PC SUPPORTED EIGHT CHANNEL SURFACE STIMULATOR
FOR
PARAPLEGIC WALKING - FIRST RESULTS
M. Bijak, C. Hofer, H. Lanmüller, W.
Mayr, S. Sauermann, E. Unger, *H. Kern
Department of Biomedical Engineering and Physics, University of Vienna,
Austria
*Department of physical medicine and rehabilitation, Wilhelminenspital,
Vienna, Austria
Today
Functional Electrical Stimulation (FES) is used amongst others to restore hand
and arm function, to restore mobility of the lower extremities, for phrenic
pacing and in cardiomyoplasty. Common to all FES-applications is that they
require careful set-up of stimulation parameters. To improve these tasks a
PC-based software for stimulation parameter evaluation and data acquisition was
written. First the described software was used to mobilise paraplegic patients
in conjunction with an I2C bus controlled eight channel surface stimulator.
Electrodes
were placed on each leg on m. quadriceps and m. glutaeus for hip and knee
extension and peroneal nerve to elicit flexion reflex. The fourth channel was
used corresponding to subjects individual needs.
Stimulation
patterns for stand up, walking and sit down could be easily set up and
optimised by adjusting up to 128 stimulation parameters in a task specific way.
Since
Kantrowitz demonstrated in 1963 paraplegic standing by quadriceps stimulation
several groups concentrate on restoration of lower limb function by means of
FES. Some researchers use implantable stimulators and electrodes or external
stimulators and percutaneous electrodes while others concentrate on surface
stimulation. Usage of implanted components is accompanied
by surgical procedures and a stay in the hospital. The advantage is that the
system is always ready for use whilst the non-invasive surface stimulation
needs every day some time for donning and doffing /1/,/2/.
Most groups that concentrate on surface stimulation
use four or six channel stimulators to achieve standing and stepping.
Our group developed an eight channel surface
stimulator that allows in conjunction with a PC-software optimisation of a
large amount of stimulation parameters to achieve smooth FES-gained movements.
The
introduced stimulation system consists of two modular designed four channel
stimulators, a microprocessor system that controls the stimulators and can
store different stimulation sequences for distinct tasks, a PC-interface that
allows communication with stimulators internal I2C bus via serial port and
finally a PC software for easy set-up of stimulation patterns (/3/,/4/).
The idea
behind this system is to set-up and test stimulation patterns with the help of
a PC-based user interface (UI) and then to download all parameters to the
microprocessor for patients personal use.
Stimulator:
During design
phase a strictly modular concept was kept (fig. 1). Each stimulation has its
own microcontroller (PIC 16C72, Microchip. Inc. Arizona, USA) that takes care
of the timing and data acquisition, has its own output stage and offers the
possibility to check electrode impedance and to measure certain parameters of
the stimulation evoked M-wave. A complete stimulation burst like shown in fig.
2 can be loaded via Inter-Integrated circuit (I2C) bus into the
microcontroller. Triggering the channel by either pressing a switch (crutch
mounted) or sending the start command via I2C bus launches the shown
stimulation burst. In addition various modes for continuous stimulation can be
programmed. A start of the continuous stimulation after the burst is important
for standing up. Continuous stimulation before and after the burst is necessary
to set up proper gait patterns and switching of the continuous stimulation
after the burst is used for sitting down.
Stimulation
parameters can be set independently for each of the three burst regions (fig.
2) as well as for the continuous mode. Biphasic constant voltage impulses with
a peak to peak voltage up to 160V and a pulse width up to 1ms+1ms (with minor
restrictions) can be delivered with a maximum frequency of 80Hz.

M-wave parameters and
electrode impedance are measured during the second part of the burst and are
available after the end of the third part.
Fig 1: Schematic
of one part of the eight channel stimulator Fig. 2: Three-parted
stimulation burst

