IOP PUBLISHING JOURNAL OF NEURAL ENGINEERING
J. Neural Eng. 8 (2011) 056010 (8pp) doi:10.1088/1741-2560/8/5/056010
Multielectrode array recordings of
bladder and perineal primary afferent
activity from the sacral dorsal root ganglia
Tim M Bruns1, Robert A Gaunt1 and Douglas J Weber1,2,3,4
1 Department of Physical Medicine and Rehabilitation, University of Pittsburgh, Pittsburgh, PA, USA
2 Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA, USA
3 Department of Veterans Affairs, Pittsburgh, PA, USA
E-mail: tmb59@pitt.edu, rag53@pitt.edu and djw50@pitt.edu
Received 26 April 2011
Accepted for publication 28 July 2011
Published 30 August 2011
Online at stacks.iop.org/JNE/8/056010
Abstract
The development of bladder and bowel neuroprostheses may benefit from the use of sensory
feedback. We evaluated the use of high-density penetrating microelectrode arrays in sacral
dorsal root ganglia (DRG) for recording bladder and perineal afferent activity. Arrays were
inserted in S1 and S2 DRG in three anesthetized cats. Neural signals were recorded while the
bladder volume was modulated and mechanical stimuli were applied to the perineal region. In
two experiments, 48 units were observed that tracked bladder pressure with their firing rates
(79% from S2). At least 50 additional units in each of the three experiments (274 total; 60%
from S2) had a significant change in their firing rates during one or more perineal stimulation
trials. This study shows the feasibility of obtaining bladder-state information and other
feedback signals from the pelvic region with a sacral DRG electrode interface located in a
single level. This natural source of feedback would be valuable for providing closed-loop
control of bladder or other pelvic neuroprostheses.
1. Introduction
Pelvic and pudendal nerves carry sensory signals to
the lumbosacral spinal cord conveying information about the
state of the bladder, urethra and perineal region [1, 2]. These
feedback signals may be useful for providing spinal cord
injured patients information about the fullness of their bladder
or in closed-loop bladder and bowel neuroprostheses [3], a
feature that is currently not available.
Single-unit activity from individual pelvic, pudendal and
sacral root afferent fibers has been recorded in cats and shows
a high correlation to bladder pressure [4–7] as well as urethra,
colon and rectal activity [4, 8, 9]. While these studies provided
insight into the pathways involved with signal transduction
in the pelvic region, their approach of using wire or hook
electrodes to record from individually separated axons is not
readily translatable to neuroprosthetic applications.
More recently, researchers have shown the ability to detect
bladder contractions from cuffs placed around entire sacral
4 Author to whom any correspondence should be addressed.
roots [10] and the pudendal nerve [11] in cats. Intraoperative
human evaluations with nerve cuffs on sacral roots have also
suggested an ability to detect bladder contractions [12]. These
results had low signal-to-noise ratios and were unable to clearly
differentiate non-bladder activity or estimate the bladder
volume. Other cuff electrode approaches have suggested
an ability to track the bladder pressure [13, 14]; however,
they may also be susceptible to interference from non-bladder
afferent activity.
An alternate approach for accessing these afferent signals
is to implant microelectrodes in the dorsal root ganglia
(DRG) where sensory fibers are separated from motor fibers
and converge en route to the spinal cord. The DRG
contain the primary afferent neuronal somata, which are large
(20–150 μm) [15], facilitating recording of large amplitude
action potentials [16, 17]. High density microelectrode arrays
in the lumbar DRG have been shown to provide high quality
recordings of single unit activity from large populations of
hind limb afferents in both acute [18, 19] and chronic [20]
implant scenarios. These studies demonstrated the ability to
1741-2560/11/056010+08$33.00 1 © 2011 IOP Publishing Ltd Printed in the UK

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
Bladder
Urethra
Prepuce
Scrotum
Rectal
Temperature Probe
Anus
Rectum
Volume Syringe
Processor
RZ2 Signal
Arrays in
S1, S2 DRG
Amplifier
Computer
Headstage
Pressure
Transducer
& Amplifier
DRG
Spinal
Cord
Urethra
Catheter
Hand-held
Event Marker
Figure 1. Experimental setup. A urethral catheter was inserted for access to the bladder for fluid infusion and pressure monitoring. A
90-channel microelectrode array (Blackrock ICS-96 MultiPort split array: 50 and 40 electrodes) was inserted in the S1 and S2 DRG on one
side. Sensory stimuli were applied by changing the bladder volume, moving the urethra catheter, pinching the prepuce, brushing the
scrotum, moving the rectal temperature probe and brushing around the anus. Neural signals from the microelectrodes and the bladder
pressure were viewed and stored on a PC after processing with a TDT RZ2 system.
accurately estimate the multi-dimensional state of the lower
limb from the activity of a small population of afferent units
[18–21]. A similar approach may be suitable for obtaining
bladder and perineal sensory information.
