Introduction
Therapeutic interventions with electrical stimulation of the nervous system have a long and successful history. The fundamental feature of clinical applications of neurostimulation is the interaction between the stimulating electrode, its resulting electric field, and the surrounding neural tissue. Decades of research have elucidated many of the basic principles of these interactions, however, ongoing research efforts continue to be necessary to refine the details. The general goals of such studies are to expand scientific understanding on the effects of electric fields on the brain and enable more efficacious stimulation delivery for medical devices. The goal of this chapter is to provide an overview of current knowledge on the interactions between permanently implanted stimulating electrodes and the nervous system. While many of the presented details are from deep brain stimulation (DBS) examples, the general concepts are applicable to all forms of clinical neurostimulation (spinal cord stimulation (SCS), functional electrical stimulation (FES), cochlear stimulation, etc.).
Electrode–brain interface
The interface between a permanently implanted electrode and the brain is a complex entity that can dramatically affect the functioning of the clinical device. An anatomically and electrically stable interface is necessary to allow for the determination of stimulation parameters that can provide consistent therapeutic benefit. This interface is made up of many different components. The basis of this interface is the transfer of charge from the metal electrode to the ionic medium of the brain; thereby establishing a potential gradient from the cathode to the anode . This potential gradient passes through the extracellular space whose electrical conductivity is inhomogeneous, meaning it changes as a function of distance from the electrode, and anisotropic, meaning it changes as a function of direction. Very close to the electrode (within a few hundred microns) the tissue medium is typically made up of glia cells and extracellular matrix proteins that form a low conductivity encapsulation layer around the implanted foreign body. The periphery of this encapsulation layer transitions into the normal brain tissue. These biological features create a complex volume conductor for the generated electric field. This field is then indiscriminately applied to all of the neural tissue surrounding the implanted electrode. The spatial change in the electric field, or the second spatial derivative of the extracellular voltage distribution, along a neural process (e.g. axon or dendrite) directly influences its subsequent polarization . Some neural elements surrounding the electrode will generate action potentials in response to the applied stimulus. Most of these action potentials will be transmitted from their initiation zone to their axon terminals, resulting in synaptic action on their target. This interaction with an end organ, or a larger neural network, underlies the basic therapeutic function of the device. From this perspective, the fundamental goal of a neurostimulation device is to control the release of neurotransmitters.
As neurostimulation device users and engineers, our goal is to tap into the existing nervous system circuitry and rekindle its lost function or modulate its abnormal activity. There exist two basic features we can manipulate to achieve our goal. First is the design/placement of the electrode and, second, is the stimulation parameter settings we choose to apply. The details of selecting/choosing these features are covered elsewhere, but they both dictate the first step in characterizing the effects of neurostimulation – quantifying the electric field.
Electric field
The electric field generated by an implanted electrode is a three-dimensionally complex phenomenon that is distributed throughout the brain. While the fundamental purpose of neurostimulation technology is to modulate neural activity with applied electric fields, historically, much of the device design work and clinical protocols were primarily based on anatomical considerations (i.e. stimulation of a specific brain nucleus). This approach was taken because logical hypotheses could be generated to relate the effects of selectively stimulating a given nucleus to a behavioral outcome. However, without considering the complete system of electrode placement in that nucleus, stimulation parameter settings, electrical characteristics of the electrode, and electrical properties of the surrounding tissue medium, it is impossible to determine if the stimulation effects will be contained in that nucleus or if they will extend to surrounding brain regions . Therefore, the first step in predicting the effects of neurostimulation is to characterize the voltage distribution generated in the brain.
Recently, Miocinovic et al experimentally measured the spatial and temporal characteristics of the voltage distribution generated by DBS electrodes implanted in the brain of a non-human primate. Recordings were made during voltage-controlled and current-controlled stimulation. We found that three factors directly affected the voltage measurements:
- 1.
a voltage drop at the electrode–electrolyte interface
- 2.
inhomogeneity and anisotropy of the tissue medium
- 3.
capacitive modulation of the stimulus waveform.
A relatively dramatic drop in voltage occurs in the transition from the polarization of the electrode contact to the ionic medium. This voltage drop (≈40% for DBS electrodes ) is rooted in the complex non-linear and frequency-dependent reactions that actually take place at the electrode–electrolyte interface . Therefore, it is as if for a 1 V stimulus only ≈0.6 V makes it through the transition from an electron-mediated to ion-mediated polarization.
