Biomedical Engineering Reference
In-Depth Information
Some circuit techniques can be used to reduce the size of the model's equiv-
alent electric circuit. For instance, if the impedance value between two nodes is
very small in comparison with surrounding impedance values, it can be approxi-
mated with a value of zero. While this generates a small error in the calculation,
it also lumps two circuital nodes together, reducing the rank of the resulting
impedance matrix by one.
In addition to optimizing the model representation, numerical methods used
can be tuned to use less computing resources. In general, since in this type of
simulation the resulting transfer function of the linear system can be represented
as a large sparse symmetric matrix, and taking advantage of the fact that sparse
matrices can be stored in compact data structures, computing storage space can
be saved by solving the linear system using iterative methods that keep the
matrix sparse.
Results
There are several different types of information that can be obtained using a
current spread simulation through the impedance method or similar methods.
One matter of interest is to relate the current density recorded in the ganglion
cell layer of the retina with the particular electrode array geometry and
intensity of injected currents. Further, the effect that the location of the
electrode arrays' current return has on the ganglion cell layer excitation pattern
may be studied with this type of simulation. Models with resolution fine
enough to describe the geometrical characteristics of actual retinal cells can
also be developed. Results below show how current spread simulations can
provide information about the response to excitation by stimulating electrode
arrays.
The current spread simulation provides the quantitative data needed to under-
stand what would be the excitation pattern - and thus possibly the visual pattern -
induced by a particular electrode arrangement and activation pattern.
The simulation results in Figure 15.17 show a transversal cut of a retinal
model excited by an electrode array, and the resulting current densities inside
the ganglion cell layer of the retina. The 4
4 electrode array is composed
of electrodes measuring 75m per side and is backed by a dielectric material.
Individual electrodes are 75m apart. Each electrode is injecting a current of
200A. The resolution of the model is 2m.
The location of the current return changes the current path and affects
the current density in the ganglion cell layer. As in the previous simulation,
Figure 15.18 shows a transversal cut of a retinal model excited by an electrode
array, and the resulting current densities inside the ganglion cell layer of the
retina. The setup of the model is the same as the previous simulation, with
the exception of the current return placement, which is not centered over the
electrode array in this case.
×
 
Search WWH ::




Custom Search