The primary objective of this study was to identify and
record neural activity from bladder afferents within sacral
DRG using penetrating microelectrode arrays. The secondary
goal was to identify and record from primary afferent units
from other sources within the pelvic and perineal region.
Both objectives were achieved as neural recordings were
obtained from numerous units that responded to specific
stimuli, demonstrating the feasibility of this approach.
2. Methods
2.1. Experimental setup
Three intact, adult male cats were used in this study, with
all procedures approved by the University of Pittsburgh
Institutional Animal Care and Use Committee. Each animal
was initially sedated with ketamine (10–15 mg kg−1) before
continuous anesthesia with isoflurane (1.5–2.5%). Subsequent
anesthesia maintenance varied between each experiment. In
the first experiment, after completion of all surgical steps,
the animal was transitioned to α-chloralose (70 mg kg−1
initial dose; 20 mg kg−1 maintenance doses) augmented with
low levels of isoflurane (0.1–0.7%). The second animal
was decerebrated by transecting the brainstem anterior to
the superior colliculus at a 50◦ angle to the horizontal,
and isoflurane remained at 1–1.3% for the remainder of the
experiment. The third experiment was transitioned completely
to α-chloralose after completion of all surgical steps
(70 mg kg−1 initial dose; 20 mg kg−1 maintenance doses,
augmented with bupernex 0.01 mg kg−1 every 12 h) without
isoflurane. Variations in the anesthetic regimen are not
expected to affect the sensitivity of the mechanoreceptive
afferents examined in this study as they do not receive efferent
inputs.
After induction of anesthesia, a tracheotomy was
performed and the animal was connected to a ventilator. Vitals
were monitored and maintained within normal physiological
ranges. An intravenous line was inserted in one or both
forelimbs for fluid infusion. A polypropylene catheter (3.5 Fr,
Sovereign) was inserted intraurethrally to the bladder
for controlling the bladder volume and measuring the
bladder pressure (figure 1). A stop cock at the end of the
urethra catheter was used to select pressure monitoring or
volume control. A saline-filled tube for pressure monitoring
was connected to a pressure transducer (DTXPlus DT-4812,
Beckton Dickinson) and transducer amplifier (TA-100, CWE).
Bladder volume control was performed with a refillable
syringe connected to the other output of the stop cock. A
laminectomy was performed to expose the S1–S2 sacral
dorsal root ganglia. The cat was positioned in a custom-
built frame with the torso and pelvis suspended. Penetrating
microelectrode arrays (90 channels: 4 × 10 and 5 × 10
ICS-96 MultiPort split planar arrays, 1 mm shaft length,
0.4 mm interelectrode spacing, Blackrock Microsystems)
were inserted in the S1 and S2 DRG of each cat with a
pneumatic inserter (Blackrock). In the first and second cats, the
electrode tips were coated with sputtered iridium oxide and had
average impedances of 46.9 ± 5.3 k. In the third cat,
2

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
the electrode tips were coated in platinum and had average
impedances of 227.3 ± 91.2 k. In the first cat, the 5 × 10
array was inserted in S1 and the 4 × 10 array was inserted
in S2. In the second and third cats, the arrays were switched.
Neural data from the microelectrode arrays as well as the
analog bladder pressure were recorded using a biopotential
processor for data sampling and storage (RZ2, TDT). Neural
signals were sampled at 25 kHz and bandpass filtered (300–
3000 Hz) before online thresholding to generate spike snippets,
which were stored for offline analysis. The bladder pressure
signal was captured at 100 Hz.
2.2. Experimental protocol
In the second and third experiments, neural activity was
recorded at incremental constant bladder volumes. First, the
bladder was emptied via the urethral catheter. Next, 1 min
long constant bladder volume trials were conducted starting
at 0 mL, and continued at 5–10 mL increments until there
was leakage of the infused saline around the urethra catheter.
Recording commenced in each trial after the bladder pressure
had settled following an infusion of saline.
Several perineal mechanical stimulation trials were
performed in all three experiments: 0.5–1.5 mm shifts in
the position of the urethra catheter and rectal temperature
probe, light brushing around the anus with a q-tip and pinching
of the prepuce. In the second and third experiments, light
brushing of the scrotum with a q-tip was also performed. The
time of application for each sensory stimulus was recorded
with a hand-held trigger button that generated an event marker
in the TDT system.
