication	Description or formula
Hjorth mobility28	Estimate of the signal’s mean frequency (Equation SI 1)
Hjorth complexity28	Estimate of the signal’s bandwidth (Equation SI 2)
Fractal dimension28	Computed using Higuchi’s algorithm61 with $k_{\mathit{max}}=17$
Shannon entropy (SE)28	Uncertainty of information content based on probability distribution (Equation SI 3)
Approximate entropy (approxEnt) *	Quantifies regularities of signal
Sample entropy (sampEnt) *	approxEnt computed without self comparisons
Multi-level sample entropy (MSE) *	Course approximations of sample entropy, computed at scales 2, 4, 6, 8

* Indicates features that, to the authors’ knowledge, have not been previously implemented for BP estimation. Full details of the algorithms and equations used for computing each feature can be found in Supplementary Information SI: 2.