- Research
- Open Access
Analytic Modeling of Neural Tissue: I. A Spherical Bidomain
- Benjamin L. Schwartz^{1},
- Munish Chauhan^{1} and
- Rosalind J. Sadleir^{1}Email author
https://doi.org/10.1186/s13408-016-0041-1
© Schwartz et al. 2016
- Received: 3 December 2015
- Accepted: 1 September 2016
- Published: 9 September 2016
Abstract
Presented here is a model of neural tissue in a conductive medium stimulated by externally injected currents. The tissue is described as a conductively isotropic bidomain, i.e. comprised of intra and extracellular regions that occupy the same space, as well as the membrane that divides them, and the injection currents are described as a pair of source and sink points. The problem is solved in three spatial dimensions and defined in spherical coordinates \((r,\theta,\phi )\). The system of coupled partial differential equations is solved by recasting the problem to be in terms of the membrane and a monodomain, interpreted as a weighted average of the intra and extracellular domains. The membrane and monodomain are defined by the scalar Helmholtz and Laplace equations, respectively, which are both separable in spherical coordinates. Product solutions are thus assumed and given through certain transcendental functions. From these electrical potentials, analytic expressions for current density are derived and from those fields the magnetic flux density is calculated. Numerical examples are considered wherein the interstitial conductivity is varied, as well as the limiting case of the problem simplifying to two dimensions due to azimuthal independence. Finally, future modeling work is discussed.
Keywords
- Bidomain
- Analytic modeling
1 Introduction
The purpose of this paper is to model the electric potentials in and around a finite volume of excitable tissue that result from externally applied injection current. Our motivation toward quantitative understanding of the distributed electrophysiology of excitable tissue is due to the emergence of magnetic resonance electrical impedance tomography (MREIT) [1]. The contrast in MREIT—as well as in another MR technique, Electrical Properties Tomography (EPT) [2]—depends on the electrical property distribution throughout the region of interest. Briefly, in an MREIT scan, current is injected into an object in concert with the pulse sequence of an MRI scanner. This current will induce a magnetic field [3] whose distribution throughout the entire region can be captured via the phase component of the reconstructed MR images. Electrical conductivity maps may then be constructed from the phase data using the Laplacian of the z component of the induced magnetic field, \(\nabla^{2} B_{z}\) [4, 5]. MREIT has already shown clinical promise, e.g. lesion characterization [6], but it is the possibility of monitoring brain activity with MREIT [7] that especially motivates this study. If MREIT is to be used to detect neural activity it is useful to estimate the influence of MREIT imaging currents on both active and passive tissues. Therefore, we have constructed from first principles an analytic mathematical model of tissue stimulated by injection currents, not unlike that of an MREIT scan.
Excitable tissues are comprised of cells, discrete units through which electric signals may propagate via action potentials [8]. While many have studied and modeled the behavior of individual cells in both sub- and supra-threshold conditions, it is also very important to understand excitability behavior at the tissue level. This approach has been particularly useful in understanding cardiac activity [9]. The bidomain model [10], a generalization of the cable equation [11], has been employed in this area avoiding the discrete constructs of tissue, assuming instead a continuum of two domains, intra- and extracellular, connected by a membrane and that occupy the same volume [12]. Each domain represents an average, then, of all its individual components. MR imaging also necessarily involves averaging over tissues. If we seek to image neural activity using MREIT it is convenient to use a geometrically simple model to predict changes in these images created by neural activity.
