Effect of number of electrodes, electrode displacement, and RMS measurement noise on the localization accuracy of ECG inverse problem

The effect of the number of electrodes, electrode displacement and RMS measurement noise was evaluated using an anatomically-detailed computer model of the thorax as a volume conductor. The body surface potential distributions due to cardiac dipole sources were calculated by applying five different electrode montages: the eight electrodes representing the independent leads of the standard 12-lead electrocardiogram (ECG), a modified 24-lead configuration, a Lux 32-lead full-body configuration, a Montreal 64-lead configuration, and a Brussels 120-lead configuration. Inverse solutions were computed using the lead field concept in the presence of both erroneous locations of the electrodes and of RMS measurement noise added to the torso surface potentials. The results indicate that increasing the number of leads enhances the localization accuracy of the inverse problem. With 32 or more electrodes, the localization accuracy remained stabilized despite the added RMS measurement noise. Similarly, increasing the number of displaced electrodes to 32 improved the localization accuracy as compared to cases with fewer electrodes.