DEVELOPMENT
OF PARAPLEGIC LEG POWERED CYCLING WITH THE LUMBO-SACRAL ANTERIOR ROOT
STIMULATOR IMPLANT.
T.A.Perkins*, N.deN.Donaldson*,
A.L.Dunkerley, N.A.C.Hatcher,
A.M.Tromans & D.E.Wood
*Implanted Devices Group,
UCL Medical Physics,
ABSTRACT
In
December 1994 a woman with a complete T9 spinal lesion had a 12-channel
stimulator implanted with electrodes on the anterior L2 to S2 spinal roots
bilaterally. Some restoration of cycling ability has already been demonstrated
in [1] and [2]. She achieved leg powered cycling for 1.2km at a time using a
recumbent tricycle in May 1998. The cycling system has since been developed
further. Initially, static measurements of pedal forces due to stimulating each
implant channel (and also some lumbar and sacral root combinations) were
plotted against cycle crank angle. From these tests, stimulation patterns were
derived to give the most positive static pedal forces for every crank angle.
These stimulation patterns were then used in dynamic cycling tests, the delay
between stimulation and muscle response being accounted for by advancing the
crank angle for selecting patterns by an angle equivalent to 0.25 seconds at
the measured speed. Co-ordinated pedalling was then possible at up to 86rpm. To
optimise control program further, we made dynamic pedal force measurements,
both on our paraplegic subject and an able-bodied cyclist for comparison. We
found that a reduction in simultaneous lumbar and sacral root stimulation
improved power output during rapid pedalling.
Key words: Cycling, paraplegia,
Functional Electrical Stimulation, spinal root stimulator implant.
INTRODUCTION
Numerous attempts to restore leg function
in paraplegia by Functional Electrical Stimulation (FES) have been made,
including [3] for standing/walking with peripheral nerve implants and [4] for
cycling with surface stimulation. The purpose of the British Medical Research
Council’s Lumbo-sacral Anterior Root Stimulator
Implant (LARSI) project, as outlined in [5] was to
evaluate root stimulation responses and determine their suitability primarily
for restoring standing in paraplegia. In the current project we seek to add the
function of cycling in a recumbent tricycle. We hope to obtain sufficient
utility for the patient to be encouraged to maintain significant daily leg muscle
training, especially as others have found health
benefits from sustained regular
METHODS
In December 1994, a complete T9 paraplegic
was implanted bilaterally with electrodes for stimulating the anterior spinal
roots from L2 to S2. The stimulation hardware needed was developed from that of [5]. Since the surgical wounds healed shortly
after the implantation, there has been no break in the skin: the implant is
controlled from outside via RF coupling.
When
cycling, the ankles are stabilized by Ankle Foot Orthoses
(AFOs) attached to the tricycle’s pedals. To choose
the correct patterns of stimulation, propulsive pedal force responses to
stimulation (ie at right angles to the crank) were
measured at 16 crank angles for 18 different patterns. For the stimulation
program, crank angle is measured by a shaft encoder and the relevant
stimulation pattern is then read from a look up table, in the stimulation
controller’s memory. The effect of the dynamic response of the muscles is
accounted for by noting the delay between stimulus at the spinal root and the
consequent peak muscle force generated. A crank angle phase advance in
stimulation, which is proportional to pedalling
speed, is obtained by angle used for the look up table being adjusted forward
by this delay.
Initially, power output was estimated from
the combined weight of rider and tricycle, rolling resistance, road gradient
and cycling speed. Dynamic measurements of crank angle and crank strain (via RF
telemetry from strain gauges in the cranks) give more direct power estimates.
Crank strains were also monitored with able-bodied cyclists for comparison.
Several
variants of the cycling program were tried, while monitoring power output. The
best of these programs was then chosen for use. To further assess the cycling
program, measurements of the net static pedal forces available from stimulation
were added to calculated forces due to the weight of the limb segments and the
AFO/pedal combination, to obtain a predicted plot of crank strain versus angle.
This was then compared with the actual crank strains obtained in cycling.
RESULTS
Our patient has had the tricycle home for
regular static leg powered cycling exercise (with the drive wheel supported on
a “resistance trainer” stand) since April 1998. In May1998, with the trainer
removed, she was able to cycle with the aid of her implant for over a kilometre
for the first time. She then cycled on gently undulating open road for
1200metres at a time with a mean power of 34Watts, corresponding to a speed of
12kph on level ground.
She has
subsequently reported a 3km. cycling ability. The patient maintains training at
about 30 minutes daily, alternating cycle training with that for standing.
