This Article Abstract Full Text (PDF) Data Supplement All Versions of this Article: 91/4/331    most recent 01.RES.0000031957.70034.89v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hooks, D. A. Articles by Hunter, P. J. Search for Related Content PubMed PubMed Citation Articles by Hooks, D. A. Articles by Hunter, P. J. Related Collections Structure Electrophysiology Pacemaker Arrythmias-basic studies Quantitative modeling

(Circulation Research. 2002;91:331.)
© 2002 American Heart Association, Inc.

Integrative Physiology

Cardiac Microstructure

Implications for Electrical Propagation and Defibrillation in the Heart

Darren A. Hooks, Karl A. Tomlinson, Scott G. Marsden, Ian J. LeGrice, Bruce H. Smaill, Andrew J. Pullan, Peter J. Hunter

From the Bioengineering Research Group, University of Auckland, Auckland, New Zealand.

Correspondence to Darren Hooks, PhD, Department of Physiology, School of Medicine, University of Auckland, Private Bag 92019, Auckland, New Zealand. E-mail d.hooks{at}auckland.ac.nz

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Our understanding of the electrophysiological properties of the heart is incomplete. We have investigated two issues that are fundamental to advancing that understanding. First, there has been widespread debate over the mechanisms by which an externally applied shock can influence a sufficient volume of heart tissue to terminate cardiac fibrillation. Second, it has been uncertain whether cardiac tissue should be viewed as an electrically orthotropic structure, or whether its electrical properties are, in fact, isotropic in the plane orthogonal to myofiber direction. In the present study, a computer model that incorporates a detailed three-dimensional representation of cardiac muscular architecture is used to investigate these issues. We describe a bidomain model of electrical propagation solved in a discontinuous domain that accurately represents the microstructure of a transmural block of rat left ventricle. From analysis of the model results, we conclude that (1) the laminar organization of myocytes determines unique electrical properties in three microstructurally defined directions at any point in the ventricular wall of the heart, and (2) interlaminar clefts between layers of cardiomyocytes provide a substrate for bulk activation of the ventricles during defibrillation.

Key Words: bidomain model • defibrillation • finite elements • anisotropy • discontinuous conduction

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
There is now compelling evidence that ventricular cardiomyocytes are not arranged in a uniformly connected continuum as has often been assumed in the past. Microscopic studies of ventricular myocardium have revealed that myocytes are assembled in distinct layers^1,2 with extensive clefts, or cleavage planes, between these layers. Such laminar architecture has been observed in hearts of different species^3–7 and is characterized as an ordered system of interconnected muscle layers or sheets^2,5,6 that have a predominantly radial orientation in apex-base transmural sections.^3–7 The extent to which the discontinuous laminar structure of ventricular myocardium influences the electrical properties of the heart remains unclear.

The effects of discontinuous structures on propagation velocity^8,9 and the safety of conduction¹⁰ have been studied in a variety of two-dimensional tissue preparations. However, it is difficult to extrapolate from these data to the complex three-dimensional architecture of ventricular myocardium. It is commonly assumed that propagation proceeds with equal velocity in all directions transverse to the myofiber axis.^11,12 This notion of transverse electrical isotropy is at odds with recent descriptions of myocardial structure. It has been proposed that the sheet structure of myocardium may lead to orthogonal anisotropy with electrical propagation across muscle layers slower than propagation transverse to the myofiber axis within layers.⁶ This hypothesis has not previously been tested.

Structural discontinuity has been widely invoked to explain the myocardial response to electrical shock.^13–17 It is acknowledged that defibrillating shocks applied to the surface of the heart would not depolarize a sufficient volume of tissue to achieve cardioversion if myocardium behaved as a continuum.¹⁶ An explanation is that discontinuities such as gap junctions, collagenous septa, and blood vessels give rise to localized regions of depolarization during defibrillation, thereby activating a critical mass of tissue.^13,15–17 These local depolarizations are termed secondary sources of excitation, where the primary source is at the electrode-tissue interface. Secondary sources arise at positions where the current flux between the stimulating electrodes is forced to cross the cellular membrane in order to traverse discontinuity of the intracellular resistive network (see Figure 7 of Reference ¹⁷). Fast et al¹³ have suggested that the cleavage planes between muscle layers may provide an important substrate for the formation of such secondary sources. This hypothesis was supported by studies using cultured cardiac myocytes, but it has yet to be addressed in three-dimensional ventricular myocardium.

