MONITORING BLADDER
ACTIVITIES IN PARALYSED DOGS:
SYSTEM DESIGN AND ACUTE
EXPERIMENTS
A. Harb, M. Sawan, M-A. Crampon, M. Abdel-Gawad†, M. Elhilali†
École Polytechnique de Montréal, Department of
Electrical and Computer Engineering
Email: harb | sawan@vlsi.polymtl.ca
†
ABSTRACT
Acquiring and processing sensory
bladder signals is an important component to integrate within the stimulator of
a complete urinary implant. This paper presents the results of recording
signals from sacral nerve S2 innervating the bladder. The recordings were
performed during acute experiments on dogs. A new tripolar cuff electrode has
been designed and fabricated for our application. A Shape Memory Alloy Armature
is used to perform the opening and closing of the cuff electrodes around the
nerve. The two end contacts of the electrode are tied together to cancel the
recording of the EMG activity. The bladder is then filled and the signal is
recorded at the same time. A very low noise amplifier (based on the
instrumentation amplifier INA103 from Burr-Brown) is used to amplify the signal
up to 1 Million times. The signal is then on-line bandpass filtered between 400
Hz and 10 kHz. A data acquisition card from Data Translation Inc (DT301) is
used to convert the signal into digital data and to store it in a PC Hard Disk.
The signal is then off-line analyzed.
The analysis with Matlab software
includes 2 steps: signal rectifying and bin-integrating (RBI). The results show
an increasing of nerve activity amplitude with increasing of the bladder
volume. The processed signal will be used as a feedback to the local stimulator
to transfer the information about the filled volume to the patient and to
command the stimulator parameters in incontinent patients.
INTRODUCTION
Recording the electrical nerve signals (electroneurogram
ENG) is necessary to design an autonomous electronic implantable system. Using
the electrical nerve signal as a feedback loop in rehabilitation system was
reported in many papers [[1],[2]]. Recording the ENG is one of
the challenges in such systems. The reason is that ENG has very low amplitude
and ranged between 1 and 10 mV and
the recording electrodes can pick up many other unneeded signals such as
electromyogram (EMG) and artifact signals. In order to design a complete
integrated rehabilitation system to restore the bladder dysfunction [[3]] a
discrete electronic system was designed to record the signal from sacral nerve S2 innervating
the bladder. Once the behavior of that signal with the bladder volume is known,
an integrated circuit can be built to perform the restoration function.
In this paper the discrete
electronic system is described, as well as the results of nerve signals
recording and processing are presented.
SYSTEM DESIGN
Figure 1 shows the block diagram of
the nerve signal acquisition system. It is composed of:
1.
A
tripolar cuff electrodes based on a Shaped Memory Alloy Armature that performs
the opening and closing (function of temperature) of the electrodes around the
nerve. The two end contacts of the electrode are tied together to cancel the
recording of the EMG activity;
2.
A very
low noise, high CMRR instrumentation amplifier (IA). This programmable gain (up
to 106) IA is based on very low noise IA INA103 from Burr-Brown
corp.;
3.
A
band-pass filter [400 Hz-10 kHz] (BPF) built around the universal active filter
UAF42 from Burr-Brown corp.;
4.
A data
acquisition card (DAQ) from Data Translation Inc (DT301) is used to convert the
signal into digital data and to store it in a PC Hard Disk. The acquisition is
performed at 30 kHz with 12 bits resolution.
Figure 1: The
nerve signal recording system.
ACUTE EXPERIMENTS
AND PRELIMINARY RESULTS
The above-described system is used
to record signal from sacral nerve S2 innervating the bladder during acute
experiments on dogs. The dog is anesthetized, the spinal cord is opened and the
sacral nerve is exposed and the tripolar electrode is installed. The ENG was
recorded during the bladder filling for several minutes. The recorded signal is
then processed off-line. The analysis with Matlab software includes 2 steps:
signal rectifying and bin-integrating (at 200 ms).
Figure 2 illustrates the obtained
signal and the resulting bladder volume (i.e. the volume of the urine filled
into the bladder) versus time. These graphs show an increasing in nerve
activity amplitude with increasing of the bladder volume. However, there is
short delay between bladder filling beginning and start of signal increasing.
This effect could be attributed to:
1.
a
certain volume threshold below it the ENG does not respond;
2.
a delay
between the injected volume and the response of the bladder.
A delay after the bladder becomes
full, the RBI signal stops to increase and it can be noticed a weak decreasing
at the last seconds which can be attributed to the urine leakage [[4]].
The peaks are due to various
artifacts and it is easy to eliminate them in an integrated system to find the
right volume of the bladder.
Figure 2: Bladder
volume and processed ENG signal variation with the filling time.
CONCLUSION
The ENG signal from sacral nerve
innervating the bladder was recorded using a tripolar ecuff electrode and a
discrete electronics system. By processing the ENG, it can shown that and
easy-to-use signal could be obtained. It will help us to design an integrated
volume monitoring device. The resulting signal will be used as a feedback to
the local stimulator to transfer the information about the filled volume to the
patient and to control the stimulator parameters in incontinent patients
ACKNOWLEDGMENTS
The authors would
like to express their thanks to the Kidney Foundation of Canada and McIntyre
Animal Center of McGill University for their support during this project.
REFERENCES