Leo G. Horan

Surface reflections of cardiac excitation and the assessment of infarct volume in dogs. A comparison of methods.

Ventricular depolarization was analyzed in intact dogs by simultaneously recording body surface potential maps, McFee axial vectorcardiograms, and a 5 × 4 lead precordial grid of QRS complexes. The purpose of this study was to compare the effectiveness of subtraction approaches, using the simultaneously acquired data. The totally closed chest approach avoided the problem of volume conductor alteration by thoracotomy. Infarct volume was calculated morpho-logically from measurements of serial ventricular sections. The maximal correlation with anatomic infarct size using the precordial QRS grid approach was 0.51, using cumulative difference data between 1 and 38 msec when the postinfarction grid was subtracted from the preinfarction grid. A correlation coefficient of 0.80 was achieved using the numerically integrated data between 1 and 31 msec from the vectorcardiogram, and the body surface potential map achieved a correlation coefficient above 0.88 when the electrical difference of msec 16 was used. These data suggest that estimates of infarct size from selected surface reflections of the activation process are feasible if some sort of preinfarction control data are available. Caution must be exercised to avoid inclusion of electrical effects late in the activation process which contain contamination by highly variable alterations in the excitation sequence due to delayed conduction or alteration in conduction pathway in or near the infarct zone.

Correlative analysis of vectorcardiograms and serial instantaneous surface potential maps in normal young men.

Abstract This study compared and analyzed the similarities and disparities between the instantaneous body surface potential maps (BSMs) and vectorcardiograms (VCGs) obtained from 30 healthy young men. Changes in the VCG from that of the classic normal usually consisted of scalloping of the smooth arch in the VCG loops and usually occurred in inferior and posterior loops after 35 msec. Changes in the BSM that differed from what we considered the normal were usually double positive peaks or double negative valleys. There was good correlation between the BSM and the VCG as long as the heart acted as a single dipole. With multiple dipolar activity of the heart, the BSM usually showed double sinks or peaks of voltage representing the more complex heart activity. During this time the VCG usually showed scalloping. We concluded that much of the multipolar activity of the heart centers around the various times of electrical breakthrough to the epicardium in the normal sequence of depolarization of the heart. We feel that this study has started a basis to help clarify the relationships between inflections and indentations of the VCG and the complex surface activity in detailed potential maps which will be important in distinguishing these normal changes from the more common notching and scalloping with known heart disease.

Comparative surface potential patterns in obstructive and nonobstructive cardiomyopathy

Abstract Extensive noninvasive electrophysiologic studies have been performed on patients with cardiomyopathy. Body surface potential maps in the nonobstructive patients mimicked the normal in the general progression of the surface manifestations of activation, but demonstrated relatively more prominent patterns of electrical breakthrough in terms of absolute magnitude and in terms of the perimeter of negativity at a given instant in time. The pattern of the obstructive group was distinctly different, being marked by increased magnitude and duration of positivity with an obvious delay in septal, right ventricular, and left ventricular breakthrough contiguous to the septum, with attenuated negativity at the time of those break-throughs. Maximum negativity also usually occurred 15 to 20 msec. later in the patients with IHSS. The patients with nonobstructive disease characteristically had posteriorly and equatorially oriented VCG loops and often lacked evidence of normal early left-to-right activation of the interventricular septum. Loops of the patients with obstructive disease were usually oriented below the horizontal and were less posteriorly oriented. Additional abnormalities noted in some patients with obstructive disease included slowing of the loop through at least 24 msec. The most extreme form of abnormal sequence of ventricular activation was seen in one third of the patients with IHSS in the form of false infarction patterns. An electrophysiologic rationale of this latter abnormality has been suggested by van Dam and Durrer and is thought by us to represent the best current explanation.