Many authors have modeled excitable tissue with the bidomain equations, choosing the coordinate system that most closely resembles the tissue geometry. In circular cylindrical coordinates, Altman and Plonsey modeled a bundle of nerves as an infinite cylinder in an infinite conducting bath, studying first the steady state [13] and transient stimulation [14]. In the former they incrementally increased the realism of their model, going from an isotropic monodomain to an anisotropic bidomain, while in the latter they investigated the effect of fiber diameter on stimulation and impulse propagation. Henriquez et al. [15–17] and Trayanova et al. [18] investigated the merits of assuming a single fiber vs. a bundle, i.e. bidomain, of fibers when modeling an infinite cylinder of tissue excited by either a disk or a line source. They showed that the single fiber core conductor model is not an unreasonable approximation of the control region of a large bundle of fibers, but loses its validity toward the periphery of the bundle and is entirely unsatisfactory for small bundles. Plonsey and Barr showed in a two dimensional rectangular framework, except for special cases, the bidomain approach to modeling tissue electrophysiology is not a mere generalization of one dimensional cable theory [19, 20]. They found that current flowed very differently in isotropic tissue compared to anisotropic tissue with unequal anisotropy ratios. Roth gave approximate analytic solutions to the problem of bisyncytia with unequal anisotropy ratios [21], using rectangular coordinates. His perturbation method involved expansion in a parameter defined through the anisotropy ratios. He considered two sources: an expanding wave front that was approximated with a step function, and a point source. Trayanova et al. considered the case of bidomain tissue in a uniform electric field, modeling the heart as a sphere of anisotropic tissue with a core of blood [22]. The uniform field meant that they could assume azimuthal independence, leaving only a two dimensional problem in the spherical coordinates r and θ. Heretofore none has studied a three spatial dimension bidomain problem in spherical coordinates.
Our present study is motivated by the need to understand the effect on MREIT images of excitable tissue—specifically, a ganglion excised from the abdomen of a sea slug (Aplysia californica)—affected by injection currents injected through electrodes set into the boundary of its artificial sea water bath [7]. We develop a model that is a dramatic simplification of the actual experiment but which still is novel for its generalization to three spherical dimensions. In and of itself this model will depict basic electrophysiological phenomena and can act as a standard against which numeric simulations such as finite element models (FEM) are held, lending credibility to those in concurrence. Seen in a broader context, this work can serve as the foundation for more and more sophisticated analytic modeling, e.g. nonlinear transmembrane currents and mixed boundary conditions.
In this first study of three dimensional analysis of distributed neural tissue we model the Aplysia abdominal ganglion (AG), known to be electrically coupled by gap junctions [23], as an isotropic bidomain sphere, the artificial sea water bath as an infinite isotropic conducting medium, and the injection currents as source as sink points. We assume isotropic conductivity here for simplicity. However, anisotropy may be the subject of future work, as active tissue is generally anisotropic.
2 Problem Formulation
2.1 Geometry
2.2 Bidomain Tissue
2.3 Infinite Medium
3 Solutions
3.1 Transmembrane Potential
3.2 Monodomain Potential
3.3 External Potential
4 Boundary Conditions
5 Current Densities
6 Magnetic Flux Density
7 Numeric Examples and Discussion
Modeling inputs
Parameter | Value |
---|---|
Bath resistivity, \(\rho_{e}\) | 0.29 Ωm |
Intracellular resistivity, \(\rho_{i}\) | 0.19 Ωm |
Membrane resistance times unit area, \(R_{m}\) | 0.15 Ωm^{2} |
Ratio of surface area to volume, β | 20,000 m^{−1} |
Source and sink magnitude, \(I_{0}\) | 1 mA |
Tissue radius, a | 2 mm |
Point source position, \(\mathbf {p}_{+} \) | \((5,\frac{\pi}{2},0)\) |
Point sink position, \(\mathbf {p}_{-} \) | (5,π,0) |
Summation upper bound, μ | 10 |
In this article we set out to model the electromagnetic fields in and around a volume of neural tissue stimulated by current that is injected in close proximity to it. The geometry selected is expected in an in vitro MREIT scan, where an AG may be submerged in a bath of artificial seawater contained in a cylindrical sample chamber that has injection current ports on opposing sides [7]. We note that the effect of the applied external field is to simultaneously depolarize and hyperpolarize portions of the simulated tissue nearby the current sources. If a portion of tissue is sufficiently depolarized to form an action potential it may propagate throughout the tissue from these regions. It has been suggested [36] that modest depolarizations or hyperpolarizations caused by weak external currents applied to the skull are sufficient to excite or inhibit neural excitability in brain structures. More complex models of the tissue and field geometry used here may prove useful methods of exploring the mechanisms of such neuromodulation techniques. In these numeric examples we have held the source and sink points to be equidistant from the sphere of tissue, \(r_{+}=r_{-}\). This gives the problem a symmetry about the lines \(y=-x\) in Figs. 2, 3, and 4 and \(y=0\) in Fig. 5. However, the AG is smaller than the diameter of the sample chamber; so, it will not necessarily be directly between the ports, spoiling this axial symmetry. Thus a complete three dimensional treatment of this type of problem is finally required.
