#28 · Systems biology

Calibrated in silico electrophysiology models as individual digital twins

The dominant thread is parameterized in silico electrophysiology models tied to individual-level observables. Bayesian history matching personalises reaction-eikonal models of 17 hypertrophic cardiomyopathy patients from 252-electrode body surface potentials. A neural network trained on large synthetic electrophysiology datasets infers biophysical parameters and builds digital twins of excitable cells. A separate preprint models 117 transcriptomic-guided pediatric atrial cardiomyocyte populations and reports that age and heart rate jointly shape electrophysiology. Its calibration strategy is not described in the summary, so it contributes model-state heterogeneity rather than demonstrated fitting to patient data. What is new is the coupling of parameter inference to clinical or synthetic observables at cell and tissue scales.