OPTIMIZATION OF A FES CYCLING NEUROPROSTHESIS ON

STROKE PATIENTS BY MEANS OF THE LEFT AND RIGHT

CRANK MEASUREMENTS

Ferrante Simona

, Comolli Lorenzo

, Pedrocchi Alessandra

, Bocciolone Marco

, Ferrigno Giancarlo

and Molteni Franco

Bioengineering Department, NITLab - TBMLab, Politecnico di Milano, Via Garofalo 39, Milano, Italy

Department of Mechanical Engineering, Politecnico di Milano, Via La Masa 34, Milano, Italy

Valduce Hospital, Villa Beretta Rehabilitation Center, Via N.Sauro 17, Costamasnaga (LC), Italy

Keywords: FES cycling, rehabilitation, hemiplegia.

Abstract: The use of functional electrical stimulation (FES) is a well established method in the rehabilitation of stroke

patients. In particular, a bilateral movement such as cycling induced by FES would be crucial for these

patients who had an unilateral motor impairment and had to recover an equivalent use of the limbs. To

improve the rehabilitative effects of the FES cycling, a metrologically qualified cycle-ergometer was used,

so that the left and right crank torque values are measured in real-time. Three protocols were evaluated.

First, healthy subjects performed voluntary pedaling. Second, healthy subjects were stimulated one muscle

individually to study the contribution of each single muscle to the cycling. Third, stroke patients executed a

complete FES cycling trials. Results demonstrated that the proposed sensors could be successfully used to

monitor online the unbalance of the cycling. Single muscle tests showed that only the quadriceps and the

hamstrings provided a significant contribution to the crank torque. Patient trials confirmed the difficulty for

stroke subjects to carry out symmetrical cycling. The use of the proposed sensors, hence, could offer a good

signal for biofeedback neuroprostheses and for closed loop controllers.

1 INTRODUCTION

A neuroprosthesis is defined as a functional

electrical stimulation (FES) device used as a

substitute for lost neurological function. The use of a

neuroprosthesis is a well established method in the

rehabilitation of individuals with spinal cord injuries

and stroke (Kralj, Acimović & Stanic, 1993,

Petrofsky, 2004). Among all the existent

neuroprostheses, the cycling movement induced by

the stimulation of the major muscles of the lower

limbs is now becoming a spread application (Hunt et

al., 2004, Trumbower, & Faghri 2005, Szecsi et al.,

2007). In the following this artificial movement will

be named FES cycling.

One of the most difficult and appealing

improvements in FES research is the integration of

hardware and software and the development of

control systems for the neuroprosthesis in order to

enhance the clinical benefits for the patients. Some

of the rehabilitation objectives for spinal cord

injured patients are an increase in muscular tone, an

improvement in the peripheral and cardiac

circulation avoiding the occurrence of decubitus

ulcera and the prevention of joint rigidity (Hunt et

al. 2004). In stroke patients, the neuroprosthesis also

becomes therapeutic and in fact, it permits one to

learn new motor strategies for Central Nervous

System (CNS), exploiting residual capacity (Lee &

van Donkelaar, 1995, Sheffler & Chae, 2007).

Indeed, FES cycling providing the complete

afference of the task to the stroke patients could re-

educate the synaptic controls needed to produce a

well organized movement. Because of the laterality

of the pathology, one of the rehabilitation aims is the

recovery of walking and particularly the recovery of

the symmetry of the movement. Therefore, the use

of FES cycling was investigated as a method to

re-learn the bilateral and symmetrical use of the legs.

206

Simona F., Lorenzo C., Alessandra P., Marco B., Giancarlo F. and Franco M. (2008).

OPTIMIZATION OF A FES CYCLING NEUROPROSTHESIS ON STROKE PATIENTS BY MEANS OF THE LEFT AND RIGHT CRANK MEASURE-

MENTS.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 206-211

DOI: 10.5220/0001054202060211

 SciTePress

The feasibility of applying FES cycling to stroke

patients was analyzed in a previous study and

encouraging results were reached (Ferrante et al.,

2006). The stimulation strategy delivered to the

quadriceps, hamstrings, gluteus maximus and tibialis

anterior of the two legs was exactly the same for

both the legs, but shifted of 180° in respect to the

crank angle. The patient was asked not to voluntary

contribute to the motor task. In this study the total

motor torque produced at the crank was measured.

Therefore, it was possible to quantify the

performance obtained in terms of power output but

not in terms of unbalance at the pedals which is a

crucial aspect in the rehabilitation of individuals

with stroke.

The aim of this study is to develop and test

sensors able to measure the torque at the right and

left crank in order to provide a step forward in the

application of FES cycling on individuals with

stroke. In fact, monitoring these signals during the

movement would be a crucial starting point in the

design of an automatic time variant controller for

symmetry.