After completion of all experimental objectives, the
animals were euthanized. The first two animals were given
an intravenous dose of potassium chloride (10 mL of 2 mEq
mL−1) while deeply anesthetized. The third animal was deeply
anesthetized with an intravenous dose of ketamine (15 mg
kg−1) before transcardial perfusion with 4% paraformaldehyde
toward histological analyses related to other experimental
objectives.
2.3. Data analysis
Offline analyses were performed to identify neural units that
responded to each sensory stimulus. Spikes were sorted in
OpenSorter (TDT) using a two-stage process: a k-means
algorithm was first used to generate initial spike clusters and
then each channel was reviewed and manually sorted further,
if necessary, to obtain only units that could be consistently
differentiated across all trials run in an experiment. In
the two experiments with fixed bladder volume trials, the
average number of spikes per second for each unit and the
average bladder pressure was calculated in Matlab for each
bladder volume trial. Units identified during the spike-
sorting process were designated as ‘bladder units’ using
criteria established in previous studies of recordings from
individual pelvic nerve and sacral root filaments. These criteria
include units where spiking was absent when the bladder was
empty and that had an approximately linear increase in the
firing rate at increasing bladder pressures. Bladder afferents
have been shown to have maximum firing rates below 20
spikes per second for bladder pressures below 50 cm H2O
[4, 5, 7]. A similar approach was followed in the first
experiment, using several short 10–15 s periods of constant
bladder pressure.
For the perineal stimulus methods, a different analysis
approach was followed, since these stimuli were in one of
two states (on, off). Spike counts in 1 s bins were used
to estimate the ‘stimulus on’ spike rate per unit for each
trial. For non-continuous stimuli (urethra catheter and rectal
probe movement), 0.5 s on either side of each event marker
were used. For continuous stimuli (anus brushing, prepuce
pinching, scrotum brushing), contiguous 1 s intervals, starting
1 s after and ending 1 s before start/stop event markers,
were used. Spike counts during a corresponding number
of 1 s bins when no stimulus was being applied were also
counted to estimate the ‘stimulus off’ spike rate. At least
20 bins were used for both ‘stimulus on’ and ‘stimulus off’
periods for each trial, with additional bins used when available
from both periods. The stimulus on and off estimated spike
rates for all units within a given trial were compared for
significant difference using the Wilcoxon rank-sum test. The
significance level (α <0.05) was adjusted according to the
Bonferroni correction, to account for multiple comparisons
(i.e. multiple neurons were tested) within each trial. This
typically resulted in significant p-values less than 0.0005 when
identifying units that responded to each stimulus type. Values
for a particular result are given as the average ± standard
deviation.
3. Results
In all experiments, DRG units were identified that responded
to one or more pelvic stimuli. Across the three experiments,
290 units (114 on S1; 176 on S2) were identified that responded
to one or more stimuli, including many units that responded
to multiple stimuli (71%). Table 1 provides a summary by
experiment and trial type of the number of responding units.
Also identified were 69 units (43 on S1; 26 on S2) that did
not show a significant response to any applied stimuli. The
results from each cat were obtained during an approximately
2 h period at the end of each experiment. No differences in the
amplitudes or firing rates of individual units were observed in
this time.
3.1. Bladder units
In the two experiments that included extensive tests of bladder
volume, 11 (Cat 2) and 31 (Cat 3) channels had units that
were linearly related to bladder pressure (table 1). Six units in
each of these experiments (10 of 12 on S2) were well-isolated
units that were quiet at low bladder pressures and fired more
frequently at higher bladder pressures (figure 2). An additional
36 units (28 on S2) were identified that responded to bladder
pressure changes but had non-zero firing rates at low bladder
pressures (figure 3). Thirty-two of these ‘multi-units’ also
responded to one or more of the other applied stimuli and could
not be separated into bladder-responding and non-bladder-
responding units. When the firing rates of these multi-units at
3

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
Table 1. Number of electrode channels (in parentheses—number of units) responding to each stimulus trial for each DRG and experiment.
S1 had 50 electrodes in Cat 1 and 40 electrodes in Cats 2 and 3. S2 had 40 electrodes in Cat 1 and 50 electrodes in Cats 2 and 3.
