Content uploaded by Eric P Knott
Author content
All content in this area was uploaded by Eric P Knott on Feb 12, 2014
Content may be subject to copyright.
30 IEEE PUL SE
▼
JANUARY/FEB RUARY 2012
By Wolfgang J. Streit,
Qing-Shan Xue,
Abhishek Prasad,
Viswanath Sankar,
Eric Knott, Aubrey Dyer,
John R. Reynolds,
Toshikazu Nishida,
Gerald P. Shaw,
and Justin C. Sanchez
2154-2 287/12/$31. 00©2012 I EEE
Electrode Failure
Date of publi cation: 6 Februar y 2012
The development of invasive, rehabilitative neuropros-
thetics for humans requires reliable neural probes that
are capable of recording large ensembles of neurons for
a long period of time.
Recent advances in the development of neuropros-
thetics i n animals and humans have shown that com-
munication and control can be directly derived from the cen-
tral nervous system (CNS) for restoring lost motor ability [1].
This proof of concept has opened the possibility of new thera-
pies for the millions of individuals suffering from neurological
disorders of the nervous system. The success of these therapies
hinges on the ability to reliably access the relevant signals from
the brain with high quality for the lifetime of the patient. As a
result, research has focused on the cascade of events that follow
chron ic implantation of microelec trodes a nd temporal degrada-
tion in the signal and electrode quality: signal-to-noise ratio,
noise floor, peak amplitude, and neuronal yield. Implanted mi-
cr oe lectrodes have b een rep or ted to suffer f rom ti me -dependent
degradation in signal quality due to unknown issues related to
tissue interfaces.
The long-term reliability of microelectrode recording im-
plants has been correlated with two broad categories of factors:
biotic (issues related to the electrode itself) and abiotic (issues
related to the neural tissue). To conduct a comprehensive evalu-
ation of both abiotic and biotic aspects of electrode performance,
we have assembled a team of investigators composed of neu-
robiologists, biomedical and electrical engineers, and electro-
chemists working toward elucidating key mechanisms involved
in chronic microelectrode implantation. Previous studies in this
field have stressed, on the abiotic side, physical changes such as
damaged insulation, change in surface area, and oxidation and
Tissue, Electrical, and
Material Responses
Digital Obj ect Identifie r 10.1109/MPUL. 2011.2175632
JANUARY/FEB RUARY 2012
▼
IEEE PUL SE 31
corrosion of electrodes; and on the biotic side, neuroinflamma-
tion, blood brain barrier disruption, tissue encapsulation, and
astroglial scarring [2]. We will build on and ex-
pand these prior studies by performing coupled
biotic–abiotic investigations. These include scan-
ning electron microscopy of electrodes both pre-
implantation and postimplantation, monitoring
changes in electrode impedance pre, post, and
during implantation, and measuring a number of
biological markers of injury in the cerebral spinal
fluid (CSF) and serum. These combined metrics
support our idea that degeneration of microg-
lial cells is a cellular event that occurs following
prolonged implantation and can be critically tied
to the degradation of neuronal signals. Because
microglial degeneration represents a relatively
recent concept in neuropathology [3] and still
has not been studied in the current context,
we briefly outline our rationale for postulating that microglial
degeneration is a pivotal cellular event in electrode failure.
Microglial Activation Versus
Degeneration for Neural Electrodes
In flammation i s defi ned as the cel lular response to tissue injury,
and the key cellular element involved in neuroinflammation
following CNS injury is the microglial cell. In short, microglia
constitute a population of endogenous CNS glial cells that are
neuroprotective and can perform immunological functions, i.e.,
they comprise the brain’s immune system [4]. Because microglia
are ubiquitous throughout the CNS, any acute injury quickly
triggers the activation of local microglia whose main task is to
restore tissue homeostasis and orchestrate the wound healing
process. Clearly, the implantation of electrodes into the living
brain constitutes an injury, and thus previous studies have re-
ported microglial activation as a conspicuous cellular event [2],
[5]. The problem that arises with long-term implantation of
electrodes is that the injurious stimulus does not go away, and
therefore activation of microglial cells endures and leads to a
chronic neuroinflammatory response. Although little is known
about the fate of microglia persisting in such a chronic state of
activation, there is reason to believe that long-term activation
eventually takes its toll and causes m icroglial cells to degenerate
[6]. This idea primarily stems from prior work done in neurode-
generative diseases such as Alzheimer’s disease and amyotrophic
lateral sclerosis, which have shown a direct correlation between
microglial and neuronal degeneration [3]. Thus, we believe that
degeneration and loss of microglia results in loss of neuroprotec-
tion that contributes to the onset of neurodegenerative changes,
which in the case of chronically implanted recording electrodes
becomes manifest in a loss of neuronal signal strength and qual-
ity. Based on this hypothesis, the main (biotic) goals of this proj-
ect are to 1) ascertain the time line of microglial degeneration, 2)
define molecular mechanisms that cause microg-
lial degeneration, and 3) find ways of interfering
with these mechanisms, thereby preventing both
microglial and neuronal damages and thus ex-
tending the working life of implanted electrodes.
