2 Faculteit Bewegingswetenschappen, Vrije Universiteit Amsterdam, van de .... 5. 6. 7. IPI [ms]. Q uad ru plets nFT. Ip. P c). Fig.2: Mean nFTIpP± S.E.M (n=6) for ...
MUSCLE LENGTH DEPENDENCE OF OPTIMAL STIMULATION PATTERNS 1
P.Mela1, P.H.Veltink1, P.A.Huijing1,2, S.Salmons3 , J.C.Jarvis3.
Institute for Biomedical Technology (BMTI), Biomedical Signals and Systems, Department of Electrical Engineering, University of Twente P.O. Box 217, 7500 AE Enschede, The Netherlands. 2 Faculteit Bewegingswetenschappen, Vrije Universiteit Amsterdam, van de Boechorststr 9 1081 BT The Netherlands 3 Department of Human Anatomy and Cell Biology, University of Liverpool, New Medical School, Ashton Street, Liverpool L69 3GE, United Kingdom.
Abstract Stimulation patterns can be optimized by maximizing the force-time integral per stimulation pulse of the elicited muscle contraction. Such patterns, designed to provide the desired force output with the minimum number of pulses, may contribute to reduce muscle fatigue, which has been shown to correlate to the number of pulses delivered. Applications of electrical stimulation to use muscle as a controllable biological actuator may, therefore, be improved. Although muscle operates over a range of lengths, patterns have been determined only at optimum muscle length. In this study four-pulse patterns that produced the highest force-time integral were determined at different muscle lengths for rabbit tibialis anterior muscles isometrically stimulated. In agreement with previous studies, at high muscle length, the optimal stimulation pattern consisted of an initial short IPI (doublet) followed by longer IPI’s. However, at low length, the third pulse still produced more than linear summation (triplet). Furthermore, the relative enhancement of the force time integral per pulse was considerably larger at low length than at high length, suggesting that optimal patterns are length dependent. .
Introduction Skeletal muscle has attracted interest as a biological force generator in a number of clinical applications in which its activation is provided artificially by means of electrical stimulation. The main drawback of functional electrical stimulation (FES) of muscle is fatigue, operationally defined as loss of force in time, despite constancy of stimulation parameters. Since fatigue is correlated to the number of pulses delivered to the neuromuscular system [1], optimal patterns of stimulation have been investigated in a variety of applications with the aim of eliciting the desired force output with the minimum number of pulses [2-4]. Despite the fact that muscle works over a range of lengths, optimal stimulation patterns have been determined only at muscle optimum length (i.e. the length at which the muscle develops the highest isometric twitch force) [2-4].
The purpose of this study was to determine optimal four-pulse stimulation patterns at ten muscle lengths using an iterative technique: the interpulse interval (IPI) between two pulses producing the maximum force time integral per unit impulse (FTIpP) is selected and then further pulses are added successively following the same optimization criterion. An account of the results at two muscle lengths (low and high) is presented here.
Methods a) Set up The experiments were carried out on the tibialis anterior muscle of rabbits (n=6), stimulated by means of bipolar electrodes placed immediately beneath the common peroneal nerves of both legs. The nerves were divided between the electrode site and the spinal cord. The tibialis anterior tendon was transected in the foot and linked to the arm of a servo motor functioning as a force and length controller. A personal computer was used to provide synchronized control of the servo motor, muscle stimulation and data acquisition. The control code was written in Labview. b) Iterative method The muscle length range for the experiments extended from the lowest length at which a twitch force could be measured to the length at which passive force rose to approximately 4N. The optimum muscle length was then determined by delivering twitches at ten lengths evenly distributed within the length range. At the same series of lengths doublets with various IPIs (2, 4, 6, 8, 12, 24, 42 ms) were delivered, allowing one minute between contractions for recovery. The sequence was randomized for both muscle length and IPI in such a way as to cancel out any possible fatigue effect when averaging over all the tested muscles. For each doublet, isometric force was measured, the force time integral (FTI) calculated and plotted against the interval between the two pulses. For each muscle length the IPI which gave the highest force-time integral was selected and fixed. In a second randomised sequence an additional pulse was added to the train at IPIs of 2, 4, 8, 16, 24, 28, 32, 36, 42, 54 ms to find the optimal triplet pattern. The
same procedure was repeated for a fourth pulse (IPIs=8, 12, 24, 32, 42, 54 ms). c) Data processing If the muscle behaved as a linear system, the force time integral of its response would increase in direct proportion to the number of pulses applied and the ratio between the FTI of the response to a n-pulse train and n times the FTI of the response to a single pulse (normalized FTIpP: nFTIpP) would be unity. A nFTIpP greater than unity represents a more than linear summation and it is an index of the effectiveness of the stimulation train.
reaction: the response to such doublet was, therefore, equivalent to the contraction produced by a single twitch (hatched area). The IPI that produced the highest peak force and FTI was 4 ms at both lengths. However, the effect of the second pulse was greater at low length (four fold increase in peak force above single twitch) than at high length (2.5 fold increase). The pattern of summation was similar at the two muscle lengths, resulting in a close to exponential trend of mean nFTIpP plotted against IPI (fig. 2a).
Results Results are presented for two muscle lengths towards the extremes of the length range (low and high muscle length) Doublets Fig.1 illustrates the muscle response to two pulses delivered at different interpulse intervals at a low (a) and a high (b) muscle length.