In the present study, we approach these questions using a novel combination of extended-volume confocal reconstruction of tissue microstructure⁷ and finite element computer modeling of electrical propagation. This combination provides a powerful means of relating structure to function and overcomes the lack of an experimental technique for examining three-dimensional electrical events in the heart at appropriate spatial resolution. We describe a model of electrical propagation solved in a structurally discontinuous domain, which accurately represents the measured three-dimensional microstructure of a transmural segment of left ventricular myocardium. The model describes discrete discontinuities at the interlaminar cleft level while averaging the effects of gap junctional discontinuities at the cellular level. The governing equations are the bidomain equations that have been well validated in a number of experimental studies.¹⁸ We use the model to address the hypotheses that (1) early propagation from a focal activation can only be accurately described by a discontinuous model of the myocardium, (2) the laminar organization of myocytes determines unique propagation velocities in three microstructurally defined directions at any point in myocardium, and (3) interlaminar clefts, or cleavage planes, provide a means by which an externally applied shock can influence a sufficient volume of heart tissue to terminate cardiac fibrillation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Structural Representation
Extended confocal microscopy was used to reconstruct the three-dimensional architecture of myocytes in a transmural segment of rat left ventricular myocardium.⁷ The image volume was 0.8x0.8x3.7 mm and consisted of 6.07x10⁸ cubic voxels with 1.56-µm sides. The reconstructed volume (Figure 1; see movie 1 in the online data supplement, available at http://www.circresaha.org) confirmed that ventricular myocytes are arranged in branching layers separated by extensive interlaminar clefts (or cleavage planes).⁶ Myofiber orientation varied smoothly through the ventricular wall from around 75° below the circumference at the epicardial surface to 75° above the circumference at the endocardial surface. Three orthogonal unit vectors (a_l, a_t, and a_n) that reflect this microstructure can be identified at any point¹⁹: a_l lies parallel to myofibers, a_t is transverse to the myofibers but in the plane of the muscle layer, and a_n is normal to the layer. For further analysis, the tissue reconstruction was reduced to a stack of 18 transmural images, each with square pixels of side dimension 4.69 µm. One of these images is shown in Figure 1. These images were used to reconstruct the three-dimensional geometry of the cleavage planes that coursed through the tissue volume. A network of bilinear finite elements²⁰ describing the cleavage plane geometries was generated from the image stack by manual assignment of element nodes to regions devoid of myocyte-to-myocyte connections. Each finite element spanned a distance of 47 µm in the z direction (Figure 1), which was the distance between successive images in the stack. In total, 1540 elements and 2803 nodes were required for accurate geometrical description of the cleavage planes (Figure 1).

View larger version (66K): [in this window] [in a new window]   Figure 1. Top, Reconstructed volume of rat left ventricular free-wall myocardium. Middle, Transmural slice from the reconstructed volume showing a complex network of cleavage planes that course between layers of myocytes. Each layer is of an order of 80 µm in thickness. Bottom, Bilinear finite element geometrical description of the cleavage planes through the entire rat tissue block and a smaller midwall subsection. Myofiber orientation is shown on the epicardial (epi) and endocardial (endo) surfaces. See online movies in the online data supplement for a three-dimensional reconstruction of the volume.

Discontinuous Bidomain Model Formulation
A bidomain approach¹⁸ was used to model electrical propagation in the reconstructed tissue volume so that the effects of the interlaminar clefts (cleavage planes) on the spread of activation could be investigated. With this approach, the coupling between ventricular myocytes and extracellular space is represented on a continuum scale by defining two domains that coexist at each point throughout the volume. Conductivity tensors G_i and G_e for these intracellular and extracellular domains, respectively, are defined by

equation

where g_l;i,e, g_t;i,e, and g_n;i,e are effective intracellular and extracellular conductivities along the local microstructural axes a_l, a_t, and a_n.¹⁸

Within this context, interlaminar clefts must be viewed as discontinuities in the intracellular domain, across which there is zero intracellular flux. We sought a solution to the active bidomain equations that allowed accurate spatial representation of the interlaminar clefts, in order to apply the condition of zero intracellular flux across these boundaries. A Galerkin finite element method²⁰ was used on a mesh of trilinear elements representing the reconstructed tissue volume. Disruption of cellular connections across cleavage planes was achieved by the removal of elements of the trilinear mesh from the intracellular domain and application of no-flux (Neumann) boundary conditions along the resulting internal intracellular domain boundaries. Examples of the procedure are illustrated in Figure 2. Representation of cleavage planes required removal of 11% of the solution mesh volume from the intracellular domain. The entire mesh consisted of 483 840 nodes. The finite element method generates systems of equations in V_m (transmembrane potential) and _e (extracellular potential), which were solved with a semi-implicit time integration scheme utilizing the conjugate gradient method. A finite element solution of the bidomain equations was used in preference to a finite difference approximation, as the Neumann condition could be applied at all internal intracellular boundaries with a minimum number of degrees of freedom.

View larger version (97K): [in this window] [in a new window]   Figure 2. A, Discontinuous finite element method. The bidomain model describes two domains that coexist at each point in space: the intracellular (bottom plane) and extracellular (top plane) domains. In this two-dimensional example, a discontinuity described in space by one-dimensional linear elements (yellow lines) is introduced into a bilinear element mesh by the removal of elements from the intracellular domain (region shown in green). No-flux boundary conditions are applied at the perimeter of each domain, as well as on the resulting internal boundaries of the intracellular domain (red lines). B, Superimposed on the tissue image of Figure 1 is a cross section through the bilinear finite element description of cleavage planes (shown in yellow) used in the model, with the corresponding regions of removed intracellular elements (shown in green). Inset shows subregion at higher magnification, with red lines indicating internal no-flux boundaries, as in panel A.

The model incorporated the following linear variation of myofiber angle () through the ventricle wall based on measurements from the reconstructed tissue volume⁷: =0.83x-1.28 radians (see Figure 1 for direction of x).

Intracellular and extracellular conductivities were assumed to be transversely isotropic with respect to fiber direction (g_n;i,e=g_t;i,e). Any possible transverse anisotropy of the intracellular domain was explicitly modeled by the inclusion of the cleavage plane discontinuities.

