Specific modulation of motor unit discharge for a similar change in fascicle length during shortening and lengthening contractions in humans
- 1Laboratory of Applied Biology, Université Libre de Bruxelles, 28 avenue P. Héger, CP 168, 1000 Brussels, Belgium
- Corresponding author J. Duchateau: Laboratory of Applied Biology, Université Libre de Bruxelles, 28 avenue P. Héger, CP 168, 1000 Brussels, Belgium. Email: jduchat{at}ulb.ac.be
Abstract
This study examines the effect of a change in fascicle length on motor unit recruitment and discharge rate in the human tibialis anterior during shortening and lengthening contractions that involved a similar change in torque. The dorsiflexor torque and the surface and intramuscular electromyograms (EMGs) from the tibialis anterior were recorded in eight subjects. The behaviour of the same motor unit (n = 63) was compared during submaximal shortening and lengthening contractions performed at a constant velocity (10 deg s−1) with the dorsiflexor muscles over a 20 deg range of motion around the ankle neutral position. Muscle fascicle length was measured non-invasively using ultrasonography. Motor units that were active during a shortening contraction were always active during the subsequent lengthening contraction. Furthermore, additional motor units (n = 18) of higher force threshold that were recruited during the shortening contraction to maintain the required torque were derecruited first during the following lengthening contraction. Although the change in fascicle length was linear (r2 > 0.99), and similar for both shortening and lengthening contractions, modulation of discharge rate differed during the two contractions. Compared with an initial isometric contraction at short (11.9 ± 2.4 Hz) or long (11.7 ± 2.2 Hz) muscle length, discharge rate increased only slightly and stayed nearly constant throughout the lengthening contraction (12.6 ± 2.0 Hz; P < 0.05) whereas it augmented progressively and more substantially during the shortening contraction, reaching 14.5 ± 2.5 Hz (P < 0.001) at the end of the movement. In conclusion, these observations indicate a clear difference in motor unit discharge rate modulation with no change in their recruitment order between shortening and lengthening contractions when performed with a similar change in muscle fascicle length and torque.
The force produced by a muscle is influenced by its length (Gordon et al. 1966), and the modality and velocity of the contraction (Katz, 1939; Edman et al. 1978). A common observation in whole-muscle or single-fibre studies in experimental animals is that the force achieved during a maximal contraction is greater in lengthening (eccentric) than in isometric and shortening (concentric) conditions when measured on the plateau and on the descending limb of the length–tension curve (Katz, 1939; Edman et al. 1978; Morgan et al. 2000). In the performance of voluntary actions, the contraction-type difference in the force capacity of muscle may be related to the control strategy used by the central nervous system (CNS) to activate the motor unit pool of the muscle (Westing et al. 1991; Pinniger et al. 2000). This hypothesis is supported by the observation that EMG activity recorded at the same movement velocity is usually lower during maximal voluntary lengthening compared with shortening contractions (Komi & Burskirk, 1972; Westing et al. 1991; Aagaard et al. 2000). Furthermore, the maximal torque that can be achieved during a lengthening contraction is increased by the addition of electrical stimulation superimposed over the voluntary effort (Westing et al. 1991; Pinniger et al. 2000). The incomplete activation during lengthening contractions is accompanied by lower excitability of the corticospinal tract to transcranial magnetic or electrical stimulation (Abbruzzese et al. 1994; Sekiguchi et al. 2003) and depressed monosynaptic (Romano & Schieppati, 1987; Abbruzzese et al. 1994; Nordlund et al. 2002) and polysynaptic reflex excitability (Nakazawa et al. 1997). Furthermore, EEG recordings indicate greater and earlier cortical activity during submaximal and maximal lengthening elbow flexor actions (Fang et al. 2001, 2004), suggesting that the CNS may plan and program lengthening movements differently from shortening contractions.
