Current source reconstructing and magnetic imaging of cardiac electrical activity during P-wave

Abstract

<sec> The current source reconstruction and magnetic imaging is a new technique to non-invasively obtain spatial information regarding cardiac electrical activity using magnetocardiogram (MCG) signals measured by the superconducting quantum interference device (SQUID) on the human thorax surface. Using MCG signals to reconstruct distributed current sources needs to solve the inverse problem of magnetic field. The beamforming is a type of spatial filter method that has been used for distributed source reconstruction and source imaging in electroencephalogram (EEG) and magnetoencephalogram (MEG). In this paper, the dipole moment of distributed current source is estimated with corresponding each spatial filter based on the cardiac source field model. The purpose is to enhance the intensity contrast of the dipole moment of distributed current sources in distributed source spatial spectrum estimation with beamforming, so that the reconstructed-pseudo sources beyond the heart can be removed for imaging cardiac electric activity well.</sec><sec> A new beamforming method of improving intensity contrast (IIC) of distributed source spatial spectrum estimation is developed for imaging cardiac electric activity in P-wave, due to cardiac magnetic signals in P-wave lower than that of the peak value of R-wave, which has a relatively low signal-to-noise ratio (SNR). For enhancing the accuracy of current source reconstruction in P-wave, the IIC divided into two steps: firstly, to introduce the lead-field matrix, which represents the measurement sensor-array sensitivity to magnetic field current sources, into a weight matrix of the spatial filter for making the output estimation of the filter more sensitive to the current sourcedistribution, so as to improve the intensity contrast of the reconstructed distributed sources.Secondly, by setting a threshold of source intensity from experience, to extract the reconstructed source with locally-maximal dipole strength at each time for eliminating the relatively weak pseudo sources in other locations, so as to enhance the accuracy of current source reconstruction during P-wave. In this paper, the IIC and three other methods, including minimum variance beamforming (MVB), suppressing spatial filter output noise-power gain (SONG) and trust region reflective (TRR), are compared by using the theoretical analysis and simulation experiments of MCG current source reconstruction during P-wave. The results show that the IIC has higher intensity contrast of the single source spatial spectrum estimation, and possesses better accuracy of the current source reconstruction. The 61-channel MCG signals of two healthy subjects and their imaging of cardiac electrical activity during P-wave also are analyzed. The result shows that the IIC is better than the other three methods. It is indicated that two healthy subjects have stronger electrical activity in the atrium than that in the ventricle at Ppeak time, also that the electrical activity has the direction feature when the right-atrium is depolarized during P-wave. </sec><sec> In summary, the IIC is useful for imaging the cardiac electrical activity. However, it is needed to carry out a further research on patients with local myocardial ischemia and left or right coronary artery stenosis, and to establish the evaluation index for imaging of cardiac electrical activity in such patients.</sec>