Two sets of simulations were carried out to investigate both the response of the tissue to propagation after point stimulation (part I) and the response to defibrillation-strength shock (part II).

Part I: Propagated Response
The reconstructed tissue volume was stimulated at twice diastolic threshold by applying transmembrane current for 2 ms at its center. The propagated response was followed through time, until the volume was fully depolarized (18 ms). A simple cubic ionic current model²¹ was used in the simulation to reduce computational expense. Neumann boundary conditions were applied along all external boundaries of the reconstructed tissue in both intracellular and extracellular domains, and _e was measured relative to position {x=3.7, y=0, z=0.8} (see Figure 1). Solutions required 85 hours of CPU time with 8 processors of an SGI Origin 2000 (250 MHz R10000 CPU units).

A major premise underlying this part of the study is that electrical propagation through myocardium perpendicular to muscle layers is slower than propagation within layers transverse to the myofiber axis because of the discontinuities introduced by interlaminar clefts. To obtain spatially averaged estimates of propagation in the discontinuous bidomain model, results were matched with a continuous model in which orthotropy was incorporated through unequal transverse intracellular conductivities (g_n;ig_t;i). Comparison of these results also resolved the effects of structural discontinuity on local propagation. In the continuous model, a conventional bidomain formulation was used and the following continuous description of the transmural variation in cleavage plane angle (ß; defined in Costa et al²²) was used: ß=-0.0892x³+0.5595x²-0.8114x+0.8669 radians.

The microstructural axes a_t and a_n were then oriented according to the cleavage plane angle description. The intracellular conductivity normal to the cleavage plane, g_n;i, was set at a value (g_n;i^MIN) that minimized the root mean square (rms) difference between the continuous and discontinuous model solutions for transmembrane potential while the same value for g_t;i was used for both models.

Part II: Shock-Induced Response
A defibrillation-strength shock was applied across the reconstructed tissue volume. Constant current (10 ms in duration; uniform density of 0.14 mA/mm²) applied to the extracellular domain at the epicardial (cathodal) and endocardial (anodal) surfaces induced an extracellular potential gradient of 1 V/mm across the ventricle wall. The shock response was modeled using the Drouhard-Roberge–modified Beeler-Reuter ionic current model (BRDR)^23,24 with recent revisions.²⁵ During the shock, the transmural boundaries of the reconstructed tissue volume acted to redistribute current between the extracellular and intracellular spaces because of the oblique angle of approach of the myocyte laminae to these boundaries. To minimize these artifactual boundary effects, the domain was expanded laterally (in y and z directions; expanded volume=4x4x3.7 mm), allowing continuity of current at the transmural boundaries of the reconstructed volume. For computational efficiency, a continuous description of the laminar structure was included in this additional region bounding the central tissue volume in which the discontinuities were explicitly represented.

As in part I, the shock-induced response of this discontinuous model was compared with that of a fully continuous model in which the same cubic description of cleavage plane angle was applied throughout the entire tissue volume.

Model Parameters
The model parameters used in this study are given in the Table and discussed in the online data supplement, available at http://www.circresaha.org.

View this table: [in this window] [in a new window]   Table 1. Parameters of the Model

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Part I
The activation sequence after point stimulation at the center of the tissue volume is shown for the discontinuous and continuous models in Figures 3A and 3B, respectively. Potential maps are given on seven planes through the heart wall at 2.7-ms intervals after stimulation (see movie 2 in the online data supplement for the complete activation sequence). Discontinuous propagation around the cleavage planes is evident from the irregularity in isopotential lines in Figure 3A. Important differences between the two models are apparent. First, activation in the discontinuous model is markedly asymmetric about the site of stimulation in its early stages (see the 5.4-ms panel of Figure 3A) compared with the symmetrical activation of the continuous model. Second, the insulating boundaries formed by cleavage planes allow juxtaposition of fully depolarized and nondepolarized tissue regions in the discontinuous model (eg, at 10.8 ms).

View larger version (103K): [in this window] [in a new window]   Figure 3. Potential maps for the discontinuous model (A) and continuous model (B). Transmembrane potentials are mapped on 7 equally spaced surfaces through the reconstructed rat tissue volume, at 2.7-ms intervals after midwall stimulation. Isopotential lines at 5-mV intervals are shown in black. Site of stimulation is shown with black dot at center of volume. The cleavage plane obstacles in panel A lead to a highly discontinuous form of propagation that is, however, approximated well by the continuous model, in terms of global directions and rates of propagation. The full activation sequence may be viewed in the online movies (see online data supplement).

Despite local differences between the two models, the continuous model reproduces key features of the global pattern of excitation in the discontinuous model. For example, earliest activation skews toward the edges (y=0, z=0) and (y=0.8, z=0) as the wavefront propagates through the tissue volume in both models. The conductivity parameter g_n;i^MIN used in the continuous model of Figure 3B was that which minimized the rms difference between the continuous and discontinuous solutions over the 18 ms required for full activation of the tissue block. The rms difference between models is 5.2 mV, and the best-fit parameter g_n;i^MIN was 0.0107 S/m, giving the ratio g_t;i/g_n;i^MIN=2.44. With this optimization procedure, it is assumed that g_n;i is uniform throughout the block in the continuous model. However, there were fewer cleavage planes in subepicardial and subendocardial regions than in the midwall, suggesting that g_n;i^MIN may overestimate conductivity normal to muscle layers in the midwall region. Excluding the final 6 ms of propagation (where the wavefront enters subepicardial and subendocardial regions) from the optimization procedure reduced g_n;i^MIN to 0.0080 S/m (g_t;i/g_n;i^MIN=3.25, minimized rms difference=4.95 mV). The nonunity value of g_t;i/g_n;i^MIN indicates that the tissue block behaves in an electrically orthotropic fashion: propagation across layers of myocytes is substantially slowed because of the convoluted path the wavefront must take around cleavage planes. In fact, one effect of this orthotropy is the tendency of earliest activation to skew toward the upper and lower left edges of the domain as the wavefront progresses (Figures 3A and 3B).

