IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. BME-29, NO. 10, OCTOBER 1982 (c) (d) (e) (t) Fig. 7. (Continued.) Flow marker histograms of subepicardial and subendocardial regions in a normal dog heart. The sub- epicardial region outlined in the upper left portion of each image produces (c) a Thioflavin histogram and (d) an auto- radiograph histogram, while the adjacent outlined subendocardial region yields (e) another Thioflavin histogram and (f) another autoradiograph histogram. The histograms are representative of those that were obtained also for the other regions. For this analysis, the amplitudes were logarithmically scaled into a range centered at 140 with a standard devia- tion equal to 16. 14 143 12 5 10 6 10 UVAR Count Fig. 8. Regional averages of flow marker. Numbers proportional to Thioflavin fluorescence, autoradiograph density, and counted volume radioactivity are plotted for each of four epicardial regions (1, 3, 5, and 7) and four endocardial regions (2, 4, 6, and 8) outlined in the flow marker images shown in Fig. 7. (AR = autoradiographic analysis, UV = Thioflavin fluorescence, Count = gamma counter determination in counts/min/g of wet tissue 10-3, a = standard deviation). images which have different detail. This appears to merit further investigation and repeated trials of our computer-based approach to obtain an improved understanding of myocardial blood flow. REFERENCES [1] R. J. Domenech, J.I.E. Hoffman, M.I.M. Noble, K. B. Saunders, J. R. Henson, and S. Subijanto, "Total and regional coronary blood flow measured by radioactive microspheres in conscious and anesthetized dogs," Circ. Res., vol. 25, pp. 581-596, 1967. [2] L. C. Becker, N. J. Fortuin, and B. Pitt, "Effect of ischemia and antianginal drugs on the distribution of radioactive microspheres in the canine left ventricle," Circ. Res., vol. 28, pp. 263-269, 1971. [3] T. Yipintsoi, W. A. Dobbs, Jr., P. D. Scanlon, T. J. Knopp, and J. B. Bassingthwaighte, "Regional distribution of diffusible tracers and carbonized microspheres in the left ventricles of isolated dog ,hearts," Circ. Res., vol. 33, pp. 573-587, 1973. [4] M. L. Marcus, R. E. Kerber, J. Ehrhardt, and F. M. Abbound, "Three dimensional geometry of acutely ischemic myocardium," Circ. Res., vol. 52, pp. 254-263, 1975. [5] P. M. Malsky, P. S. Vokonas, S. J. Paul, S. L. Robbins, and W. B. Hood, Jr., "Autoradiographic measurement of regional blood flow in normal and ischemic myocardium," Amer. J. Physiol., vol. 232, pp. H576-H583, 1977. [6] L. M. Partlow, L. G. Bush, L. J. Stensaas, and R. P. Kesner, "A novel technique for quantitative autoradiography of labeled histological specimens," J. Histochem. Cytochem., vol. 29, pp. 79-83, 1981. [71 R. A. Kloner, C. E. Ganote, and R. B. Jennings, "The 'no reflow' phenomenon after temporary coronary occlusion in the dog," J. Clin. Invest., vol. 54, pp. 1496-1508, 1974. [8] W. K. Pratt, Digital Image Processing. New York: Wiley, 1978. [9] R. B. Blackman and J. W. Tukey, The Measurement of Power Spectra from the Point of View of Communication Engineering. New York: Dover, 1959, p. 98. A Simulation Study of the Single Moving Dipole Representation of Cardiac Electrical Activity PIERRE SAVARD, GUY E. MAILLOUX, FERNAND A. ROBERGE, RAMESH M. GULRAJANI, AND ROBERT GUARDO Abstract-A simulation study of the accuracy of the single moving dipole (SMD) cardiac representation was done using a realistic com- puter model of the human torso. This model included regions of different electrical conductivities such as the lungs and the intraven- tricular blood masses. Dipolar potential distributions generated on the surface of this model were sampled with different electrode arrange- ments and were subsequently contaminated with random noise. The location, magnitude, and orientation of the SMD which optimally fitted these potentials were then compared to those of the original dipole source in order to assess the errors in the SMD parameters. The SMD computations were based on a least squares estimation of a variable number of multipolar components recovered inside a torso model which was either finite and homogeneous or finite with lungs. The results showed that the SMD accuracy can be significantly improved by adding higher order multipolar components, e.g., with 120 electrodes and the absence of noise, the rms position error was 15.5 mm with eight multi- poles and 0.6 mm with 24 multipoles when the same homogeneous model was used for both the simulation and the recovery. It was also shown that the addition of lungs in the recovery model permitted a higher accuracy when either the lungs, or the lungs with blood masses, were present when the dipolar potential distributions were generated; Manuscript received October 15, 1981; revised March 4, 1982. This work was supported by the Medical Research Council of Canada, the Canadian and Quebec Heart Foundations, and the Conseil de Recherche en Sant6 du Qu6bec. The authors are with the Institut de G6nie Biomedical, Ecole Poly- technique et Universit6 de Montr6al, Montreal, P.Q., Canada H3C 3T8. 0018-9294/82/1000-0700$00.75 ( 1982 IEEE 150 a 140 130 120 ............ *UV 0- - - -OAR ~ Count 2 3 4 5 6 7 8 Area # 700