Relation between the ventriculographic silhouette and topography of thoracic potential in coronary artery disease

The body surface potential map obtained within 30 days of cardiac catheterization was examined in 180 patients with coronary artery disease. Radii to the systolic and diastolic boundaries of the right anterior oblique ventriculogram were measured at 18 degrees intervals; isointegral voltages were tabulated for early and late halves of the QRS complex at 35 definitive electrode sites. Multivariate analysis showed all ray lengths depended on all 70 voltage values. Linear transformation matrices to predict ray length from voltage distribution were calculated for a training set which was successively expanded from 80 to 160 at increments of 20 patients. Training set expansion led to a progressive decrease in the error of reproduction of the ray lengths for patients outside the training set. There is a strong relation between ventriculographic contours in patients with coronary artery disease and body surface potential values during early and late QRS complexes. Even in simplified linear formulation, the relation is detectable throughout a large population despite interindividual variations in anatomic geometry.

Recovery of the moving dipole from surface potential recordings

Abstract Two methods of recovering the equivalent dipole of variable locus and moment from voltages of 140 or fewer surface recording sites have been examined: (1) summation of all available elements (Gabor and Nelson 8 ) and (2) lead-vector search (Helm and Chou 11 ). Summation requires many recording sites but is rapidly done on a digital computer, whereas lead-vector search requires fewer recording sites but involves much longer computing times. Recovery of a single noise-free factitial input dipole by summating over 140 points was good as to locus and moment. Recovery by lead-vector search was also good whether 7 or 28 electrodes were employed. Increasing error of estimation—especially of locus—was noted, however, with the lead-vector search with (1) increasing the noise level, (2) diminishing the number of electrode sites (from 28 to 20 to 7), and (3) increasing the complexity of the generator (from single dipole input, to double dipole input, to live data). Clinical application awaits improved methodology—perhaps by compromise between slow-recording, fast-computing summation, and fast-recording, slow-computing lead-vector search.

A dynamic electrical record of the pathway of human His bundle activation from surface mapping.

Body surface potential maps of human His bundle activity have been difficult to produce for two reasons: (1) The peak surface potentials are often less than 5 μv, and (2) the simultaneous atrial repolarization potentials frequently exceed 100 μv. We have therefore amplified surface signals 25 times the standard gain of 1000, and then removed by cross-correlation the static pattern of atrial repolarization for serial 1-msec maps of the P-R segment in five normal men. A consistent finding emerged: a positive anterior chest peak appeared 40 msec before QRS onset, and then-within 10 msec-spread out into a long, low transverse mound before disappearing in 5 more msec. The map data were analytically converted to serial electrical sources: the center of electrical activity moved first slightly down, then directly forward, before retracing its path and disappearing. The retrace and accompanying surface spread-out strongly suggests diverging dipolar sources. Thus the data fit a simple heart source which moves anteriorly and then breaks into two (right and left)-as expected from activation of the bundle of His and its bifurcation into left and right bundle branches.

The electrical sequelae of aerosol inhalation

Abstract Of the 16 animals, then, one was allowed to inhale glue which contained toluene and acetone but no fluoroalkane gas and developed a mild sinus tachycardia only. One animal died accidentally from ventricular fibrillation due to improper pacing on the part of the investigators. A third animal was deliberately hyperoxygenated but did experience sinus slowing. In the remaining 13 animals, evidence of suppression of the rate of the sinus node occurred within seconds to minutes of aerosol inhalation. The most frequent chain of events included sinus bradycardia, A-V dissociation with progressively lower escape rhythms, and ultimate electrical asystole or ventricular fibrillation. The import of these data would strongly suggest that, whatever the offending agent proves to be, the would-be sniffer of aerosols in high concentration cannot depend upon subjective sensations in order to cease inhalation in time, for in half our animals we were unable to reverse the relentless development of fatal rhythm disturbances in spite of reinstitution of artificial respiration and oxygenation when sinus bradycardia first occurred.

An electrocardiographic and spatial vectorcardiographic pattern associated with diffuse myocardial damage and ventricular aneurysm.