From Eqs. (4a)-(4b) we can see that each region, intracellular and extracellular, has a monodomain component ψ that obeys the Laplace equation. Plots of this potential produce results similar to those of Rush and Driscoll [37, 38]. They solved for the electric potential in a brain from electrodes placed directly on a scalp, modeling the brain, skull, and scalp as different layers of monodomain tissues, i.e. a sphere encased in a thin shell of bone which was itself encased in a thin shell of skin. We could amend our model to include similar surrounding layers, each with its own expressions for ϕ and J and coefficients determined from the boundary conditions. The boundary conditions themselves would change, e.g. the interstitium would have continuity of potential and normal current with the skull rather than with artificial sea water. Such changes would be appropriate for a model on the scale of e.g. a dog’s head [39].
We have modeled both domains as being ohmic, i.e. their impedivities \(z=\rho\) are only real valued, but z can be made complex by introducing a frequency dependence [34]. In their extensive literature review [40] and experimental measurements [41], Gabriel et al. show that most tissues have frequency dependent electrical properties. More recently, Bédard et al. [42] and Bazhenov et al. [43] explored frequency dependence in local field potentials. In a series of theoretical papers Bédard and Destexhe provide a general framework for modeling electromagnetic fields in brain tissue without assuming the interstitium to be purely resistive. Absent those assumptions, they developed a generalized formalism of current source density analysis with the goal of relating the extracellular potential to current sources in the tissue [44], considering monopolar sources, dipolar sources, and combinations thereof. Next they incorporated frequency dependent extracellular and intracellular impedivities, \(z_{o}\) and \(z_{i}\), to generalize the cable theory [11] for neurons embedded in a complex interstitium [45]. They showed that \(z_{o}\) and \(z_{i}\) have a non-trivial impact on the properties of neurons, e.g. voltage attenuation with distance and the spectral profile of \(V_{m}\). Finally, they calculated the magnetic fields generated from a current-carrying neuron and, using superposition, a population of neurons [46]. They showed that since the electrical properties of neural tissue impact the transmembrane and axial currents of a neuron, they will necessarily also impact the magnetic fields these currents create. By contrast, in our study we have concerned ourselves with the interaction of neural tissue and an aphysiologic stimulus. The effects of this stimulus will naturally also depend upon complex tissue properties, but over a larger scale determined by the stimulus geometry. Future work should explore the impact non-ohmic impedivities have on a tissue interactions with applied external fields.
Let us now consider possible next steps to build on this first study. As well as modeling tissue properties as complex, it should also be possible to examine the transient behavior of excitable tissue using a Hodgkin–Huxley-like model [47]. These analytical models of spheres could then be validated and used to estimate the scale of changes expected in MREIT images due to different neural activity patterns.