The mechanisms that give rise to specific muscle activation during shortening and lengthening contractions (see Enoka, 1996) involve modulation in motor unit recruitment and rate coding. While some studies reported that lengthening contractions are associated with a selective activation of high-threshold fast-twitch motor units and a derecruitment of low-threshold slow-twitch units (Nardone et al. 1989; Howell et al. 1995; Linnamo et al. 2003), others reported a recruitment order that is consistent with the size principle (Henneman, 1957) for both shortening and lengthening contractions (Garland et al. 1994; Søgaard et al. 1996; Bawa & Jones, 1999; Stotz & Bawa, 2001). Although, Stotz & Bawa (2001) reported the recruitment of additional higher threshold units during some lengthening contractions, this occurred only when the force or movement profile was erratic. In contrast to the similarity in recruitment order, motor unit discharge rate does vary with contraction type. Average discharge rate is usually lower during submaximal lengthening contractions compared with shortening contractions (Tax et al. 1989; Howell et al. 1995; Søgaard et al. 1996; Kossev & Christova, 1998; Semmler et al. 2002; Del Valle & Thomas, 2005), even when the number and properties of identified active motor units were similar (Søgaard et al. 1996).
To obtain a more complete understanding of the functional organization of the motor unit pool during shortening and lengthening contractions, it would be instructive to determine if a similar change in torque is reached mainly by selective recruitment of high-threshold motor units, the modulation of discharge rate, or by both mechanisms. Some of the discrepancy in the existing literature could be explained because populations of units were often compared, instead of analysing the behaviour of the same unit during both contraction types. Furthermore, movement velocity was not always carefully controlled, and the change in muscle length during movement was only estimated from the recording of joint position and not from direct measurement of fascicle length. This is a critical issue because it has been shown that fascicle length during maximal shortening and lengthening contractions is not linearly related with joint angle in the tibialis anterior (Reeves & Narici, 2003). Another major advantage of measuring fascicle length during movement is that it can provide length information from the portion of muscle where the motor units are recorded, in contrast with the estimate of the whole muscle length from changes in joint angle. Nonetheless, the association between changes in fascicle length with joint angle may vary across muscles, which could explain the divergent results on motor unit recruitment and discharge rate in shortening and lengthening contractions.
Therefore, the purpose of this work was to examine the effect of a change in fascicle length of the tibialis anterior muscle on the recruitment and discharge rate of the same motor unit during submaximal shortening and lengthening contractions for a similar change in torque. Angular velocity about the ankle joint was constant during both types of contractions. It was hypothesized that differences in the rate of change in fascicle length could explain some of the previously reported differences in motor unit discharge rate observed during shortening and lengthening contractions.
Methods
Subjects
Eight subjects (6 men and 2 women) age 22–48 years, participated in this investigation and were tested on several occasions for a total of 24 experimental sessions. Two successive sessions were separated by at least 1 week. Prior to the experimental sessions, all subjects were familiarized with the procedure and contraction modalities during one or two sessions. None of the subjects had any known neurological or motor disorder prior to testing. They were all volunteers and gave their informed consent before participating in the study. This investigation was approved by the University Ethics Committee and all the experimental procedures were performed in accordance with the Declaration of Helsinki.
Ergometric device
A motor-driven computer-controlled ergometer (Type HDX 115C6; Hauser Compax 0260M-E2; Offenburg, Germany) was used (Pasquet et al. 2000). This device, equipped with a footplate that was fixed to the rotational axis of the motor, recorded the torque generated by the dorsiflexor muscles under static and dynamic (isokinetic) conditions. The subject was secured on an adjustable chair in a slightly reclined position. The right foot was strapped to the plate so that the axis of rotation of the ankle joint was aligned with the shaft of the torque motor. The plate was inclined at an angle of 45 deg relative to the floor and the position of the subject was adjusted to obtain a 90 deg angle for the ankle (neutral position or 0 deg) and a 120–130 deg knee angle. This position was duplicated across sessions. The foot was held in place by a heel block and was tightly attached to the plate by means of two straps. One strap was placed around the foot, 1–2 cm proximal to the metatarsophalangeal joint of the toe, and the second strap was placed around the foot, just below the ankle joint.
