Modeling internal charge effects on capacitor dynamics for non-invasive estimation of membrane potential.

Accurate and non-invasive measurement of cell membrane potential is essential for studying physiological processes and disease mechanisms. In this study, we propose a conceptual and computational model for estimating membrane potential based on the electrical behavior of two series-connected capacitors, simulating a cell-attached patch-clamp configuration. We hypothesize that the presence of a net intracellular charge-representing the membrane potential-affects the charging and discharging characteristics of the capacitive circuit by introducing asymmetry in voltage distribution. To test this, we used LTSpice to simulate 202 different capacitor configurations, varying the internal potential from -100 mV to -10 mV in 10 mV increments. For each configuration, we applied voltage pulses and extracted four key current features: maximum and minimum amplitudes, and total charge and discharge durations. These features were used to train a machine learning model (XGBRegressor), which, despite the limited dataset size, demonstrated strong predictive performance (R 2 = 0.90, RMSE = 13.79 mV) in estimating the internal potential. Our findings support the hypothesis that membrane potential can be inferred from capacitive current responses in a non-invasive, cell-attached configuration. This simulation-based framework offers a promising foundation for the non-invasive estimation of membrane potential and warrants further validation in experimental systems.