We apply the electric cell-substrate impedance sensing (ECIS) technique to monolayers of Madin-Darby canine kidney type II cells cultured on microelectrodes of different sizes. We analyze the effect of the microelectrode radius on the parameters provided by existing ECIS models. The cellular properties inferred from the models should be invariant to the change in the microelectrode radius used for the measurements, since these properties are inherent to the type of cells studied. The current standard model, the Giaever-Keese (GK) model, derived from electrical balances of a single cell extended to infinity by suitable boundary conditions, assumes an infinite microelectrode. The model is fitted to experimental data acquired with a large-radius microelectrode, which can be considered infinite for practical purposes. We compute the impedance of the other cell-covered microelectrodes from the parameters obtained with the GK model, resulting in values strongly discrepant with the experimental data for small microelectrodes. We repeat the process with the mean field (MF) model, an alternative model that depends on the microelectrode radius but not on the cell radius. In this paper we introduce the mesoscopic model, an analytical model that simultaneously includes the properties of an individual cell and the sizes of the microelectrode and the insulator (region between the microelectrode and the ground). The impedances calculated with the mesoscopic model are in excellent agreement with experimental data. Finally, the mesoscopic model reduces to the MF model when the insulator goes to infinity and to the GK model when it goes to zero.