It is well recognized that discontinuity in wavefront propagation may lead to complex polyphasic (fractionated) extracellular potential recordings.^9,26 To investigate the possibility of fractionation in extracellular recordings from ventricular tissue, we plotted extracellular potential waveforms at 28 sites throughout the rat tissue volume. The signals from the discontinuous and continuous models are shown in Figure 4 (top and bottom panels, respectively). All signals are referenced to the point {x=3.7, y=0, z=0.8} and hence approximate unipolar recordings as might be performed in real tissue. The distance of each recording location from the central site of stimulation is given by the signal color (red indicating close; blue indicating far; see Figure 4 legend), while the y and z coordinates of the recording locations are shown above the traces in parentheses. Several observations can be made from these traces. First, nearly all signals from the discontinuous model show some degree of fractionation. The extent of fractionation appears highest in the early stages of propagation. Second, it is evident that the downstroke duration (time over which signal derivative is negative) of signals recorded close to the stimulus site is considerably longer in the discontinuous model than in the continuous model. Downstroke duration was measured by the time required for the downstroke to change from 90% to 10% of its maximum amplitude. The mean duration across red and yellow signals (Figure 4) was 7.25 and 2.81 ms for the discontinuous and continuous models, respectively. A comparable difference in transmembrane potential upstroke duration (measured by the time required for the upstroke to change from 10% to 90% of its maximum amplitude) was not seen at the same recording locations (mean times were 3.23 and 2.77 ms for discontinuous and continuous models).

View larger version (34K): [in this window] [in a new window]   Figure 4. Plots of extracellular potential versus time for the discontinuous model (top panel) and continuous model (bottom panel). Signals are grouped according to their y, z locations (shown in millimeters, in parentheses above the traces). Within each group, the distance from the recording point to the central stimulus site is noted by signal color (see legend). Open circles indicate the activation time defined by the maximum negative slope of the extracellular potential signal, and solid circles indicate the local activation time defined by the maximum positive slope of the transmembrane potential signal from the same site. Where the open circle is not visible, it is coincident with the solid circle. The time scale is the same for both top and bottom panels; however, the potential scale varies, as indicated by the scale bars.

Traditionally, the time of maximum negative slope (-V'_max) of a unipolar extracellular recording is used to pick the time of activation at the site of measurement. ²⁷ In the fractionated recordings of Figure 4, there are often two distinct regions of similar negative slope. To test the ability of -V'_max to discern local activation time, we compared the time at -V'_max (open circles, Figure 4) with the time of maximum positive slope of the transmembrane potential recording from the same site (closed circles, Figure 4). The two markers of activation are shown on each trace where time differences were greater than the interval used in the slope calculation (0.1 ms). Local time of activation was correctly determined by -V'_max in all but two traces of the continuous model. There were more cases of inaccurate assignment in the discontinuous model; however, in only two cases was the error >0.6 ms.

Part II
The activation sequence during the shock delivered across the ventricular wall is shown for the discontinuous model in Figure 5A and for the continuous model in online Figure 1, available in the online data supplement at http://www. circresaha.org. Transmembrane potentials are mapped on a central plane (z=0.4 mm) at each time step, and the color spectrum is scaled to distinguish depolarized from hyperpolarized regions. Numerous islands of depolarization are formed on the endocardial side of the cleavage plane discontinuities. These localized excitatory regions, or secondary sources,¹³ arise from redistribution of current between intracellular and extracellular compartments, allowing current flow around the insulated cleavage plane boundaries of the intracellular domain. They are responsible for activation of the bulk myocardium within the 10-ms duration of the applied shock. In the continuous model (online Figure 1), transmembrane potential varies smoothly, and most of the tissue remains hyperpolarized during the shock. Figure 5B shows the three-dimensional architecture of secondary sources 2.5 ms into the shock period. An extensive set of surfaces on which transmembrane potential is equal to -60 mV is distributed throughout the tissue volume (see movie 3 in the online data supplement for the full activation sequence). The largest region of secondary source depolarization arises at the junction between midwall and subendocardial regions (Figure 5A, 4 ms), where there can be maximal redistribution of current between intracellular to extracellular spaces because of the lack of further discontinuity toward the endocardium.

View larger version (62K): [in this window] [in a new window]   Figure 5. A, Progression of activation in the discontinuous model during the 10 ms of applied shock. Transmembrane potentials are mapped on a single midvolume plane (z=0.4 mm) at 2-ms time increments after onset of shock. The black lines depict the cleavage plane discontinuities. B, Three-dimensional view of secondary sources at 2.5 ms into the shock. Surfaces on which transmembrane potential is -60 mV are drawn throughout the volume. The progression of these surfaces through time is shown in the online movies (see online data supplement).

