A numerical study on the equivalence of complete and shunt electrode models for transfer impedance in electrical impedance tomography

The objective of this study is to numerically compare the Complete Electrode Model (CEM)-widely regarded as the most general and physically realistic approach in Electrical Impedance Tomography (EIT) forward problem solutions-with the shunt electrode model, specifically in terms of transfer impedance (Z_t). The analyses were performed on a two-dimensional heterogeneous thorax-like geometry. Sixteen surface electrodes were utilized, and adjacent and opposite current injection patterns were applied separately. Within the scope of CEM, the contact impedance (z_c) was varied within the range of 10⁻⁴-1 Ω·m² while all other parameters were kept constant. The validity of the forward problem findings was tested through the addition of measurement noise (40-60 dB SNR) and image reconstruction studies. The results demonstrated that while z_c plays a decisive role at the injection electrodes where current is directly applied, both models produced practically identical results at the transfer impedance level under all examined conditions. Even under challenging noise conditions, the difference between the transfer impedances produced by the two models remained below 3%, and this equivalence was further proven at the inverse problem level by achieving a Pearson correlation of 0.989 for the adjacent pattern and 0.993 for the opposite pattern in the image reconstruction analyses. These findings quantitatively reveal that in wearable and real-time EIT systems utilizing high-input-impedance measurement chains, the shunt electrode model provides a computationally more efficient alternative without causing any significant loss of accuracy during the image formation stage.