Cat 1 Cat 2 Cat 3
Trial S1 S2 S1 S2 S1 S2
Urethra catheter 9 (11) 10 (21) 8 (9) 5 (8) 9 (15) 28 (52)
Prepuce pinch 24 (30) 19 (62) 15 (22) 10 (16) 18 (33) 32 (78)
Scrotum brush Not tested 5 (6) 6 (9) 16 (27) 32 (75)
Rectal probe 7 (15) 10 (13) 7 (8) 10 (21) 13 (18) 29 (61)
Anus brush 14 (25) 19 (44) 5 (6) 9 (12) 13 (20) 32 (69)
Bladder pressure 1 (1) 0 (0) 2 (2) 9 (12) 6 (7) 25 (26)
B
la
dd
er
P
re
ss
u
re
(cm
H
2O
)
Time (s) 600
0
30
Im
p
ulses/s
0
12
Im
pu
lse
s/s
0
10
0 45
Im
p
ulses/s
0
10
450
Bladder Pressure (cm H2O)
(a)
(b)
Figure 2. Relationship between firing rate and bladder pressure for
well-isolated bladder units. (a) An example unit from Cat 2 is
shown at a bladder volume of 20 mL with a baseline firing rate of
about one spike per second except for a burst of spikes during a
distension evoked contraction. The small oscillations in the bladder
pressure are due to movement of the abdomen as the diaphragm
expanded with each breath. (b) The average firing rate for a single
bladder unit, at 5 bladder volumes (left), and all bladder units
(n = 6) from the same experiment (Cat 3; right) are shown.
an empty bladder volume were subtracted from the firing rates
at other volumes, they had similar response characteristics to
the well-isolated units that responded to changes in bladder
volume (figure 4). In the experiment that did not have fixed
bladder volume tests (Cat 1), only one bladder multi-unit was
identified. Other bladder units may have been present within
the recordings but were not apparent due to the lack of extended
constant bladder volume trials.
Across all bladder units, the minimum pressure at which
the firing rate was observed to increase by at least one spike
per second from the baseline pressure was 11.5 ± 8.6 cm H2O
(range 3.8–40.6). Ninety-two percent of the observed bladder
units first increased their firing rates at pressures below 20 cm
H2O.
Table 2. Breakdown of 71 units across Cats 1, 2 and 3 that
responded to a single non-bladder stimulus by root.
Root
count
Change in firing rate
(sp s−1)
Trial S1 S2 Average Range Example
Urethra catheter 3 2 7.7 ± 5.2 1.1–15.3
Prepuce pinch 28 15 7.6 ± 12.8 −12.3–75.3
Scrotum brush 0 4 2.3 ± 1.1 1.5–3.8 Figure 5—unit 3
Rectal probe 3 9 1.6 ± 3.6 −8.1–6.3 Figure 5—unit 2
Anus brush 1 6 5.8 ± 4.3 1.1–12.4 Figure 5—unit 5
Table 3. Firing rate responses per trial type for 69 units across
Cats 1, 2 and 3 that responded to all non-bladder stimuli. Average
firing rates during prepuce pinch and scrotum brush were each
significantly different (p <0.05) from all other trials.
Firing rate increase
(sp s−1)
Trial Average Range
Urethra catheter 16.3 ± 11.9 0.8–51.6
Prepuce pinch 60.5 ± 54.4 4.2–238.0
Scrotum brush 31.9 ± 24.7 1.5–111.8
Rectal probe 16.3 ± 17.4 1.6–96.9
Anus brush 23.4 ± 34.5 0.9–229.7
3.2. Non-bladder units
Across the three experiments, 274 DRG units had significant
changes in their firing rates during one or more applied non-
bladder stimuli (110 on S1, 164 on S2; figure 5). Seventy
one (24% of total unit count) of these units responded to a
single non-bladder stimulus. See table 2 for a breakdown of
these single-trial units by type.
The most frequently observed DRG units were non-
specific and had significant changes in their firing rates during
more than one stimulus. Fifty units (17%) responded to
all non-bladder stimuli (33 on S2; figure 5—units 8 and
10). Nineteen of the thirty-six bladder multi-units responded
to bladder pressure and all non-bladder stimuli (17 on S2;
figure 3). See table 3 for the average and the range of firing rate
increases for units that responded to all non-bladder stimuli.
In addition to units that responded to all perineal stimuli,
105 units (36%) responded to two or more trials spanning
the anal and genital regions (65 on S2; figure 5—units 1, 4, 7
and 9).