Figure 6 presents transmembrane potentials along a line drawn through the center of the reconstructed tissue volume (from epicardium to endocardium), for both discontinuous and continuous models, at 6 ms into the shock. This figure demonstrates the full range of transmembrane potentials present in the tissue volume. The continuous model shows an activation wavefront moving away from the depolarized epicardium and located at 1 mm from it.

View larger version (17K): [in this window] [in a new window]   Figure 6. Plot of transmembrane potential versus tissue depth (distance from epicardium) for both models. Potential varies smoothly in the continuous model, and much of the tissue remains hyperpolarized. In the discontinuous model, cleavage plane obstructions to intracellular current flow lead to multiple inflections on the potential profile, and depolarized regions exist throughout most of the tissue volume.

It is clear from Figures 5 and 6 that cleavage plane discontinuities facilitate the synchronous activation of a large volume of myocardium during shock application, through the formation of localized (secondary) regions of depolarization.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Cardiac excitation and contraction have previously been modeled in our laboratory on global representations of cardiac anatomy using the finite element numerical solution technique.¹⁹ In the present study, we extend this approach to investigate how excitation processes could be influenced by structure at a much finer scale. We extracted an accurate representation of ventricular microstructure for use in the model from a high-resolution dataset previously assembled by sequential sectioning and extended confocal imaging of a transmural block of rat left ventricular myocardium. ⁷ These data clearly exhibit the laminar arrangement of myocytes in ventricular myocardium reported elsewhere.^5,6 Application of the bidomain equations to the finite element representation of this microstructure yielded interesting insights into the effects that discontinuities associated with muscle layers have on excitation.

Implications for Propagation
The principal finding of the study of electrical propagation in ventricular myocardium is that discontinuities associated with the laminar arrangement of cardiac myocytes could markedly influence the spread of electrical activity from a focal source of excitation. The spread of activation is predicted to be nonuniform and potentially asymmetric about the focus of stimulation. Moreover, propagation is not transversely isotropic but moves most slowly through the myocardium perpendicular to muscle layers. It is noteworthy that slow and nonuniform propagation are factors implicated in the genesis and maintenance of reentrant cardiac arrhythmia.

Nonuniform local propagation can be modeled only by explicit inclusion of structural discontinuities, and this should be taken into account in future computer models of reentrant arrhythmia. However, this study also shows that the global spread of electrical activation from a point excitation source is represented with reasonable accuracy by continuous orthotropic models in which separate electrical conductivity values are associated with three microstructurally defined material directions.

We have previously estimated the relative velocities of midwall electrical propagation in each of the three material directions based on images of myocardial structure and two-dimensional measurements of the pathlength a propagating wave might take to traverse cleavage planes.²⁸ Pathlength measurements do not take into account the possibility of direct stimulation of myocardium across cleavage planes due to current flows in the extracellular domain. Spread of activation between two completely disconnected intracellular spaces by such electrotonic effects was observed in the model studies reported here. For physiological extracellular conductivities, however, the delay of propagation at interlaminar clefts slows propagation in the direction normal to myocyte laminae. A conductivity ratio g_t;i/g_n;i^MIN of 3.25 was fitted in this study for propagation in the midwall. Our previous pathlength measurements suggested a ratio of propagation velocities c_t/c_n of 2 in the midwall. ²⁸ To compare these results, we use the following relationship²⁹:

equation

and obtain a velocity ratio c_t/c_n of 1.68. These results appear consistent given that the pathlength measurements were made in two-dimensional tissue sections²⁸ and probably overestimate the effective length of propagation pathways in three-dimensional myocardium.

In addition to anisotropy of conduction velocity, we expect discontinuous propagation in the ventricle to be reflected by other measures. Polyphasic extracellular potential waveforms have previously been observed in human ventricular tissue,³⁰ atrial pectinate muscle bundles from aged human subjects,⁹ and in infarct border zone or ischemic regions of myocardium.³¹ The results of this study indicate that extracellular recordings from sites close to a midwall focal stimulus, in healthy ventricular tissue, may also exhibit complex multiple inflections. These recordings may also show a much reduced rate of downstroke than previously expected. These two characteristics may lead to inaccurate assignment of the local time of activation based on the maximum negative derivative of the extracellular signal.

Implications for Defibrillation
Electrical shock applied to the heart achieves defibrillation by simultaneously resetting myocardium into a uniform electrical state. It is necessary for the applied shock to influence the membrane potential of a substantial volume of tissue. We have presented striking new evidence that interlaminar clefts within the ventricles are critical to this process, allowing activation of the bulk of the myocardium during the applied shock.

Over the past decade, many mechanisms by which an applied shock to the heart may alter transmembrane potential have been elucidated. "Virtual electrode" is now common jargon to refer to any site of shock-induced transmembrane potential change distant from the site of current injection. It has been shown that virtual electrodes may be induced by unequal anisotropy of intracellular and extracellular spaces, ³² myofiber curvature, ³³ fiber narrowing, ³⁴ spatial inhomogeneity of intracellular volume fraction, ¹⁶ and discontinuity associated with gap junctions ¹⁵ and intercellular clefts. ¹³ Any geometrical condition that induces current redistribution across the cell membrane during stimulation will lead to formation of a virtual electrode. The present day challenge is to ascertain how all these mechanisms affect the ability of a shock to terminate fibrillation.