Mechanical and EMG recordings
The torque produced by the dorsiflexor muscles during contractions was measured by a strain-gauge transducer (sensitivity, 0.018 V (N m)−1; linear range, 0–200 N m) mounted on the rotational axis of the motor. The force signal was amplified and filtered (AM 502; Tektronix, Beaverton, OR, USA; bandwidth DC–300 Hz).
Motor unit potentials were recorded by a selective electrode that comprised two 50 μm diamel-coated nichrome wires glued into the lumen of a 30 gauge hypodermic needle (Duchateau & Hainaut, 1990). The electrode was inserted in the middle part of the tibialis anterior muscle and during each experimental session the needle was inserted at different locations. At each location, the needle was manipulated to various depths and angles to obtain a recording site from which the same motor unit was monitored at the two ankle joint positions (10 deg plantarflexion and 10 deg dorsiflexion) and during the shortening and lengthening phases of the contraction. The EMG signal was amplified by a custom-made differential amplifier (×2000) and filtered (100 Hz to 10 kHz) before being displayed on a Tektronix TAS 455 oscilloscope. The surface EMG of the tibialis anterior, soleus and lateral gastrocnemius were recorded by means of two silver disk electrodes (8 mm diameter) placed 2–3 cm apart on the belly of the muscle. The electrodes over tibialis anterior were located on either side of the needle electrode. The ground electrodes (silver plates of 2 cm × 3 cm) were placed over the tibia. The EMG signals were amplified (×1000) and filtered between (10 Hz and 1 kHz) by a custom-made differential amplifier.
Experimental procedure
Prior to the recording of single motor units, the isometric torque exerted by the dorsiflexor muscles during a maximal voluntary contraction (MVC) was determined. First, the subject performed three MVCs of 4–5 s duration, separated by 2–3 min rest in a random order at ankle angles of 10 deg plantarflexion (long) and 10 deg dorsiflexion (short). This was followed by the recording of the dorsiflexion torque recorded during maximal shortening and lengthening contractions (2–3 MVCs in each condition) at a constant angular velocity of 10 deg s−1. Once a motor unit action potential was clearly identified at each recording site, subjects were asked to produce a ramp contraction at the two ankle positions at a rate of ∼5% MVC s−1 up to the recruitment of the selected unit and then to hold the torque constant to sustain a minimal, repetitive discharge of the unit for at least 5 s. Target torques were thus determined at short and long muscle lengths (Pasquet et al. 2005). Thereafter, subjects were asked to perform the following task: (1) sustain an isometric dorsiflexion torque at the target torque for the long muscle length for 5 s; (2) as the torque motor dorsiflexed the foot about the ankle, to assist the motion with the dorsiflexors and to reach smoothly the target torque for the short muscle length at the end of the movement; (3) to sustain an isometric dorsiflexion torque at the target torque for the short muscle lengths during 5 s; (4) as the torque motor plantarflexed the foot about the ankle, to resist the motion with the dorsiflexors and to reach smoothly the target torque for the long muscle length at the end of the movement (Fig. 1). This cycle was repeated at least 10 times depending on the ability of the subject to perform the task accurately (see below for the criteria used). To accomplish the task, subjects were provided with visual feedback on a digital oscilloscope of the actual torque and the torque targets for the two muscle lengths (Model 120; Nicolet, Madison, WI, USA). The subjects also received audio feedback of motor unit discharge rate to help them to recognize the selected unit. The shortening and lengthening contractions lasted 2 s at a constant angular velocity (10 deg s−1) over a 20 deg range of motion (from 10 deg plantarflexion to 10 deg dorsiflexion, and from 10 deg dorsiflexion to 10 deg plantarflexion around the ankle neutral position for shortening and lengthening contractions, respectively). The contractions were performed at a relatively slow velocity to diminish the interference of the unloading (shortening contraction) or stretch (lengthening contraction) reflexes with the central command to the muscle and to minimize torque fluctuations at the onset of movement. Two successive trials from different electrode locations were separated by at least 5–10 min of rest.