Microprocessor
module
The used
microprocessor module is based on Motorola’s 68HC16 (Motorola Inc. Phoenix,
Arizona, USA). If the connection to the PC is terminated the microprocessor
overtakes the I2C bus master function and handles communication with all eight
stimulation modules. Stimulation patterns are stored in nonvolatile RAM and
transferred to the modules according to the chosen program like standing up,
walking and sitting down. A display informs about the current stimulation
situation while buttons and a panel coder allow change of stimulation amplitude
and several basic parameters. This module is also responsible for collection
and storing of M-wave parameters and checking of electrode impedance. All
stored data can be transferred via I2C bus to the PC.
PC-software
To create a
simple to handle user interface (UI) we decided to use a personal computer as
major input device because only a computer screen offers enough flexibility for
the UI design.
We chose
Inprise Delphi (Inprise Corp. Scotts Valley, CA, USA) as development
environment to create a 'Microsoft -Windows' compatible graphical user
interface (most PC-users are used to it). Fig. 3 shows a screen shot during
set-up of a walking pattern.
The next
important goal is the proper handling of the data. We decided to use a
standardised database for data management which causes a higher programming
effort (although Delphi supports very strongly database management) but on the
other hand data can be scanned, sorted, queried and analysed with any standard
database software. Expanding of the data structure can easily be done by adding
fields or tables to the initial database without effecting the previous
structure and consequently preserving file compatibility.
The
described system (fig. 4) was used by two male subjects. Both are level T6
paraplegics, use electrical stimulation for muscle training regular for four years
and are involved in a training’s program for FES walking with six channel
stimulators for 3 years.
Stimulation
was applied with standard skin surface electrodes. First we used three channels
each leg activating m. quadriceps, m. glutaeus maximus and withdrawal reflex.
Later the fourth channel was applied according to subjects individual needs.
Since this
was the first application of the described system on human subjects the FES
outcome was evaluated using patient’s experience with FES walking and by consulting
the medical staff.

Fig. 3: Screen
shot during set-up of walking pattern Fig.4: 8
channel stimulator (2 four channel stimulation modules, electrode leads,
battery and microprocessor housing, I2C interface

The first
sessions with the patients were used to improve the handling of the system.
After all hard- and software changes were made the system proofed to be
reliable and useful for paraplegic walking. In each of the following sessions
we tried to optimise the stimulation parameters for standing up, walker
supported walking and sitting down. For each task and channel 16 parameters can
be set. That means that 128 parameters can be trimmed to fulfil one of the
mentioned tasks. The UI proved to be very useful in this environment because
parameters could be optimised in an efficient way and very short time. The
first ideas during parameter optimisation were the usage of a higher
stimulation frequency during begin of the movement (1. phase), typically 50Hz
and then switching back to less fatiguing lower frequencies, typically 30Hz and
optimisation of the duration of 1. phase and 3. phase to achieve smooth
movements. Impulse duration was usually set to 0.5ms+0.5ms.
Patients
and medical staff agreed that smoother movements and less exhausting walking
could be achieved.
Patient 1
showed external hip rotation during stand up phase and also some hip abduction
during swing phase. To compensate for external hip rotation and abduction the
fourth channel was used to stimulate the adductors of the thigh. After a trial
phase and parameter optimisation standing up with nearly constant knee distance
and stepping with significant less external rotation of the hips could be achieved.
In patient
2 we tried to improve hip stability by using the fourth channel for stimulation
of the longissimus muscles. After a short training’s phase the patient showed a
more upright posture during standing and walking.
The new
eight channel surface stimulator allows optimisation of up to 128 parameters
for one stimulation sequence. A PC based software proved to be very useful for
visual control of the optimisation process. First tests showed that an
optimised parameter set helps to gain both smoother and better coordinated
movements. The availability of a fourth channel for each leg allows to overcome
patient specific weak points in movement. In the future standardised procedures
how to apply the additional channel in case of specific problems have to be
worked out as well as stimulation parameter optimisation procedures.
/1/ Baillière’s
Clinical Neurology - Int. practice and research: Neuroprostheses, Vol 4/ Num1
T. Baillière, G.S. Brindley, D.N. Rushton
Baillière Tindall, ISSN 0961-0521, 1995
/2/ FES for mobility: The lesson learned in
30 years
A. Kralj, T. Bajd, L. Vodovnik
„5th Vienna International Workshop on Functional Electrostimulation“,
Proceedings (ISBN 3-900928-03-7), 13-20, 1995
/3/ A modular PC-based system for easy setup of complex stimulation
patterns
M. Bijak, S. Sauermann, C. Schmutterer, H. Lanmueller,
E. Unger, W. Mayr
Proc. 2nd Ann. Conf. of the Int. FES Society and 5th
Conf. Neural Prostheses: Motor Systems 5 (ISBN 0-86491-173-4), 257-258, 1997
/4/ Modulares PC-unterstütztes
Stimulationssystem zur einfachen Erstellung komplexer Stimulationsparameter
M. Bijak, S. Sauermann , W. Mayr, H. Lanmüller, D. Rafolt, C. Schmutterer, E.
Unger
Biomedizinische Technik, Bd. 43, ErgBd 2, (ISSN 0939-4990), 123-124,1998
Supported
by the ministry of science and transport, the Austrian National Bank and the
Vienna ‘Bürgermeisterfond’.
Manfred
BIJAK, PH.D.
Department
of Biomedical Engineering and Physics
AKH 4/L
Waehringer
Guertel 18-20
1090
Vienna, Austria
Tel.:
+43-1-40400/1992 (1983)
Fax.:
+43-1-40400/3988
E-Mail:
m.bijak@bmtp.akh-wien.ac.at