Declarations
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Authors’ Affiliations
References
- Seo JK, Woo EJ. Magnetic resonance electrical impedance tomography (MREIT). SIAM Rev. 2011;53:40–68. MathSciNetView ArticleMATHGoogle Scholar
- Zhang X, Liu J, He B. Magnetic-resonance-based electrical properties tomography: a review. IEEE Rev Biomed Eng. 2014;7:87–96. View ArticleGoogle Scholar
- Moon P, Spencer DE. Foundations of electrodynamics. Princeton: Van Nostrand; 1960. Google Scholar
- Seo JK, Woo EJ. Electrical tissue property imaging at low frequency using MREIT. IEEE Trans Biomed Eng. 2014;61:1390–9. View ArticleGoogle Scholar
- Seo JK, Woo EJ, Katscher U, Wang Y. Electro-magnetic tissue properties MRI. London: Imperial College Press; 2014. View ArticleGoogle Scholar
- Oh TI, Jeong WC, McEwan A, Park HM, Kim HJ, Kwon OI, Woo EJ. Feasibility of magnetic resonance electrical impedance tomography (MREIT) conductivity imaging to evaluate brain abscess lesion: in vivo canine model. J Magn Reson Imaging. 2013;38:189–97. View ArticleGoogle Scholar
- Sadleir RJ, Grant SC, Woo EJ. Can high-field MREIT be used to directly detect neural activity? Theoretical considerations. NeuroImage. 2010;52:205–16. View ArticleGoogle Scholar
- FitzHugh R. Mathematical models of excitation and propagation in nerve. In: Schwan HP, editor. Biological engineering. New York: McGraw-Hill; 1969. p. 1–85. Google Scholar
- Sobie EA, Susil RC, Tung L. A generalized activating function for predicting virtual electrodes in cardiac tissue. Biophys J. 1997;73:1410–23. View ArticleGoogle Scholar
- Malmivuo J, Plonsey R. Bioelectromagnetism: principles and applications of bioelectric and biomagnetic fields. New York: Oxford University Press; 1995. View ArticleGoogle Scholar
- Rall W, Agmon-Snir H. Cable theory for dendritic neurons. 2nd ed. In: Koch C, Segev I, editors. Methods in neuronal modeling: from ions to networks. Cambridge: MIT Press; 1988. p. 27–92. Google Scholar
- Plonsey R. The use of a bidomain model for the study of excitable media. Lect Math Life Sci. 1989;21:123–49. MathSciNetMATHGoogle Scholar
- Altman KW, Plonsey R. Development of a model for pint source electrical fibre bundle stimulation. Med Biol Eng Comput. 1988;26:466–75. View ArticleGoogle Scholar
- Altman KW, Plonsey R. Point source nerve bundle stimulation: effects of fiber diameter and depth on simulated excitation. IEEE Trans Biomed Eng. 1990;37:688–98. View ArticleGoogle Scholar
- Henriquez CS, Trayanova N, Plonsey R. Potential and current distributions in a cylindrical bundle of cardiac tissue. Biophys J. 1988;53:907–18. View ArticleGoogle Scholar
- Henriquez CS, Plonsey R. Simulation of propagation along a cylindrical bundle of cardiac tissue–I: mathematical formulation. IEEE Trans Biomed Eng. 1990;37:850–60. View ArticleGoogle Scholar
- Henriquez CS, Plonsey R. Simulation of propagation along a cylindrical bundle of cardiac tissue–II: results of simulation. IEEE Trans Biomed Eng. 1990;37:861–75. View ArticleGoogle Scholar
- Trayanova N, Henriquez CS, Plonsey R. Limitations of approximate solutions for computing the extracellular potential of single fibers and bundle equivalents. IEEE Trans Biomed Eng. 1990;37:22–35. View ArticleGoogle Scholar
- Plonsey R, Barr RC. Current flow patterns in two-dimensional anisotropic bisyncytia with normal and extreme conductivities. Biophys J. 1984;45:557–71. View ArticleGoogle Scholar
- Barr RC, Plonsey R. Propagation of excitation in idealized anisotropic two-dimensional tissue. Biophys J. 1984;45:1191–202. View ArticleGoogle Scholar
- Roth BJ. Approximate analytical solutions to the bidomain equations with unequal anisotropy ratios. Phys Rev E. 1997;55:1819–26. View ArticleGoogle Scholar
- Trayanova N, Roth BJ, Malden LJ. The response of a spherical heart to a uniform electric field: a bidomain analysis of cardiac stimulation. IEEE Trans Biomed Eng. 1993;40:899–908. View ArticleGoogle Scholar