Some DRG units that responded to more than one trial
had smaller receptive fields. Twenty-five units (13 on S1)
4

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
Time (s)
B
la
dd
er
P
re
ss
u
re
(cm
H
2O
)
1 10
50
0
Bladder Pressure
Time (s)1 10 Time (s)1 10
Im
pu
lse
s/s
30
0
Unit 2
Bladder: 0 mL
Im
pu
lse
s/s
90
60
Unit 1
Bladder: 30 mL Bladder: 38 mL
Bladder Pressure (cm H2O)
0 45
30
0
Across Volumes 90
60
Time (s)
Im
pu
lse
s/s
Unit 2
200
0
250
Rectal Probe
Im
pu
lse
s/s
Unit 1
450
0
Time (s)
Unit 2
200
0
0 55
Scrotum Brush
Unit 1
450
0
Probe Movement Brushing
(c)
(b)(a)
Figure 3. Examples of multi-unit bladder afferent activity responsive to bladder pressure changes and perineal stimulation. Two S2 units
from Cat 3 are shown for (a) movement of the rectal probe, (b) brushing of the scrotum and (c) different bladder volumes. In (a) and (b), the
instantaneous firing rate was calculated at 50 ms intervals by filtering the spike impulses with a linear filter [20] to provide a greater
temporal resolution of the maximum spike rates attained. In (c), spike counts and bladder pressures for 1 s bins are given for three bladder
volumes. On the right, average spikes per bin are shown at five different bladder pressures corresponding to 0, 10, 20, 30 and 38 mL.
responded to two or three of the trials involving urethral
catheter movement, brushing of the scrotum and pinching
of the prepuce (figure 5—unit 6). Four units (three on S2)
responded to only the movement of the rectal probe and
brushing of the anus.
3.3. Non-perineal units
Sixty-nine DRG units were identified that did not have
significantly different spike rates during any applied stimuli.
Among these were 18 units (12 on S1) that were likely
perineal units, having at least one stimulus test with a p-
value < 0.05; however, their change in the estimated spike rate
was not significant after the Bonferroni correction. Among
the remaining non-perineal units, three (two on S1) were
cutaneous or visceral afferent units that had cyclic firing rates,
which directly tracked the respiratory rate of about 12 breaths
per minute. Several more of these cyclic units also responded
to one or more applied sensory stimuli and were included in
those totals. Also observed were 25 units (21 on S1) which
had fairly constant firing rates, ranging from ∼1 spike every
second up to ∼50 spikes every second, that were not affected
by any applied stimuli. These may have been cutaneous units
with receptive fields in direct contact with the support frame or
muscle spindle type II afferents. Twenty-three remaining non-
perineal units (17 on S1) did not have any unique or identifying
characteristics.
4. Discussion
The goal of this study was to examine the feasibility of
recording from sacral primary afferents with receptive fields in
the pelvic region using a multielectrode array in the DRG. In
three experiments, we observed 48 units that were correlated
to bladder pressure and 274 units that responded to one or more
applied perineal stimuli. In general, the majority of bladder
units were found at the S2 level while slightly greater numbers
of non-bladder units, responsive to all stimuli, were found on
5

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
Bladder Pressure (cm H2O) 450
0
20
Im
pu
lse
s/s
(w
ith
re
sp
ec
t to
em
pty
bl
ad
de
r r
ate
)
Well Isolated Bladder Units
Bladder Multi-Units
Figure 4. Changes in bladder unit firing rates against bladder
pressure. The average firing rates for each multi-unit at an empty
bladder volume were subtracted from the average rates at other
volumes. The response characteristics of many of the bladder
multi-units (dashed line) were similar to the firing rate changes
exhibited by well-isolated bladder units (thick gray lines), with
some multi-units having a wider dynamic range. These units are
from Cat 3.
S2 than S1. As the pelvic organs have bilateral innervation,
these results suggest that a single microelectrode array in a
single DRG (S2 in cat) may be sufficient to record afferent
activity from the bladder, genitalia and rectum.
The responses in these experiments were similar to neural
recordings from individual pelvic and sacral root axons under
comparable test conditions [4–7]. Bladder afferents increased
their firing rate from zero when the bladder was empty to 3–
20 spikes per second at higher bladder pressures (figures 2,
4 and 6). The bladder pressures during these tests were
comparable to normal physiological levels [5]. Bladder
Figure 5. Examples of DRG units responsive to different perineal stimuli. At the top, raster plots for ten units (1–10, top to bottom) and the
corresponding event markers are shown for (a) movements of the urethral catheter, (b) pinching of the prepuce, (c) brushing of the scrotum,
(d) movement of the rectal probe and (e) brushing of the anus. Some units were clearly associated with more than one stimuli (units 1, 4,
6–10) while other units responded only to a single test (units 2, 3 and 5). At the bottom, the increase in spike rate during each trial is
normalized to the maximum value across trials for each unit. Black bars indicate units with significantly greater spike counts during the
stimulus, according to a Wilcoxon rank-sum test with a Bonferroni correction (p < 0.0005) for each trial. Gray bars indicate units with
non-significant increases in spike counts for that trial. Units 1–5 were from S2 while units 6–10 were from S1 in Cat 2.