2 METHODS

2.1 Experimental Setup

A current–controlled 8-channel stimulator and a

motorized cycle-ergometer were used for the

experiments. It was possible to control the

cycle-ergometer by changing the resistant torque and

the angular velocity or by directly setting the motor

voltage.

As shown in Figure 1, the cycle-ergometer was

equipped with instruments able to measure the

bending moments and the radial forces at the right

and left cranks, so that the torque can be computed

(Gföhler et al., 2001). Four Wheatstone full bridge

made up of electrical resistance strain gauges were

used to measure the strain on the right and left crank

during the cycling task. The strain gauge bridges are

conditioned through a four-channel wireless device,

which allows to transmit the signal from the rotating

shaft to the acquisition system. The angular crank

position was measured through optical encoders

mounted on the main wheel of the ergometer. More

details on the instrumentation, the acquisition chain

and the metrological characterization of the

cycle-ergometer can be found in Comolli et al., 2005

and Bocciolone, Comolli and Molteni, 2008. The

cycle-ergometer was equipped with two ankle foot

orthoses fixed to the pedals and used to stabilize the

legs and to constrain the movement to the sagittal

plane.

Figure 1: The instrumented cycle-ergometer, the wireless

device is in the box on the crank.

All the transducers were connected to a PC for data

acquisition and stimulation device control. The

real-time acquisition and control system is shown in

Figure 2. The stimulation controller set the

stimulator parameters according to the crank angle

measured in real time. Therefore, each muscle used

is activated in a particular angular phase of the

cycling movement as explained in the following.

Figure 2: The experimental setup.

2.2 Experimental Protocol

Written informed consent was obtained from all the

subjects who took part in the experiments. Three

different experimental sessions were performed and

are explained in the following paragraphs.

PC and

stimulation

controlle

8-channel

stimulator

Acquisition board

Subject &

ergometer

Wireless device

Angular

sensors

Force

sensors

OPTIMIZATION OF A FES CYCLING NEUROPROSTHESIS ON STROKE PATIENTS BY MEANS OF THE LEFT

AND RIGHT CRANK MEASUREMENTS

207

2.2.1 Tests During Voluntary Pedaling

First, the developed sensors were tested during

voluntary pedaling of healthy subjects. The protocol

required the subject to perform one trial lasting

8 min. The first 2 min the motor of the

cycle-ergometer was used and the subject was

cycling only passively at 30 rpm. Then, the subject

carried out a normal cycling for 2 min, trying to

keep the former velocity. After that, 2 min in which

the subject was asked to pedal using only the right

leg and 2 min using only the left one were

performed.

2.2.2 Stimulation of the Muscles

Individually

Once tested the sensors, a specific FES cycling

experimental protocol was carried out on a healthy

subject. Once the subject was sat on the ergometer

and the electrodes were placed, an initial trial to

choose the stimulation current of each muscle was

carried out. The chosen stimulation currents were set

at a value that produced a tetanic contraction using a

pulse width of 400 µs. The crank angle was set at 0°

in correspondence to the point of maximum flexion

of the left hip. During the trials 8 muscles, 4 per

each lower limbs (quadriceps, hamstrings, gluteus

maximus and tibialis anterior) were stimulated. The

stimulation strategy, i.e. the choice of the cycling

phases in which each muscle was stimulated in

respect to the crank angle, were selected following

Ferrante et al., 2005. An ON-OFF pulse width

profile was used: for all the muscles the pulse width

value was fixed at 400 µs during the ON phase and

at 0 µs during the OFF phase. The stimulation

frequency was set at 20 Hz and all the torque signals

were gathered at 500 Hz.

The subject performed a trial with a resistant

torque value of 3 Nm. The angular velocity was

maintained by the motor at a minimum value of

30 rpm during all the trials.

The trial lasted 10 min: during the initial and

final 60 s, the subject was pedaling passively. In the

intermediate part of the test, each muscle group was

individually stimulated for 60 s within its

stimulation range.

2.2.3 First Tests on Stroke Patients

Finally, a feasibility test was carried out on 2

post-acute stroke patients at the beginning of their

rehabilitation. Both the patients performed a

complete FES cycling trial according to Ferrante et

al., 2006.

2.3 Data Analysis

The measured bending moments and radial forces

were used to compute the torque at the right and left

crank. The torque was expressed as a function of the

crank angle. The median and the 5

-95

percentiles

of the right and left active torque were computed in

each test condition of the trials. Then the left and

right active torque at the crank were computed as the

difference between the torque produced during the

active phases and the one generated during passive

pedaling. The total torque was obtained adding

together the right and left torque.