We assert that the most important form of structural discontinuity in terms of cardioversion is the interlaminar cleavage plane not the gap junction. Modeling studies of gap junction–induced virtual electrodes have indicated spatial fluctuations in transmembrane potential of a few millivolts¹⁵; however, these sawtooth patterns have not been seen experimentally.¹³ There is further previous evidence for our assertion. The threshold required for diastolic cathodal point stimulation in the midwall has been measured as 0.8 V/cm.³⁵ It is necessary to assign discontinuity at the level of bundles of myocardial fibers rather than gap junctions to match this value in a discontinuous syncytial model of myocardium.¹⁴ Our structural model incorporates transmural fiber rotation and unequal anisotropy, as well as structural discontinuity. Under conditions of uniform field stimulation, it is the cleavage plane discontinuities that are responsible for the activation of the bulk of the myocardium within the duration of the shock.

It is expected that a shock applied to the body surface will induce a relatively uniform current density field within the myocardium. The shock field from an implantable cardioverter-defibrillator (ICD) is likely to be less uniform, however, and the relative role of structural discontinuity in this form of far-field stimulation remains to be investigated. In the setting of nonuniform field stimulation, unequal anisotropy between intracellular and extracellular spaces can by itself generate virtual electrodes in the form of a dog-bone shape of polarization, with adjacent regions of opposite polarization. ³² It is interesting to note that the experimental studies confirming these patterns have been carried out on the epicardial surface of the heart,³⁶ where coupling between myocytes is increased with respect to the midwall, and large intercellular clefts are absent. It is yet to be seen whether a dog-bone pattern of polarization is in fact visible at all, deep within the midwall, during nonuniform field application. The development of a transmural probe able to record transmembrane potentials deep within myocardium³⁷ may yield experimental information on these processes, albeit at a lower spatial resolution than the scale of cleavage plane discontinuity.

Limitations of the Study
The most severe limitation of this work stems from the restricted volume over which the cleavage plane geometries could be reconstructed. The length of the rat tissue volume from epicardial to endocardial surfaces was 4.63 times its width, which resulted in the propagating wavefront reaching the transmural extents of the tissue block in the early stages of propagation from the central stimulus site (Figure 3). The presence of the transmural boundaries acts to reduce the available paths of propagation around the cleavage plane obstacles, in turn enhancing the effect of the cleavage planes on the slowing of propagation. As a result, it is likely that we have obtained an overestimate of g_t;i/g_n;i^MIN for this tissue block. The boundary effect does not, however, change our overall conclusion that the tissue behaves in an electrically orthotropic fashion. Other limitations of the study relate to the piecewise cube approximation of the gap space between myocyte laminae that was used, and the assumption of isotropy of the extracellular space in the plane orthogonal to myofiber direction that was made. These limitations are discussed further in the online data supplement.

Conclusions
In summary, this study contributes two observations that we believe are of primary importance to cardiac electrophysiology. First, we have shown that the standard notion^12,13 that myocardium is electrically isotropic in the plane orthogonal to myofiber direction is incorrect; the laminar organization of myocytes leads to orthogonal anisotropy of electrical properties within the ventricular wall. Second, we have demonstrated that the discontinuities associated with muscle layers could play a significant role in the termination of fibrillation by an externally applied shock.

	Acknowledgments

 
This study was supported by the Marsden Fund and the Maurice and Phyllis Paykel Trust of New Zealand.