Experimental Test Beds
In addition to the biotic goals, the observed
abiotic effec ts, which include the electrode’s
physical changes (damaged insulation, change
in sur face area, oxidation/reduction, and cor-
rosion) [7], will also impact the electrical
recordi ng properties of the neural probe fol-
lowing prolonged ex posure to the brain mi-
croenvironment in vivo. Although many of
these effec ts are typically studied individual-
ly, a deeper understanding of the true nature of the cause(s)
of electrode failure likely requires the coupling of both biotic
and abiotic perspectives. This combined scientific perspec-
tive upon which we have based our work may enable us to
explain the underlying mechanisms governing the effects
of chronic electrode implantation as well as derive from the
quantification of the various parameters new predictions
that would improve electrode performance in the future.
© BRAND X PICTURES
Recent advances in
the development of
neuroprosthetics in
animals and humans
have shown that
communication and
control can be derived
directly from the
CNS for restoring lost
motor ability.
32 IEEE PUL SE
▼
JANUARY/FEB RUARY 2012
Thus, results from a multifaceted approach can provide a
more complete understanding of the effects of biotic–abiotic
interactions [8]. To better understand the spatiotemporal
responses that occur at the most fundamental levels, it is
necessar y to develop data acquisition methods that best cap-
ture the biochemical, structural, and electrophysiological
responses in real time. A key problem i n investigating the
failure of chronic microelectrodes is the temporal resolution
FIGURE 1 Coupling of biotic and abiotic metrics throughout the lifetime of electrodes.
Quantitative Time Line for In Vivo Studies
Acute Recovery Chronic Failure
Reported Electrophysiological Signal Quality
Simultaneous Electrophysiology, Histopathology, and Microdialysis
(Dots Indicate Sacrifice Points)
Surgery
Impact
Inflammation
Impact
Chronic Impact Between Neural Matrix
and Electrodes
Aggregation of Metrics for
Modeling and Statistical
Analysis to Rank Importance
of the Following:
Abiotic Abiotic Abiotic Biotic
Biotic
Biotic
• Geometry
• Impedance
• Signal-to-
Noise (SNR)
• Impedance
• SNR
• Insulation Failure
• Recording Site
Failure
• Impedance
• SNR
• Microglia Numbers
• Neuron Numbers
• Morphometry
• Biomarkers of
Injury
• Neuronal, Axonal
Integrity
• Region of Interest
• Effective Rate of Failure
• Variance in Failure
• Linear–Nonlinear
Relationship Among
Factors
• Cells of Interest
• Signaling Molecules of
Interest
• Electrode Integrity
• BBB
• Biomarkers
of Injury
• Impedance
• BBB
• Biomarkers
of Injury
• Synaptic
Changes
Hours 1–14 Days Time
1, 6, 12, and 18 Months
FIGURE 2 Scanning electron microscopy of electrode recording sites, (a) preimplant and (b) postimplant. Note that the recording tip
has undergone a transformation in its morphology. In addition, the amount of tungsten relative to the polyimide insulation has
been reduced. Implant duration here is seven days.
(a) (b)
JANUARY/FEB RUARY 2012
▼
IEEE PUL SE 33
required to assess chronic performance. In an
effort to solve this problem, we have devel-
op ed t est b eds t hat e n a ble the rea l-ti me c olle c-
tion of abiotic/biotic metr ics and bring to bear
new probing capabilities. Included in these
metrics is the essential evaluation of neuronal
function, which is characterized with respect
to the electrode–tissue interaction.
Figure 1
provides a summary of metrics and shows a
time line of key phases known to be associ-
ated with electrode failure: acute, recovery,
chronic, and failure. Figure 1 is an average
representation of identifiable phases useful for guiding our
investigation.
This test bed that we have developed has the capability
to support large parallel studies of electrode failure mecha-
ni sms with pr ec ise ti me re so lution. I n add it io n to the tiss ue
reactions, there is a strong possibility that this test bed can
yield insight on microstructural features of electrodes that
evoke tissue reactions leading to failure. I n these studies,
we are investigating arrays of metal microwires. As shown
in
Figure 2
, surface modification of polyim ide-insulated
tungsten microwires (50 μm diameter) not only alters the
electrical properties, but also changes the surface area and
adhesive properties. The changes shown in Figure 2 are
representative of short implant durations (seven days) and
likely a contributing factor of failure for long-term implan-
tation (years). As electrode surface chemistry changes, it’s
affinity for adhesive molecules in the extracellular fluid
also changes. Electrode corrosion (physical and chemical)
and surface adhesiveness (biochemical and biological) can
be effectively evaluated. Every electrode that enters into
this study undergoes a quantitative preimplantation i n-
vestigation to show how certain electrode characteristics
correlate with the failure rate in vivo. By comparing these
rates, as in the pre- and postimages of Figure 2, we can
determine the corrosive feature responsible (surface charge
or porosity), demonstrate the adhesive molecules that bind
the corroded electrode, and then show which cell type pref-
erential ly attaches to the fai lure electrode to result in func-
tional changes. These features become predictive of failure
and can be correlated with the in vivo analyses of perielec-
trode cellular reactions.
