Comparison of EIT current electrode locations in three-dimensional head reconstruction

We compared several 16-electrode current patterns on their ability to reconstruct and quantify small amounts of bleeding. The current patterns chosen were: adjacent injection 'ring' patterns, where electrodes were placed in a single plane about the center of the phantom; and current patterns based on a standard EEG electrode layout. Within the EEG layout, we compared the performance of two 'Cz' current patterns (where all currents applied were returned via an electrode on the apex of the head) where one (Cz1) involved making all possible measurements and the other (Cz2) involved using the same current positions but only adjacent voltage measurements. Structures mimicking electrically important structures in the infant skull were included and their effects on Cz reconstructions were determined. We found that as expected, the ring electrode pattern was not able to reconstruct anomaly location well, since it could not distinguish different axial positions. The EEG based current patterns reconstructions localized anomalies well, but performed unevenly in terms of quantification, even within a central region of interest. When a structure similar to the fontanel was included underneath the apex electrode, the quality of Cz2 referred images, both in terms of reconstruction and QI stability, was somewhat better than for Cz1 or ring-based patterns.