pressures leading to an increase of at least one spike per
second varied between 4 and 41 cm H2O, with most of the
units first responding below 20 cm H2O. This range is in
agreement with studies reporting threshold ranges of 7–20 cm
H2O when recording directly from pelvic nerves or sacral roots
[4, 5, 7]. Bladder afferent firing rates tracked changes in
bladder pressure (figures 2(a), 3(c) and 6) in a linear manner
(figures 2(b) and 4) and could be used to differentiate between
bladder and non-bladder activity (figure 6).
We observed bursts of action potentials for each type
of perineal stimulation (figures 3, 5 and 6). These changes
in firing rates were usually clear, as 93% of the 69 units
responding to all perineal stimuli had a spike rate increase
of at least ten spikes per second in one or more trials and
30% increased at least ten spikes per second in all trials
(example in figure 3(a)). Similar changes in afferent firing
rates have been reported when recording from individual
pelvic [4] and pudendal [9] axons during the same types of
stimuli. Some lumbosacral DRG afferents are dichotomizing
afferents that innervate more than one pelvic organ [22–24].
In addition to recording from these multi-organ afferents, it is
likely that DRG units responding to multiple stimuli included
perineal cutaneous units that had large receptive fields that
were activated in some manner during each of the different
trials. Among these multi-responsive units, pinching the
prepuce and brushing the scrotum led to significantly greater
spike rate responses than other stimuli (table 3). The tip of
the penis, squeezed during pinching of the prepuce, has a
high concentration of afferent fibers [25], which may have
contributed to the higher spike rates. Further studies evaluating
conduction velocities and anatomical specificity would be
6

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
0
20
60
Pr
es
su
re
(c
m
H 2
O
)
100Time (s)
Bladder Pressure
0
16
Im
pu
lse
s/s
S2 Bladder Unit
0
10
Im
pu
lse
s/s
S1 Non-Bladder Unit
0
60
Im
pu
lse
s/s
S2 Non-Bladder Unit
Prepuce Pinching
Figure 6. An example showing how multielectrode array recordings
from the sacral DRG can monitor complex pelvic activity such as
this continence reflex, where reductions in bladder pressure were
evoked by manual stimulation of genital nerves in the penis. The
non-bladder DRG units responded only to the pinching of the
prepuce while the bladder unit modulated its firing rate in response
to bladder pressure fluctuations. Unit firing rates are given as spike
counts per 1 s interval. These units are from Cat 2.
useful in verifying the type and source(s) of each DRG
unit.
As demonstrated here, DRG recordings provide an
opportunity for probing the neurophysiology of the lower
urinary tract and other pelvic functions. Microelectrode
arrays inserted in DRG provide access to hundreds of units
at a location where primary afferent fibers converge upon the
spinal cord. Studies that looked at the responses of individual
sacral afferents with cuff or hook electrodes were limited to
recording from only a few units per experiment [4, 5], a much
lower efficiency than is possible with a DRG microelectrode
approach. Recordings from the sacral DRG could be the
central focus of studies to gain further understanding of pelvic
diseases and neurophysiology, such as investigating changes
in the properties of DRG afferents after spinal cord injury
[26], studying the mechanisms of afferent convergence at
the DRG [22] and evaluating neural connections in cases of
neurogenic lesions [27]. Chronic recording studies performed
with microelectrode arrays implanted in the sacral DRG of
awake, behaving cats, as has been done in lumbar DRG
[19], would provide further insights without the influence of
anesthesia and allow for evaluations of the stability of sacral
DRG recordings, which has yet to be demonstrated.
