Ischemic heart disease is closely related to stenosis on the coronary artery (CA), and the volume change of the left ventricle (LV) in a cardiac cycle is an important index for the evaluation of cardiac function. Coronary angiograms are usually used to identify stenotic position and its level, in clinical practice. However, a coronary angiogram also has information on heart motion in its time-series record. Usually, physicians only feel an abnormal motion from his empirical model and perform another inspection to confirm it. In this paper, we developed a method to estimate deformation and motion of the left ventricle by ellipsoidal fitting to three-dimensional shapes of the CA from time-series images of a biplanar CA angiogram. We had demonstrated this procedure against a clinical CA angiogram.

抄録全体を表示

In the information transmission mechanism of excitable cells such as neurons and cardiac cells, electrophysiological phenomena play an important role. To analyze the action potential propagation, many numerical simulations have been reported for modeling excitable cells and tissues into a complex of conductive and dielectric media. In this paper, the numerical simulation of action potential propagation in a two-dimensional model of anisotropic cardiac muscle is presented. We considered the three types of the transmembrane ionic current. Hodgkin-Huxley (HH) model, fast sodium current (FS) model and the Beeler-Reuter (BR) model. In the case of the BR model, intracellular calcium-ion density is also calculated, which is an important parameter for muscle contraction in cardiac tissue. For the two-dimensional model of cardiac tissue, we assumed that 1) muscle fibers are parallel to the x-direction and the conductivity differs in x- and y-direction, 2) the tissue is so thin that intra- and extracellular currents are parallel to the membrane surface while transmembrane current is normal to it, and 3) the extracellular-to-intracellular conductivity ratio is constant in both x and y directions. As a result, the propagation velocities in the x and y directions, v_x and v_y respectively, are found to be v_x=43.9cm/sec and v_y=31.3cm/sec in HH model, while v_x=30.1cm/sec and v_y=20.8cm/sec in BR model, which are in good agreement with experimental data. Total computation times for FS and BR models are found to be, respectively, 2 and 109 times longer than that for the HH model.

抄録全体を表示

This study describes a simplified confidence limit only for the radial distance of estimated multiple equivalentcurrent dipoles (ECDs), by which the localization accuracy is quantitatively evaluated. By applying multiple ECD localization to three kinds of actual visual event-related potentials (ERPs), this measure was compared with normalized root mean square (nRMS), and its dependence on dipole position and moment was investigated. The present results show that the confidence limit depends on the radial distance of the dipoles, that the confidence limit is a localization accuracy measure conceptually different from the nRMS and that larger dipole moment yields misleadingly higher accuracy, independently of the number of ECDs.

抄録全体を表示