Static
pedal forces in response to stimulation are shown in Figure 1. Having tried
muscle
response delay
allowances from 100-300mS, we found that the best power output was obtained
with a 0.25 second advance. At 60rpm this corresponds to adding 90 degrees of
phase-advance to the crank angle measurement. By removing some of the
co-activation in lumbar and sacral root stimulation that seems necessary to
obtain the best static pedal forces in Figure 1, we found that power output at
higher pedalling speeds was improved. Our patient could then pedal in a
co-ordinated manner at up to 86rpm. Figure 2 shows dynamic right crank strains
at increasing pedalling rates in response to stimulation. The spikes on the
response seen at higher angular speeds seem to be due to knee flexion reflexes,
which fortunately aid the cycling motion, but disturb the measured phase of
crank strain compared with the predicted response shown in Figure 3. Below
60rpm, however, Figure 4 shows that predicted and measured dynamic responses
correlate well in both phase and amplitude. We found that the crank strain data
from an able-bodied cyclist also closely matched that of our paraplegic subject
at these lower speeds, when cycling with the same speed and resistance.
Figure 2:
Position and crank strain versus time for LARSI cycling.
Figure 3:
Measured and predicted crank strain, LARSI cycling at greater than 60rpm.
Figure 4:
Measured and predicted crank strain LARSI cycling at less than 60rpm.
CONCLUSIONS
The stimulation program used for our
paraplegic subject produces cyclic changes in torque which match ‘normal’
cycling closely up to 60rpm. Considerable work remains on optimising the
cycling stimulation program further. The present
program nevertheless shows LARSI can already provide a paraplegic with a useful
and enjoyable leg powered cycling capability.
ACKNOWLEDGEMENTS
We
are grateful to the Wellcome Trust, Stanmore Royal
National Orthopaedic Hospital (RNOH) and INSPIRE for funding this work. We are
also grateful to the following contributors:- R. Fitzwater, RNOH, for assembly shaft encoder modules; From
Salisbury District Hospital:- E.Askew and S.Morant, for the AFOs; A.Lamb, for cycle
equipment design; I.Swain,
for planning and encouragement; S. Crook and P.Taylor,
for useful advice and J. Norton and P.Wright, for
static pedal force measurement. Most of all we would like to thank our
volunteer patient and our families for their tolerance of the long hours of
testing.
REFERENCES
1. Donaldson N, Rushton D & Tromans T (1997)
“Neuroprostheses for leg function after spinal-cord
injury” Lancet [letter], Sep 6, 350(9079),
p711.
2.
Perkins T.A., Donaldson N.deN.,
Harper V.J., Tromans A.T., Norton J.A., Rushton D.N. & Wood D.E.(1998) “Standing, stepping and
cycling for a T9 paraplegic with a Lumbo-sacral
Anterior Root Stimulator Implant (LARSI).”, Abstract Book of the INS 4th
International Congress / joint meeting with IFESS, Lucerne, Switzerland, Sept,
p. 229.
3. Brindley G.S., Polkey C.E. & Rushton D.N.
(1979) "Electrical splinting of
the knee in
paraplegia." Paraplegia 16, pp. 428-435.
4.
Petrofsky J.S., Heaton H.
3rd, Phillips C.A. (1983) “Outdoor bicycle for exercise in
paraplegics and quadriplegics.”J.Biomed.Eng. 5(4), pp. 292-6
5.
Donaldson N., Perkins T.A.
& Worley A.C.M. (1997) "Lumbar root stimulation for restoring leg
function. Methods: Stimulator and measurement of muscle actions."
Artificial Organs, 21(3), pp. 247-249.
6.
Mohr T., Podenphant J., Biering-Sorensen F., Galbo H., Thamsborg G. & Kjaer M. (1997) “Increased bone mineral density after
prolonged electrically induced cycle
training of paralyzed limbs in spinal cord injured man.” Calcified Tissue
International, 61(1), pp.22-25.
7.
Rushton
D.N., Donaldson N.deN., Barr F.M.D., Harper V.J.,
Perkins T.A., TaylorP.N. & Tromans
A.M. (1996) “Lumbar anterior root stimulator for lower limb control in
paraplegics.” in Neuroprosthetics: from basic research to clinical
application, M. Ferrarin, A. Pedotti,
J. Quintern & R. Riener
(eds), Springer-Verlag,
Munich, pp. 611-622.
8.
Perkins T.A., Donaldson N.deN., Worley A.C., Harper V.J., Wood D.E., & Rushton D.N. (1996) “Initial results with a lumbar/sacral
anterior root stimulator implant.” Ibid, pp. 623-634.
9.
Rushton D.N., Perkins T.A., Donaldson N.deN., Wood
D.E., Harper V.J., Tromans A.M., Barr F.M.D. &
Holder D.S. (1997) ‘LARSI: how to obtain favourable muscle contractions’, Proc
2nd Ann Conference of IFESS (IFESS’97) & 5th Trienn
Conference Neural Prostheses (NP’97) in Burnaby, British Columbia, Canada, Aug
1997, Popovic D (Ed), ISBN: 0-86491-173-4,
IFESS, La Canada Flintridge,
California, pp. 163-164.