An application of this DRG microelectrode approach
is the development of neural prostheses integrating sensory
feedback. Recordings from DRG bladder units could be used
to regulate conditional genital nerve stimulation to inhibit the
bladder when the bladder pressure gets high or as soon as
contractions occur [10]. The use of high-density recordings
that simultaneously track bladder and non-bladder units may
yield a more accurate estimation of bladder activity than
non-specific recordings from entire nerve fibers [10–12] that
indicate only gross activity such as bladder contractions or are
susceptible to interference from non-bladder activity. Sacral
DRG units can be sorted to differentiate bladder activity
from perineal activity (figure 6), reducing the occurrence of
incorrect bladder contraction detections. DRG recordings
from non-bladder units may be useful for determining when
the colon is full, flow through the urethra is occurring or
when there is undesirable cutaneous pressure or friction in
the pelvic region. Dichotomizing afferent signals could be
used to provide general state information on the entire pelvic
region while individual-organ responsive units would provide
finer resolution. A sacral DRG interface that tracks some
of these sensations may be instrumental in defecation neural
prostheses, bladder voiding neural prostheses, sexual function
neural prostheses or for detecting the onset of pressure ulcers.
Although these microelectrode arrays penetrate neural
tissue, the potential exists for recording from the surface of
the sacral DRG. An electrode interface that does not penetrate
neural tissue may have fewer undesirable tissue responses and
may be easier to evaluate in human subjects. Cell bodies
within the DRG are packed closely below the epineurium
[28], suggesting that recording electrodes may not need to
penetrate the DRG to record neural activity. We have recently
demonstrated the feasibility of this approach and have recorded
lumbar afferent activity that correlates to limb position and
movement [29]. A similar approach at the sacral level may
lead to a minimally invasive sacral DRG neural interface.
Several factors within this experiment may have led to
an under-representation of the number of primary afferents
that can be recorded with a penetrating microelectrode array.
These tests were conducted at the end of experiments that
evaluated other objectives with the lumbar DRG and used the
same electrode arrays. Repeated use of these electrodes led to
at least 30 non-functioning electrodes on the S2 array in Cat 2
and an unknown number of failures on the other arrays. Thus,
in each experiment, we were recording from many fewer than
90 total electrodes. Also, the effect of an extended period
under anesthesia for each cat, by the time these tests were
performed, may have had a negative impact on the bladder or
neural tissue. Furthermore, significance tests for each non-
bladder unit may not have been accurate representations of
the actual physiology as stimulus event markers were human-
generated and the time counts of spikes were for 1 s intervals, a
much lower resolution than at which the units actually respond
to stimuli. Nonetheless, as suggested by figures 2, 3, 5 and 6,
a simple visual review was sufficient to identify the majority
of afferent units.
7

J. Neural Eng. 8 (2011) 056010 T M Bruns et al
5. Conclusions
This work is the first demonstration of bladder and perineal
primary afferent recordings using microelectrode arrays
implanted in the DRG. This approach may be beneficial for
probing the neurophysiology of the pelvic organs and perineal
region or for obtaining sensory signals for the use in closed-
loop control of bladder and bowel neuroprostheses. Future
experiments will study the underlying physiology further and
evaluate signals from the surface of sacral DRG.
Acknowledgments
The authors thank members of the Rehabilitation Neural
Engineering Lab for their assistance during the animal
experiments (Ingrid Albrecht, Tyler Simpson) and with data
analysis (Chris Ayers, Jim Hokanson, Joost Wagenaar). This
work was supported in part by NIH grants 1R01-EB007749
and 1R21-NS-056136 and TATRC grant W81XWH-07-1-
0716.
References
[1] de Groat W C 2006 Integrative control of the lower urinary
tract: preclinical perspective Br. J. Pharmacol. 147 25–40
[2] Vodusek D 2004 Anatomy and neurocontrol of the pelvic floor
Digestion 69 87–92
[3] Haugland M and Sinkjaer T 1999 Interfacing the body’s own
sensing receptors into neural prosthesis devices Technol.
Health Care 7 393–9
[4] Bahns E, Halsband U and Janig W 1987 Responses of sacral
visceral afferents from the lower urinary tract, colon and
anus to mechanical stimulation Pflug. Arch. Eur. J. Phys.
410 296–303
[5] Habler H, Janig W and Koltzenburg M 1993 Myelinated
primary afferents of the sacral spinal cord responding to
slow filling and distension of the cat urinary bladder
J. Physiol. 463 449–60
[6] Iggo A 1955 Tension receptors in the stomach and the urinary
bladder J. Physiol. 128 593–607
[7] Winter D 1971 Receptor characteristics and conduction
velocities in bladder afferents J. Psychiatr. Res.
