Modeling of Non-invasive Cell Membrane Potential Measurement: LTSpice Simulation and Machine Learning Analysis

Abstract

AbstractThis paper presents a novel simulation approach using electric pulses to measure cell membrane potential. The primary objective was to develop a non-invasive method for accurately assessing membrane potential without altering the cell membrane or its internal components. Traditional techniques, such as voltage-sensitive dyes, often require cell incubation, which can affect membrane properties and reduce measurement accuracy. To overcome these limitations, we modeled the cell as two series-connected capacitors with a negative potential between them. By manipulating this negative potential and analyzing the charge and discharge characteristics, simulations conducted with LTspice demonstrated the feasibility of predicting membrane potential based on these characteristics.We tested 202 groups with various capacitor combinations and measured four key parameters related to charge and discharge currents: maximum current, minimum current, total charge time, and total discharge time. Using the XGBRegressor model, we achieved a strong fit with an R2score of 0.9. This indicates a robust correlation between the measured charge and discharge characteristics and the internal potential of the cell. Our findings suggest that accurate measurement of cell membrane potential is possible by attaching an electrode to the cell without disrupting membrane integrity. Thus, this simulation-based approach offers a promising and non-invasive alternative for measuring cell membrane potential.