- Bodmer R, Verselis V, Levitan IB, Sprat DC. Electronic synapses between Aplysia neurons in situ and in culture: aspects of regulations and measurements of permeability. J Neurosci. 1988;8:1656–970. Google Scholar
- Moon PH, Spencer DE. Field theory handbook: including coordinate systems, differential equations, and their solutions. Berlin: Springer; 1981. MATHGoogle Scholar
- Plonsey R, Heppner DB. Considerations of quasistationarity in electrophysiological systems. Bull Math Biophys. 1967;29:657–64. View ArticleGoogle Scholar
- Tung L. A bi-domain model for describing ischemic myocardial D-C potentials. Ph.D. dissertation. Cambridge, MA, USA: Massachusetts Institute of Technology; 1978. Google Scholar
- Plonsey R. The nature of sources of bioelectric and biomagnetic fields. Biophys J. 1982;39:309–12. View ArticleGoogle Scholar
- Ben-Menahem B, Singh SH. Seismic waves and sources. New York: Springer; 1981. View ArticleMATHGoogle Scholar
- Frannklin P. Basic mathematical formulas. In: Menzel DH, editor. Fundamental formulas in physics. New York: Dover; 1960. p. 1–106. Google Scholar
- Roth BJ. A comparison of two boundary conditions used with the bidomain model of cardiac tissue. Ann Biomed Eng. 1991;19:669–77. View ArticleGoogle Scholar
- Moon PH, Spencer DE. Vectors. Princeton: Van Nostrand; 1965. Google Scholar
- Bell WW. Special functions for scientists and engineer. Mineola: Dover; 2004. Google Scholar
- Smythe WR. Static and dynamic electricity. New York: McGraw-Hill; 1939. MATHGoogle Scholar
- Sneddon IN. Fourier transforms. Mineola: Dover; 1995. MATHGoogle Scholar
- Schwartz BL, Sadleir RJ. Analysis of bipolar external excitation of spherical tissue by spatially opposed current source and sink points. In: Conf proc IEEE eng med biol soc, 37th, Milan. 2015. p. 2299–302. Google Scholar
- Nitsche MA, Paulus W. Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. J Physiol. 2000;527:633–9. View ArticleGoogle Scholar
- Rush S, Driscoll DA. Current distribution in the brain from surface electrodes. Anesth Analg. 1968;47:717–23. View ArticleGoogle Scholar
- Rush S, Driscoll DA. EEG electrode sensitivity–an application of reciprocity. IEEE Trans Biomed Eng. 1969;16:15–22. View ArticleGoogle Scholar
- Kim YT, Lee BI, Woo EJ, Kim HJ. In vivo MREIT experiment of canine brain. In: Intl conf electr bioimp, 13th, Graz. 2007. p. 512–5. Google Scholar
- Gabriel C, Gabriel S, Corthout E. The dielectric properties of biological tissues: I. Literature survey. Phys Med Biol. 1996;41:2231–49. View ArticleGoogle Scholar
- Gabriel S, Lau RW, Gabriel C. The dielectric properties of biological tissues: II. Measurements in the frequency range 10 Hz to 20 GHz. Phys Med Biol. 1996;41:2251–69. View ArticleGoogle Scholar
- Bédard C, Kröger H, Destexhe A. Does the \(1/f\) frequency scaling of brain signals reflect self-organized critical states? Phys Rev Lett. 2006;97:118102. View ArticleGoogle Scholar
- Bazhenov M, Lonjers P, Skorheim S, Bédard C, Destexhe A. Non-homogeneous extracellular resistivity affects the current-source density profiles of up–down state oscillations. Philos Trans R Soc A, Math Phys Eng Sci. 2011;369:3802–19. MathSciNetView ArticleMATHGoogle Scholar
- Bédard C, Destexhe A. Generalized theory for current-source density analysis in brain tissue. Phys Rev E. 2011;84:041909. View ArticleGoogle Scholar
- Bédard C, Destexhe A. Generalized cable theory for neurons in complex and heterogeneous media. Phys Rev E. 2013;88:022709. View ArticleGoogle Scholar
- Bédard C, Destexhe A. Generalized cable formalism to calculate the magnetic field of single neurons and neuronal populations. Phys Rev E. 2014;90:042723. View ArticleGoogle Scholar
- Hodgkin AL, Huxley AF. A quantitative description of membrane current and its application to conduction and excitation in nerve. J Physiol. 1952;117:500–44. View ArticleGoogle Scholar
- Arfken G. Mathematical methods for physicists. Orlando: Academic Press; 1985. MATHGoogle Scholar