3 RESULTS

3.1 Tests During Voluntary Pedaling

Figure 3 shows the results obtained by an healthy

subject during the voluntary pedaling protocol in

terms of the right and left active torque measured at

the crank.

Figure 3: Measurement of the right (solid line) and left

(dashed line) passive (a) and active (b,c,d) torque; the

median on 15 revolutions are shown. In dotted lines the 5

and 95

percentiles are reported.

(d) only left leg

(b) voluntary cycling

(a) Passive cycling

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

208

In panel (a) the repeatability of the signals during

passive cycling is noticeable. In panel (b), the

subject was performing a normal pedaling. The

effect of the two legs on the right and left active

torque was quite similar in amplitude but it showed a

shift of 180° in phase. The asymmetrical cycling

presented in panels (c) and (d) highlighted a dual

behaviour. When only the left leg was cycling

actively (c), only the left active torque had a positive

peak and the right torque was quite zero during the

whole revolution. The peaks of the active torque in

panels (b), (c) and (d) were produced when the right

quadriceps was pushing. This behaviour agrees with

the fact that during a voluntary pedaling the

quadriceps muscle gives the greatest contribution to

the movement.

The results obtained in the same trials in terms of

total torque are reported in Figure 4. The panels are

referred to the same working conditions of Figure 3.

The total torque seems to be a good signal to

distinguish an asymmetry in the cycling movement.

Figure 4: Measurement of the total passive (a) and active

(b,c,d) torque; the median on 15 revolutions are shown in

solid line. The 5

and 95

percentiles are reported in

dotted lines.

3.2 Stimulation of the Muscles

Individually

Figure 5 shows the results obtained by the healthy

subjects during the FES protocol when only the right

quadriceps (a), hamstrings (b), gluteus maximus (c)

and tibialis anterior (d) were selectively stimulated.

In all the panels of Figure 5 the grey area represent

the angular range in which the muscle was

stimulated.

Figure 5: Measurement of the right (solid line) and left

(dashed line) active torque; the median on 15 revolutions

are shown. The 5

and 95

percentiles are reported in

dotted lines.

In panels (a) and (b), a positive peak of the right

active torque can be noticed exactly in

correspondence to the muscular stimulation range.

This confirms that the muscles were stimulated

in their functional range, i.e., when they were

assisting the cycling motion.

The active left torque was nearly zero during the

whole revolution because the left leg was pedaling

passively. The gluteus and the tibialis anterior

instead did not produce an effective right active

torque at the crank, thus their contributions to the

drive torque were negligible. Such a result was

(d) only left leg

(b) voluntary cycling

(a) Passive cycling

(b) Hamstrings

(a) Quadriceps

(d) Tibialis anterior

OPTIMIZATION OF A FES CYCLING NEUROPROSTHESIS ON STROKE PATIENTS BY MEANS OF THE LEFT

AND RIGHT CRANK MEASUREMENTS

209

expected because the gluteus, as a proximal muscle

acting only on the hip, can transfer a low

contribution to the crank, while the effect of the

tibialis anterior is limited by the use of the ankle foot

orthosis.

These results are confirmed by the total torque

shown in Figure 6.

Figure 6: Measurement of the total torque; the median on

15 revolutions is shown in solid line. The 5

and 95

percentiles are reported in dotted lines.

3.3 First Tests on Stroke Patients

Figure 7 reports the results obtained by a post-acute

stroke patient during a FES cycling trial performed,

stimulating 8 muscles following the stimulation

strategy described in Ferrante et al., 2006. As it is

shown in panel (a) the positive peak of the left

torque is the 40% greater then the one of the right

torque, which was correspondent to the pushing of

the impaired leg.

Figure 7: The right (solid line) and left torque (dashed

line) are reported. Results are shown in terms of the

median torque of 50 revolutions obtained by a stroke

patient during an FES cycling trial. This result suggests

the idea to use these signals as control signals for a

real-time automatic controller to minimize the asymmetry

of the movement in stroke patients.

4 CONCLUSIONS

The possibility of using a rhythmic and bilateral

movement such as cycling induced by a stimulation

pattern delivered in both legs seems to be a good

rehabilitative method for post acute stroke patients.

These patients, having an unilateral motor

impairment, need to be re-educated to the correct use

of both the lower limbs together in order to recover

the motor control symmetry in more complex and

demanding task, such as walking.

The aim of this study was to develop and test

sensors designed for a FES cycling neuroprosthesis

specific for stroke patients. Custom designed sensors

able to measure independently the torque produced

by the right and the left leg during cycling were

tested both on healthy subjects and stroke patients.