Models to Predict Electrode Health
One of the hallmarks of this work is that a variety of new failure
markers and measurement test beds are being developed. With
these data, it is possible to systematically aggregate and evalu-
ate these metr ics in terms of their abil ity to pred ict electrode
health. These models will accept as inputs real-time biomark-
ers, impedance, and corrosion measurements, and these met-
rics will be used to predict thresholds for signal-to-noise-ratio,
electrode yield, and action potential peak-to-peak amplitude.
With these models, it will be possible to estimate the transi-
tion probabilities between specific neuronal marker states and
among the markers themselves to indicate the mechan isms of
action of electrode–tissue interactions. The knowledge gained
from these studies will help to elucidate the de-
sign of future neural electrodes that are capable
of high-performance function for long duration.
Acknowledgment
This work was sponsored by the Defense Ad-
vanced Research Projects Agency Microsystems
Technology Office under the auspices of Dr. Jack
Judy (jack.judy@darpa.mil) through the Space
and Naval Warfare Systems Center, Pacific Grant
N66001-11-1- 4 0 0 9
.
The views, opinions, and/or findings con-
tained in this article are those of the author and should not be
interpreted as representing the official views or policies, either
expressed or implied, of the Defense Advanced Research Projects
Agency or the Department of Defense.
Wolfgang J. Streit (pschorr@ufl.edu), Qing-Shan Xue (qsxue@
ufl.edu), and Gerald P. Shaw (shaw@mbi.uf l.edu) are with
the Department of Neuroscience, University of Florida, Gainesville,
Florida. Toshikazu Nishida (nishida@ufl.edu) and Viswanath
Sankar (vsanka1@ufl.edu) are with the Department of Electri-
cal and Computer Engineering, University of Florida, Gainesville,
Florida. Eric Knott (eknott@chem.ufl.edu), Aubrey Dyer (au-
breydyer@gmail.com), and John R. Reynolds (reynolds@chem.
ufl.edu), are with the Department of Chemistry, University of Flor-
ida, Gainesville, Florida. Abhishek Prasad (apr009@gmail.com)
is with the Neuroprosthetics Research Laboratory at the University
of Miami, Florida. Justin C. Sanchez (jcsanchez@ miami.edu)
is with the Departments of Biomedical Engineering, Neuroscience,
and the Miami Project to Cure Paralysis, University of Miami, Coral
Gables, Florida.
References
[ 1] R. A. Normann, “Technology Insight: Future neuroprosthetic therapies
for disorders of the nervous system,” Nature Clin. Pract. Neurol., vol. 3,
no. 8 , pp. 444–452, 2007.
[ 2] V. S. Polikov, P. A. Tresco, a nd W. M. Reichert, “Response of brain
tissue to c hron ical ly impla nted neural elect rodes,” J. Neurosci.
Methods, vol. 148, pp. 1–18, Oct. 2005.
[ 3] W. J. Streit and Q. S. Xue, “ Life and death of m icrog lia,” J Neuroim-
mune Pharmacol., vol. 4, pp. 371–379, Dec. 2009.
[ 4] W. J. Streit and C. A. Kincaid-Colton, “The brain’s immune sys-
tem,” Sci. Amer., vol. 273, no. 5, pp. 54– 61, 1995.
[ 5] B. D. Winslow and P. A. Tresco, “Quantitative analysis of the tis-
sue response to chronically implanted microwire electrodes in rat
cortex,” Biomaterials, vol. 31, no. 7, pp. 1558–1567, Mar. 2010.
[ 6] M. B. Graeber and W. J. Streit, “Microglia: Biology and pathol-
ogy,” Acta Neuropathol., vol. 119, no. 1, pp. 89–105, Jan. 2010.
[ 7] L. A. Geddes and R. Roeder, “Criteria for the select ion of materi-
als for implanted elec trodes,” Ann. Biomed. Eng., vol. 31, no. 7, pp.
879– 890, June 2003.
[ 8] T. W. Berger, J. K. Chapin, G. A. Gerhardt, D. J. McFarland, J. C.
Pri ncipe, W. V. Soussou, D. M. Taylor, a nd P. A. Tresco, “Interna-
tional assessment of research and development in brain –computer
interfaces,” Baltimore, M D, WT EC Panel Rep., 2007.
Degeneration and
loss of microglia
results in a loss of
neuroprotection
that contributes
to the onset of
neurodegenerative
changes.