8 225–35
[8] Todd J 1964 Afferent impulses in the pudendal nerves of the
cat Exp. Physiol. 49 258–67
[9] Cueva-Rolon R, Munoz-Martinez E, Delgado-Lezama R
and Raya J 1994 The cat pudendal nerve: afferent fibers
responding to mechanical stimulation of the perineal skin,
the vagina or the uterine cervix Brain Res. 655 1–6
[10] Jezernik S, Grill W M and Sinkjaer T 2001 Detection and
inhibition of hyperreflexia-like bladder contractions in the
cat by sacral nerve root recording and electrical stimulation
Neurourol. Urodynam. 20 215–30
[11] Wenzel B J, Boggs J W, Gustafson K J and Grill W M 2005
Detecting the onset of hyper-reflexive bladder contractions
from the electrical activity of the pudendal nerve IEEE
Trans. Neural. Syst. Rehabil. 13 428–35
[12] Kurstjens G A M, Borau A, Rodriguez A, Rijkhoff N J M
and Sinkjaer T 2005 Intraoperative recordings of
electroneurographic signals from cuff electrodes on
extradural sacral roots in spinal cord injured patients
J. Urol. 174 1482–7
[13] Saleh A, Sawan M, Elzayat E A, Corcos J and Elhilali M M
2008 Detection of the bladder volume from the neural
afferent activities in dogs: experimental results Neurol. Res.
30 28–35
[14] Jezernik S, Wen J G, Rijkhoff N J M, Djurhuus J C
and Sinkjaer T 2000 Analysis of bladder related nerve cuff
electrode recordings from preganglionic pelvic nerve and
sacral roots in pigs J. Urol. 163 1309–14
[15] Devor M 1999 Unexplained peculiarities of the dorsal root
ganglion Pain 6 S27–35
[16] Prochazka A, Westerman R A and Ziccone S P 1976
Discharges of single hind limb afferents in the freely
moving cat J. Neurophysiol. 39 1090–104
[17] Loeb G E, Bak M J and Duysens J 1977 Long-term unit
recording from somatosensory neurons in the spinal ganglia
of the freely walking cat Science 197 1192–4
[18] Aoyagi Y, Pearson K G, Stein R B, Branner A and
Normann R A 2003 Capabilities of a penetrating
microelectrode array for recording single units in dorsal
root ganglia of the cat J. Neurosci. Methods 128 9–20
[19] Stein R B, Weber D J, Aoyagi Y, Prochazka A, Wagenaar J,
Shoham S and Normann R A 2004 Coding of position by
simultaneously recorded sensory neurones in the cat dorsal
root ganglion J. Physiol. 560 883–96
[20] Weber D J, Stein R B, Everaert D G and Prochazka A 2007
Limb-state feedback from ensembles of simultaneously
recorded dorsal root ganglion neurons J. Neural Eng.
4 S168–80
[21] Weber D J, Stein R B, Everaert D G and Prochazka A 2006
Decoding sensory feedback from firing rates of afferent
ensembles recorded in cat dorsal root ganglia in normal
locomotion IEEE Trans. Neural. Syst. Rehabil. 14 240–3
[22] Malykhina A P, Qin C, Greenwood-Van Meerveld B,
Foreman R D, Lupu F and Akbarali H I 2006
Hyperexcitability of convergent colon and bladder dorsal
root ganglion neurons after colonic inflammation:
mechanism for pelvic organ cross-talk Neurogastroenterol.
Motil. 18 936–48
[23] Christianson J A, Liang R, Ustinova E E, Davis B M,
Fraser M O and Pezzone M A 2007 Convergence of bladder
and colon sensory innervation occurs at the primary afferent
level Pain 128 235–43
[24] Chaban V, Chriseensen A, Wakamatsu M, McDonald M,
Rapkin A, McDonald J and Micevych P 2007 The same
dorsal root ganglion neurons innervate uterus and colon in
the rat Neuroreport 18 209–12
[25] Johnson R D and Halata Z 1991 Topography and ultrastructure
of sensory nerve endings in the glans penis of the rat
J. Comp. Neurol. 312 299–310
[26] de Groat W C and Yoshimura N 2010 Changes in afferent
activity after spinal cord injury Neurourol. Urodynam.
29 63–76
[27] Wyndaele J 2010 Investigating afferent nerve activity from the
lower urinary tract: highlighting some basic research
techniques and clinical evaluation methods Neurourol.
Urodynam. 29 56–62
[28] Spencer P S, Raine C S and Wisniewski H 1973 Axon
diameter and myelin thickness—unusual relationships in
dorsal root ganglia Anat. Rec. 176 225–43
[29] Gaunt R A, Bruns T M, Crammond D, Tomycz N, Moossy J J
and Weber D J 2011 Single- and multi-unit activity
recorded from the surface of the dorsal root ganglia with
non-penetrating electrode arrays 33rd Annual Int. Conf.
IEEE EMBS (Boston, MA) at press
8