To understand the effect of the stimulation of

each muscle on the signals, a proper protocol was

defined. Each muscle was stimulated individually in

the phase of the revolution in which his action was

functional to the movement (Ferrante et al., 2005,

Ferrante et al., 2006). The results obtained suggested

that the developed sensors allowed to distinguish the

effect of the stimulation of the quadriceps and

hamstrings during the movement. Instead, the active

torque produced by the gluteus maximus and by the

tibialis anterior were quite negligible. Other

experimental trials using this second protocol

presented are now ongoing in our laboratory on

more subjects.

This study was crucial to verify that the signal to

noise ratio of the torque measurement is sufficient to

discern the torque induced by the stimulation of the

single muscle, which are an order of magnitude

lower than the ones induced by the voluntary active

pedaling.

(d) Tibialis anterior

(b) Hamstrings

(

)

Quadrice

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

210

The possibility to monitor the torque at the right and

left crank represents an important step forward in the

application of FES cycling on stroke patients. It

would be possible to use the acquired torques to

monitor the unbalance of the movement on line. The

unbalance could be used as a biofeedback signal to

the patient. Thus, looking at a display where the

asymmetry is shown, he could correct the exercise

on line giving a voluntary contribution.

Alternatively, starting from this unbalance signal it

would be possible to develop a closed loop

controller aiming at the maximization of the

symmetry during the movement. The achieved

results also suggest that a simplified first version of

the controller could act only on the stimulation of

the hamstrings and quadriceps and not on the gluteus

maximus and tibialis anterior, which shown a quite

negligible effect on the right and left torque.

ACKNOWLEDGEMENTS

This work was supported by the Italian Institute of

Technology (IIT), and the Fondazione Cariplo in the

framework of the research program HINT@Lecco.

This project was realized thanks to the grant

INGENIO funded by the European Social Found,

the Welfare Ministry and by the Lombardia region.

Authors would like to acknowledge Mauro

Rossini (Villa Beretta Rehabilitation Center) for his

helpful discussion.

REFERENCES

Kralj, A., Acimović, R., Stanic, U., 1993. Enhancement of

hemiplegic patient rehabilitation by means of

functional electrical stimulation. Prosthetics and

Orthotics International, vol. 17, no. 2, pp. 107-114.

Petrofsky, J.S., 2004. Electrical stimulation:

neurophysiological basis and application. Basic and

Applied Myology, vol. 14, no. 4, pp. 205-213

Hunt, K., Stone, B., Negard, N., et al., 2004. “Control

strategies for integration of electric motor assist and

functional electrical stimulation in paraplegic cycling:

utility for exercise testing and mobile cycling. IEEE

Transactions on Neural Systems and Rehabilitation

Engineering., vol. 12, no. 1, pp. 88–101.

Lee, R.G., van Donkelaar, P., 1995. Mechanisms

underlying functional recovery following stroke. The

Canadian Journal of Neurological Sciences, vol. 22,

pp. 257-263.

Sheffler, L.R., Chae J., 2007. Neuromuscular electrical

stimulation in neurorehabilitation. Muscle Nerve, vol.

35, pp. 562-590.

Ferrante, S., Pedrocchi, A., Ferrigno, G., et al., 2006. FES

cycling treatment on hemiplegic patients: preliminary

results. In IFESS 06, 11

Annual Conference of the

FES Society, pp. 77-79.

Ferrante, S., Saunders, B., Duffell, L., et al., 2005.

Quantitative evaluation of stimulation strategies for

FES cycling. In IFESS 05, 10

International FES

Society Conference, pp. 94-96.

Gföhler, M., Angeli, T., Eberharter, T., et al., 2001. Test

bed with force-measuring crank for static and dynamic

investigations on cycling by means of functional

electrical stimulation. IEEE Transactions on Neural

Systems and Rehabilitation Engineering, vol. 9, no. 2,

pp. 169-180.

Comolli, L., Cantatore, A., Zappa, E., et al., 2005.

HINT@LECCO project: metrological characterization

of a cycle-ergometer. In SIAMOC 05, 6

Congress of

the Italian Society of Movement Analysis in Clinic.

Bocciolone M., Comolli L., Molteni F., 2008.

Metrological characterization of a cycle-ergometer. In

Biodevices 08, International Conference on

Biomedical Electronics and Devices.

Trumbower, R.D., Faghri, P.D., 2005. Kinematic analyses

of semireclined leg cycling in able-bodied and spinal

cord injured individuals. Spinal Cord, vol 43, no 9, pp.

543-549.

Szecsi, J., Krause, P., Krafczyk, S., Brandt, T., Straube,

A., 2007. Functional output improvement in FES

cycling by means of forced smooth pedaling. Medicine

and science in sports and exercise, vol 39, no. 5, pp.

764-80.

OPTIMIZATION OF A FES CYCLING NEUROPROSTHESIS ON STROKE PATIENTS BY MEANS OF THE LEFT

AND RIGHT CRANK MEASUREMENTS

211