Once the polarization has been transmitted to the ionic medium, it must transverse the brain’s inflammatory response to the implanted electrode. This collection of cellular infiltrate, protein deposits, and collagen matrices increases the electrode impedance . As a result, the effective strength of voltage-controlled stimulation can be further reduced by encapsulation tissue because the injected current is inversely proportional to electrode impedance . Fortunately, this issue can be alleviated by using current-controlled stimulation .
Whether using voltage-controlled or current-controlled stimulation, the spatial distribution of voltage in the tissue medium will be dictated by the anisotropy and inhomogeneity of the brain. These factors can be estimated from diffusion tensor imaging , and have been used extensively in DBS electric field models . The effects of inhomogeneity are most apparent at and around the ventricles where the relatively low conductivity brain tissue is juxtaposed to the relatively high conductivity fluid cavities. In such regions, the voltage spread is enhanced in the directions of high conductivity. The effects of anisotropy are most apparent at and around large fiber tracts, such as the internal capsule. In such regions, the voltage spread is enhanced parallel to and hindered transverse to the fiber tract.
The above paragraphs provide an overview of the factors that impact the spatial distribution of voltage in the tissue medium, however, temporal aspects of stimulus waveform are also modified by the electrode–brain interface . The implanted pulse generator is designed to generate a stimulus pulse with a specific shape (typically a rectangular pulse of ≈0.1 ms). However, when using voltage-controlled stimulation, the capacitance of the electrode–electrolyte interface results in a non-linear decay of the potential during a voltage-controlled stimulus pulse. On the other hand, a current-controlled waveform is not affected by the interface capacitance, but instead exhibits a non-linear rising of the voltage measured in the brain due to the capacitance of the bulk tissue medium. These modifications to the stimulus waveform decrease the relative excitability of the pulse to the surrounding neural tissue .
Stimulated neural elements
Once the electrode is implanted, stimulus pulse applied, and electric field generated in the tissue medium, the next step is to characterize the neural response. In general, three classes of neurons can be directly affected by electrical stimulation: local cells, afferent inputs, and fibers of passage. Local cells represent neurons that have their cell body in close proximity to the electrode and an axon that projects locally and/or to a different brain region. Afferent inputs represent neurons that project to the region near the electrode and whose axon terminals make synaptic connections with local cells. Fibers of passage represent neurons where both the cell body and axon terminals are far from the electrode, but the axonal process of the neuron traces a path that comes in close proximity to the electrode. Experimental measurements indicate that local cells, afferent inputs, and fibers of passage have similar thresholds for activation . And, local cells can be directly excited by the stimulus (see below) and/or have their excitability indirectly altered via activation of afferent inputs that make synaptic connections on their dendritic arbor (see below).
Cable equation and activating function
Much of our understanding of how a neuron responds to extracellular stimulation has come from computational models. The modeling techniques typically used to predict the neural response to extracellular stimulation date back to McNeal , who was the first to integrate an electric field model and a neuron model to simulate action potential generation. Modern extracellular stimulation modeling still relies on those same two fundamental components:
- 1.
a model of the voltage distribution generated by the stimulating electrode(s), and
- 2.
a model of the neuron(s) being stimulated.
Voltage distribution models range from simple (i.e. theoretical point source electrode in an infinite homogeneous isotropic medium) to complex (i.e. finite element volume conductor with explicit representation of electrode geometry, time dependence, and tissue inhomogeneity/anisotropy) . Irrespective of the voltage distribution model selected, the simulated extracellular potentials (Ve[n]) at the location of individual compartments of neurons in the surrounding tissue medium can be predicted by defining a common coordinate system between the stimulating electrode and the neuron. The neural response to the stimulation can then be simulated with electrical circuits of conductances (Gm[n]) and capacitors (Cm[n]) in parallel, defining the transmembrane voltage . The individual compartments of a single neuron are then connected in series by resistors representing the intracellular resistance (Gi[n]) . Neuron models of this type are commonly referred to as multicompartment cable models. When an extracellular stimulus is applied to the neuron model, the membrane current at compartment n is equal to the sum of the incoming axial currents and the sum of the capacitive and ionic currents through the membrane:
Cm [ n ] ( dVm [ n ] / dt ) + Ii [ n ] = Gi [ n − 1 ] ( Vi [ n − 1 ] − Vi [ n ] + Ve [ n − 1 ] − Ve [ n ] ) + Gi [ n ] ( Vi [ n + 1 ] − Vi [ n ] + Ve [ n + 1 ] − Ve [ n ] )
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