Abstract Aspects of clinical, pathologic, electrocardiographic, and spatial vectorcardiographic findings in 24 patients with diffuse myocardial damage, infarction, and ventricular aneurysm have been described. A distinctive QRS pattern (0.10 to 0.16 second) in Leads I, V 5 , or V 6 characterized by a brief early upward deflection followed by a shallow downward deflection and, finally, a prolonged prominent R′ was found in 13 patients. Persistent elevation of the S-T segment was noted in 5 of the remaining patients. The same QRS pattern was found in 4 other patients with myocardial infarction and diffuse scarring but without ventricular aneurysm.

Patterns of body surface potential and ventriculograms specific to occlusion of subdivisions of the coronary arteries.

The thoracic activation map patterns, the distribution of occlusions, and the ventriculograms obtained at cardiac catheterization were examined in 166 patients with multivessel coronary disease without conduction defects or prior coronary intervention. The mean potential and ventriculographic configurations were determined for 15 groups, each formed on the basis of significant luminal occlusion (less than or equal to 90%) of an individual coronary arterial subdivision. We mathematically extracted distinctive map and wall motion patterns specific for isolated occlusion of each of the 15 major subdivisions. For these prototypes we found the following: 1) Definitive change in electrical pattern (less than 2 SD from the normal mean) occurred frequently outside the electrode sites of the standard electrocardiogram. 2) Focal akinesis systematically followed lesion site down the arterial courses; early electrical activation patterns corresponded to identifiable anatomic loss. 3) Certain paradoxes arose, for example, similar wall motion change but quite dissimilar electrical patterns from posterior descending artery occlusion of right coronary versus left circumflex origin. This technique unmasks component surface electrical patterns and ventriculographic deformities otherwise unrecognized in multisite, multivessel disease.

Body surface mapping including relationships with endocardial and epicardial mapping.

Recording and display of the temporal and spatial distribution of surface potentials recorded from multiple sites on the torso is termed body surface mapping (BSM). The concept of mapping was probably first articulated by Waller a t St. Mary’s Hospital in an inaugural address, who in 1888 published a torso map of the voltage field about the heart. H e based his description upon the concept that the cardiac generator could be represented by a bipole, resulting in positive and negative isopotential lines projected onto the thoracic surface (FIGURE l).’ The first recording we are aware of in isopotential map form of successive instants of the QRS complex was by Koch and Schneyer.* Instant-by-instant mapping of surface voltage change throughout the cardiac cycle was described by Nahum et al. who demonstrated a surface potential field that was more complex than could likely result from a dipolar s o ~ r c e . ~ Nelson first clearly portrayed multipolar sources as multiple peaks and valleys in the curves of instantaneous potential distribution recorded from a band of electrodes about the chest,4 after which Taccardi in I t a l ~ , ~ , ~ Amirov in Russia,’ and Horan et al. in the United States* published comprehensive maps from both dogs and men in which evidence of both dipolar and nondipolar activity was apparent. Maps were originally drawn from manual measurements of scalar electrocardiograms. The acceleration of interest in BSM is directly related to technological advances making computers available to handle tedious and time-consuming data aquisition and reduction, and to increasingly sophisticated instrumentation. This review will not explore the basic biophysical underpinnings of the science of mapping or details of efforts to solve the “inverse problem.” It will, in the main, deal with physiologic correlations and pathophysiologic relationships. TABLE 1 summarizes the rationale for BSM. The reader is also referred to two recent significant works in the area of BSM: an entire book edited by Mirvis on BSM9 and a chapter by Ambroggi et al. in a three-volume set on electrocardiology edited by MacFarlane and Laurie.” BSMs from our laboratory have been obtained by two methods. At an early stage we utilized a multichannel Ampex FM tape recorder to store analog signals from an array of 1000-gain low-level amplifiers from 142 equally spaced sites about the torso. Potentials were then digitized, signal-averaged seven channels a t a time, and then temporally collated (using the eighth channel) into map format by a digital computer. More recently, we have utilized an array of 35 chest electrodes which had been shown