Doublets nFTIpP
3.5 low length high length
3 2.5 2 1.5 1 0.5 2 4 6 8 12
50
100
150
Time [ms] a)
100
100
10
150
Time [ms] c)
100
14
50
100
0
150
100
250
a)
16 ms 28 ms 36 ms
5 4 3 2 1 0
24 8
16
24 28 32 36
IPI [ms]
50
100
150
Time [ms] d)
100 250
8 ms 24 ms 32 ms 54 ms
6 2
42
54
42
54
b)
2 ms
10
2 0
Time [ms] b)
100 250
54 ms
14
54 ms
6
2
150
18
32 ms
10
100
6
250
8 ms 24 ms
18
50
14
54 ms
2 2 ms 0 50
2 ms
4 ms
4 ms 16 ms 28 ms 36 ms
6
0
42
6
42 ms
4 2
250
Force [N]
10
42 ms
24
IPI [ms]
7
Triplets nFTIpP
24 ms
2 2 ms
8 6
Quadruplets nFTIpP
4
12
Force [N]
Force [N]
6
0
Force [N]
10
4 ms 8 ms
8
Force [N]
Force [N]
10
4 ms 8 ms 24 ms
7 6 5 4 3 2 1
0
50
100
150
100
250
Time [ms] Time [ms] e) f) Fig.1: Muscle force in response to two, three, four pulse train at different interpulse intervals delivered at low (a,c,e) and high (b,d,f) muscle length. The hatched areas represent the force-time integral of the doublets elicited with 2 ms IPI. . In both cases the 2 ms IPI fell within the refractory period of the muscle and did not elicit any muscle
0
8 12
24
32
IPI [ms]
c) Fig.2: Mean nFTIpP± S.E.M (n=6) for doublets a), triplets b), quadruplets c) at low (circle) and high (square) muscle length as a function of IPI. At low muscle length, on average a doublet with short IPI (4 ms) produced a contraction for which the forcetime integral was 3 times that produced by two separate
twitches at the same muscle length. At high muscle length the nFTIpP was lower, but still considerably higher than unity for short IPIs. Triplets At each length, a third pulse was added to the optimum doublet. Fig. 1 shows the experimental traces of the force elicited at c) low and d) high muscle length. The muscle response to the 2 ms IPI corresponded to the optimal doublet since no summation occurred. As for the doublets at low muscle length, a short interpulse interval produced a contraction with much higher peak and FTI, the effect being clearly dependent on the interval between the second and the third pulse. At high muscle length, the addition of a third pulse was not as beneficial (fig. 2b). The pattern of summation at low muscle length resembles that obtained for doublets, but it is different at high muscle length: the nFTIpP changed little with increasing IPIs and a maximum was achieved for a 42 ms IPI. Quadruplets Force traces elicited by four pulse trains based on the optimal triplets are shown in fig.1e and f. Different IPIs produced quite similar FTIs (fig.2c).
Discussion The purpose of this study was to determine the optimal four pulse stimulation sequence at different muscle lengths. The results show that at high muscle length the optimum stimulation pattern consists of an initial short IPI (doublet) followed by a longer IPI, in agreement with previous studies carried out at optimum muscle length [2-4]. However at low muscle length, it appears that a third stimulus after a second short IPI is also subject to more than linear summation, suggesting that the optimal pattern consists of a triplet of closely spaced pulses. There is not a single clear explanation for the existence of the ‘doublet effect’. Our findings are consistent with the idea of stiffness of the muscle-tendon complex being increased by the first pulse and therefore facilitating the transmission of the force produced by the second stimulus [4]. Although a single pulse could be enough to increase and maintain stiffness at high muscle length, two or more pulses may be required at low muscle length to take up the slack and stretch the series elastic components. An alternative mechanism for the length dependence of the ‘doublet effect’ is related to the length dependence of the efficacy of activation. It has been proposed [3] that the occupancy of Ca++-binding sites during a single twitch may be too short for all activated cross-bridges to undergo attachment and force generation, a second Ca++ transient would then allow a
longer time during which the myofilaments are exposed to high calcium concentrations. Force in response to a single twitch is produced more effectively at high muscle length. In accordance with our results the scope for force enhancement would then be more restricted than at low muscle lengths. To produce a sustained contraction sets of closely spaced doublets or triplets (n-lets) can be used [5]. Our study suggests that the ideal patterns may well be different at different muscle lengths. Optimization of the FTI of the elicited contraction is a suitable criterion to obtain the optimal stimulation pattern. However, Karu et al. [5] indicated that practical applications of sustained contractions may require a relative small force ripple and therefore repetition of the optimal sequence at an appropriate rate. Since three closely spaced pulses are required at low length, a relatively high repetition rate may be needed, particularly at the initiation of a contraction. It remains to be seen whether the optimal pattern for maintenance of force is similarly dependent on length.
References [1] Garland S.J., Garner S.H., McComas A.J.,: Relationship between numbers and frequencies of stimuli in human muscle fatigue, J Appl Physiol, 65(1) (1988), pp.89-93 [2] Zajac F.E., Young J.L., Properties of stimulus trains producing maximum tension-time area per pulse from single motor units in medial gastrocnemius muscle of the cat, Neurophysiol, 43(5) (1980), pp.1206-20 [3] Kwende M.M., Jarvis J.C., Salmons S., The inputoutput relations of skeletal muscle, Proc R Soc Lond B 261(1361 (1995), pp.193-201 [4] Stein R.B., Parmiggiani F., The stiffness of slowtwitch muscle lags behind twitch tension: implications for contractile mechanisms and behavior, J Physiol Pharmacol, 57(10) (1979), pp1189-92 [5] Karu Z.Z., Durfee W.K., Barzilai A.M., Reducing muscle fatigue in FES applications by stimulating with N-let pulse trains, IEEE Trans Biomed Eng., 42(8) (1995), pp 809-17.
Acknowledgments This study was supported by the European Union TMR project Neuros.