Received February 27, 2002; revision received July 19, 2002; accepted July 22, 2002.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Caulfield JB, Borg TK. The collagen network of the heart. Lab Invest. 1979; 40: 364–372.[Medline] [Order article via Infotrieve]
2. Robinson TF, Cohen-Gould L, Factor SM. Skeletal framework of mammalian heart muscle. Lab Invest. 1983; 49: 482–498.[Medline] [Order article via Infotrieve]
3. Robb JS, Robb RC. The normal heart: anatomy and physiology of the structural units. Am Heart J. 1942; 23: 455–467.[CrossRef]
4. Feneis H. Da gefüge des herzmuskels bei systole und diastole. Morphol Jahrb. 1943; 89: 371–406.
5. Hort W. Makroskopische und mikrometrische untersuchungen am myokard verschieden stark gefüllter linker kammern. Virchows Arch Pathol Anat. 1960; 333: 523–564.[CrossRef]
6. LeGrice IJ, Smaill BH, Chai LZ, Edgar SG, Gavin JB, Hunter PJ. Laminar structure of the heart: ventricular myocyte arrangement and connective tissue architecture in the dog. Am J Physiol. 1995; 38: H571–H582.
7. Young AA, LeGrice IJ, Young MA, Smaill BH. Extended confocal microscopy of myocardial laminae and collagen network. J Microsc. 1998; 192: 139–150.[Medline] [Order article via Infotrieve]
8. Kucera JP, Kleber AG, Rohr S. Slow conduction in cardiac tissue, II: effects of branching tissue geometry. Circ Res. 1998; 83: 795–805.[Abstract/Free Full Text]
9. Spach MS, Dolber PC. Relating extracellular potentials and their derivatives to anisotropic propagation at a microscopic level in human cardiac muscle: evidence for electrical uncoupling of side-to-side fiber connections with increasing age. Circ Res. 1986; 58: 356–371.[Abstract/Free Full Text]
10. Rohr S, Kucera JP, Fast VG, Kleber AG. Paradoxical improvement of impulse conduction in cardiac tissue by partial cellular uncoupling. Science. 1997; 275: 841–844.[Abstract/Free Full Text]
11. Muzikant AL, Henriquez CS. Bipolar stimulation of a three-dimensional bidomain incorporating rotational anisotropy. IEEE Trans Biomed Eng. 1998; 45: 449–463.[CrossRef][Medline] [Order article via Infotrieve]
12. Knisley SB, Trayanova N, Aguel F. Roles of electric field and fiber structure in cardiac electric stimulation. Biophys J. 1999; 77: 1404–1417.[Abstract/Free Full Text]
13. Fast VG, Rohr S, Gillis AM, Kléber AG. Activation of cardiac tissue by extracellular electrical shocks: formation of ‘secondary sources’ at intercellular clefts in monolayers of cultured myocytes. Circ Res. 1998; 82: 375–385.[Abstract/Free Full Text]
14. Krassowska W, Frazier DW, Pilkington TC, Ideker RE. Potential distribution in three-dimensional periodic myocardium, part II: application to extracellular stimulation. IEEE Trans Biomed Eng. 1990; 37: 267–284.[CrossRef][Medline] [Order article via Infotrieve]
15. Trayanova N. Discrete versus syncytial tissue behavior in a model of cardiac stimulation, II: results of simulation. IEEE Trans Biomed Eng. 1996; 43: 1141–1150.[CrossRef][Medline] [Order article via Infotrieve]
16. Trayanova N. Far-field stimulation of cardiac tissue. Herz Elektrophysiol. 1999; 10: 137–148.[CrossRef]
17. Newton JC, Knisley SB, Zhou X, Pollard AE, Ideker RE. Review of mechanisms by which electrical stimulation alters the transmembrane potential. J Cardiovasc Electrophysiol. 1999; 10: 234–243.[Medline] [Order article via Infotrieve]
18. Henriquez CS. Simulating the electrical behavior of cardiac tissue using the bidomain model. Crit Rev Biomed Eng. 1993; 21: 1–77.[Medline] [Order article via Infotrieve]
19. Hunter PJ, Smaill BH, Nielsen PMF, LeGrice IJ. A mathematical model of cardiac anatomy.In: Computational Biology of the Heart. New York, NY: John Wiley & Sons Ltd; 1997: 183–187.
20. Zienkiewicz OC, Taylor RL. The Finite Element Method. New York, NY: McGraw-Hill; 1994.
21. Hunter PJ, McNaughton PA, Noble D. Analytical models of propagation in excitable cells. Prog Biophys Mol Biol. 1975; 30: 99–144.[Medline] [Order article via Infotrieve]
22. Costa KD, May-Newman K, Farr D, O’Dell WG, McCulloch AD, Omens JH. Three-dimensional residual strain in midanterior canine left ventricle. Am J Physiol. 1997; 273: H1968–H1976.[Medline] [Order article via Infotrieve]
23. Beeler GW, Reuter H. Reconstruction of the action potential of ventricular myocardial fibers. J Physiol. 1977; 268: 177–210.[Abstract/Free Full Text]
24. Drouhard JP, Roberge FA. A simulation study of the ventricular myocardial action potential. IEEE Trans Biomed Eng. 1982; 29: 494–502.[Medline] [Order article via Infotrieve]
25. Skouibine KB, Trayanova NA, Moore PK. Anode/cathode make and break phenomena in a model of defibrillation. IEEE Trans Biomed Eng. 1999; 46: 769–777.[CrossRef][Medline] [Order article via Infotrieve]
26. Ellis WS, Auslander DM, Lesh MD. Fractionated electrograms from a computer model of heterogeneously uncoupled anisotropic ventricular myocardium. Circulation. 1995; 92: 1619–1626.[Abstract/Free Full Text]
27. Spach MS, Barr RC, Serwer GA, Kootsey JM, Johnson EA. Extracellular potentials related to intracellular action potentials in the dog Purkinje system. Circ Res. 1972; 30: 505–519.[Abstract/Free Full Text]
28. LeGrice IJ. A Finite Element Model of Myocardial Structure: Implications for Electrical Activation in the Heart [dissertation]. Auckland, New Zealand: University of Auckland; 1992.
29. Plonsey R, Barr RC. Bioelectricity: A Quantitative Approach. New York, NY: Plenum Press; 1988.
30. Durrer D, van der Tweel LH, Blickman JR. Spread of activation in the left ventricular wall of the dog, III: transmural and intramural analysis. Am Heart J. 1954; 48: 13–35.[CrossRef][Medline] [Order article via Infotrieve]
31. Gardner P, Ursell P, Fenoglio J, Wit A. Electrophysiologic and anatomic basis for fractionated electrograms recorded from healed myocardial infarcts. Circulation. 1985; 69: 369–381.
32. Sepulveda NG, Wikswo JP Jr. Electric and magnetic fields from two-dimensional anisotropic bisyncytia. Biophys J. 1987; 51: 557–568.[Abstract/Free Full Text]
33. Trayanova N, Skouibine K. Modelling defibrillation: effects of fiber curvature. J Electrophysiol. 1998; 31: 23–29.
34. Sobie EA, Susil RC, Tung L. Generalized activating function for predicting virtual electrodes in cardiac tissue. Biophys J. 1997; 73: 1410–1423.[Abstract/Free Full Text]
35. Frazier DW, Krassowska W, Chen PS, Wolf PD, Danieley ND, Smith WM, Ideker RE. Transmural activations and stimulus potentials in three-dimensional anisotropic canine myocardium. Circ Res. 1988; 63: 135–146.[Abstract/Free Full Text]
36. Knisley SB. Transmembrane voltage changes during unipolar stimulation of rabbit ventricle. Circ Res. 1995; 77: 1229–1239.[Abstract/Free Full Text]
37. Hooks DA, LeGrice IJ, Harvey JD, Smaill BH. Intramural multisite recording of transmembrane potential in the heart. Biophys J. 2001; 81: 2671–2680.[Abstract/Free Full Text]