The protocol was intended to compare the behaviour of the same unit during the isometric and dynamic (shortening and lengthening) contractions when the muscle produced a similar change in torque. The strategy to control the change in muscle torque was preferred to using the EMG as an index of contraction intensity. Indeed, EMG activity is usually less during lengthening contractions (Westing et al. 1991; Aagaard et al. 2000; Klass et al. 2005) and differences in the level of motor unit synchronization during shortening and lengthening contractions (Semmler et al. 2002) would probably contribute to differences in the amount of EMG cancellation in the two conditions (Keenan et al. 2005) that would confound the comparison of the EMGs.
Data analysis
Data processing was performed off-line from taped records (Sony PCM-DAT, DTR 8000; Biologic, Claix, France). All signals were acquired on a personal computer at a sampling rate of 3 kHz (force), 6 kHz (surface EMG) or 12 kHz (intramuscular EMG) by a MP150 data acquisition system (Biopac Systems, Santa Barbara, CA, USA).
MVC torque (isometric or dynamic contractions) was determined from the trial that yielded the largest value. The MVC torque and associated average full-wave rectified EMG amplitude (aEMG) were measured during a 2 s epoch during the MVC plateau (isometric contractions) or throughout the entire range of motion (shortening and lengthening contractions). Motor unit discrimination was accomplished either with a window discriminator (Duchateau & Hainaut, 1990) or, when necessary, by a computer-based, template-matching algorithm (Signal Processing Systems, SPS 8701, Malvern Victoria, Australia). Trials that contained abnormally short and long interspike intervals due to discrimination error were re-analysed on a spike-by-spike basis. Single motor unit action potentials were identified on the basis of amplitude, duration, and waveform shape. Only the motor units that: (1) were clearly identified, (2) showed waveforms and amplitudes that changed gradually and systematically over time throughout the different parts of the task, and (3) that differed by less than 20% in amplitude at the two ankle angles and movement modalities were included in the analysis. Furthermore, only trials during which torque increased linearly within a 90% confidence interval and did not deviate for more than 5% MVC at the two target levels were included in the analysis. These criteria and the technical difficulty of recording the same motor unit during movements explain the relative low yield in each session (∼3). Motor unit recruitment threshold, defined as the torque at which the motor unit began to discharge, was determined during the isometric ramp contractions at the two different ankle angles (10 deg dorsiflexion and 10 deg plantarflexion). Recruitment threshold was then expressed as a percentage of the MVC torque obtained at the same ankle angle. Motor unit discharge rate was measured during the different phases of the task.
Ultrasonography
The architectural changes of the tibialis anterior during shortening and lengthening contractions were investigated in each subject during a separate session by ultrasonography (Fukunaga et al. 2001; Reeves & Narici, 2003). Fascicle length was assessed by images obtained using real-time B-mode ultrasonographic apparatus (AU5; Esaote, Firenze, Italy; 13 MHz linear-array probe with a 38 mm scanning length) positioned on the skin along the mid-sagittal plane of the tibialis anterior muscle over the site corresponding to the location of the needle insertions. Once muscle fascicles had been clearly identified, the position of the probe was firmly held in place using a self-made resin sheath to provide a standardized measurement site and ensure that measurements were taken from the same position. The probe was coated with a water-soluble transmission gel to provide acoustic contact. A metallic marker was placed between the skin and the ultrasound probe to verify that the probe did not move during the recording. Images acquired during the movements were recorded at a frequency of 20 Hz. The signal from the footplate rotation was used to synchronize the ultrasound images with the ankle movement.
Images were obtained for each subject from ankle movements at torques corresponding to those recorded during motor unit recordings. With the help of visual feedback, subjects had to match target torques determined from previous experimental sessions. Measurements of fascicle length were obtained with digitizing software (Scion Image, National Institutes of Health, USA). Because the tibialis anterior is a bipennate muscle with a central aponeurosis (Reeves & Narici, 2003), fascicle length was determined as the distance from the central to the superficial aponeuroses. When the superficial end of the fascicle extended off the acquired ultrasound image, fascicle length was determined by trigonometry by assuming a linear continuation of the fascicles (Reeves & Narici, 2003).