This article has been cited by other articles:

				D. A. Pijnappels, M. J. Schalij, A. A. Ramkisoensing, J. van Tuyn, A. A.F. de Vries, A. van der Laarse, D. L. Ypey, and D. E. Atsma Forced Alignment of Mesenchymal Stem Cells Undergoing Cardiomyogenic Differentiation Affects Functional Integration With Cardiomyocyte Cultures Circ. Res., July 18, 2008; 103(2): 167 - 176. [Abstract] [Full Text] [PDF]

				V. Jacquemet and C. S. Henriquez Loading effect of fibroblast-myocyte coupling on resting potential, impulse propagation, and repolarization: insights from a microstructure model Am J Physiol Heart Circ Physiol, May 1, 2008; 294(5): H2040 - H2052. [Abstract] [Full Text] [PDF]

				C. Cabo Modulation of impulse propagation by fibroblasts Am J Physiol Heart Circ Physiol, May 1, 2008; 294(5): H1992 - H1993. [Full Text] [PDF]

				D. A. Hooks, M. L. Trew, B. J. Caldwell, G. B. Sands, I. J. LeGrice, and B. H. Smaill Laminar Arrangement of Ventricular Myocytes Influences Electrical Behavior of the Heart Circ. Res., November 9, 2007; 101(10): e103 - e112. [Abstract] [Full Text] [PDF]

				K. A. MacCannell, H. Bazzazi, L. Chilton, Y. Shibukawa, R. B. Clark, and W. R. Giles A Mathematical Model of Electrotonic Interactions between Ventricular Myocytes and Fibroblasts Biophys. J., June 1, 2007; 92(11): 4121 - 4132. [Abstract] [Full Text] [PDF]

				D. Noble From the Hodgkin-Huxley axon to the virtual heart J. Physiol., April 1, 2007; 580(1): 15 - 22. [Abstract] [Full Text] [PDF]

				R. A.B BURTON, G. PLANK, J. E SCHNEIDER, V. GRAU, H. AHAMMER, S. L KEELING, J. LEE, N. P SMITH, D. GAVAGHAN, N. TRAYANOVA, et al. Three-Dimensional Models of Individual Cardiac Histoanatomy: Tools and Challenges Ann. N.Y. Acad. Sci., October 1, 2006; 1080(1): 301 - 319. [Abstract] [Full Text] [PDF]

				T. M AUSTIN, D. A HOOKS, P. J HUNTER, D. P NICKERSON, A. J PULLAN, G. B SANDS, B. H SMAILL, and M. L TREW Modeling Cardiac Electrical Activity at the Cell and Tissue Levels Ann. N.Y. Acad. Sci., October 1, 2006; 1080(1): 334 - 347. [Abstract] [Full Text] [PDF]

				A. E. Pollard and R. C. Barr Cardiac microimpedance measurement in two-dimensional models using multisite interstitial stimulation Am J Physiol Heart Circ Physiol, May 1, 2006; 290(5): H1976 - H1987. [Abstract] [Full Text] [PDF]

				P. P. Lunkenheimer, K. Redmann, P. Westermann, K. Rothaus, C. W. Cryer, P. Niederer, and R. H. Anderson The myocardium and its fibrous matrix working in concert as a spatially netted mesh: a critical review of the purported tertiary structure of the ventricular mass Eur. J. Cardiothorac. Surg., April 1, 2006; 29(Suppl_1): S41 - S49. [Abstract] [Full Text] [PDF]

				M. L. Trew, B. J. Caldwell, G. B. Sands, D. A. Hooks, D. C.-S. Tai, T. M. Austin, I. J. LeGrice, A. J. Pullan, and B. H. Smaill Cardiac electrophysiology and tissue structure: bridging the scale gap with a joint measurement and modelling paradigm Exp Physiol, March 1, 2006; 91(2): 355 - 370. [Abstract] [Full Text] [PDF]

				N. Trayanova Defibrillation of the heart: insights into mechanisms from modelling studies Exp Physiol, March 1, 2006; 91(2): 323 - 337. [Abstract] [Full Text] [PDF]

				C. W. Zemlin, S. Mironov, and A. M. Pertsov Near-threshold field stimulation: Intramural versus surface activation Cardiovasc Res, January 1, 2006; 69(1): 98 - 106. [Abstract] [Full Text] [PDF]

				P. Hunter and P. Nielsen A Strategy for Integrative Computational Physiology Physiology, October 1, 2005; 20(5): 316 - 325. [Abstract] [Full Text] [PDF]