Statistics
Data are reported as means ± s.d. within the text, and displayed as means ± s.e.m. in the figures. Torque and surface EMG during MVCs, recruitment threshold, and average discharge rate of motor units at the two ankle joint angles (10 deg plantarflexion and 10 deg dorsiflexion) were analysed using Student's paired t test. Changes in aEMG, motor unit discharge rate and muscle fascicle length during shortening and lengthening contractions were analysed by a two-way ANOVA with repeated measures. A Tukey post hoc test was conducted when significant main effects were observed. Significance was set at P ≤ 0.05.
Results
The average isometric MVC torque produced by the dorsiflexor muscles at long muscle length (10 deg plantarflexion) was greater (44.3 ± 4.2 versus 35 ± 3.3 N m; P < 0.001) than at short muscle length (10 deg dorsiflexion). In contrast, the corresponding aEMG of the tibialis anterior decreased (P < 0.05) with increased muscle length (0.43 ± 0.06 versus 0.47 ± 0.08 mV). The average MVC torque during the shortening and lengthening contractions was 28.7 ± 4.2 and 45.3 ± 3.8 N m, respectively. As expected, the torque was significantly higher (P < 0.001) during maximal lengthening contractions. In contrast, the aEMG of the tibialis anterior was lower during the lengthening than shortening MVC (0.40 ± 0.06 versus 0.43 ± 0.08 mV; P < 0.01).
Behaviour of single motor unit during shortening and lengthening contractions
The behaviour of the same motor unit was tracked during shortening and lengthening contractions of the dorsiflexors performed at a constant angular velocity. A typical example of the discharge of a single motor unit is illustrated in Fig. 1. The unit was recruited at an isometric dorsiflexion torque of 3.9 N m (10.9% MVC) and 6.6 N m (18.0% MVC) in the short and long positions, respectively. In this example, the unit was activated continuously during the isometric contraction at the two ankle positions and the shortening and lengthening contractions. The discharge rate first decreased from 9.3 to 5.3 Hz at the transition between the isometric and shortening contractions (unloading reflex) before increasing progressively up to 12.1 Hz at the end of the movement. Conversely, the discharge rate during the lengthening contraction first increased to 15.0 Hz (stretch reflex) before slowly returning to its initial value (9.4 Hz) at the end of the movement.
The same behaviour was observed for all 63 motor units. The range of recruitment thresholds, expressed as percentage of their respective isometric MVC, extended from 0.2 to 21.1% (mean ± s.d.; 5.7 ± 5.9% MVC) and from 0.5 to 32.8% (9.8 ± 8.6% MVC) at short and long muscle lengths, respectively. The difference between the mean values corresponded to a significant reduction of the recruitment threshold of 45.1 ± 24.1% (P < 0.001) when the ankle joint was moved from 10 deg plantarflexion (long) to 10 deg dorsiflexion (short).
Before the onset of the dynamic contractions, the average discharge rate computed over a 2 s epoch and across all isometric contraction intensities was 11.9 ± 2.4 and 11.7 ± 2.2 Hz for short and long muscle lengths, respectively, and did not differ statistically. In Fig. 2, motor-unit discharge rate computed across all contraction intensities during shortening and lengthening contractions (from 0 to 2 s) and the isometric contractions (from 2 to 4 s) has been expressed as a percentage of change from the values recorded during the initial isometric contractions. The typical pattern observed at the onset of shortening contraction (first 10 discharges; Fig. 2, inset), was a rapid reduction in discharge rate to 72 ± 14.3% of the initial values (from 11.7 ± 2.2 to 8.2 ± 2.5 Hz; P < 0.001) for the first interspike interval. This first depression was followed by a rapid return to initial values and a progressive increase during the remaining part of the contraction (from 0.4 to 2 s; Fig. 2, main graph). The average discharge rate was 29.1 ± 16.9% greater (14.5 ± 2.6 Hz; P < 0.001) at the end of the movement when compared with initial values. In contrast, discharge rate during lengthening contractions was first enhanced to 115.2 ± 14.0% of the initial values (from 11.9 ± 2.4 to 13.7 ± 2.7 Hz; P < 0.001; Fig. 2, inset) and roughly maintained at this level to the end of the movement (Fig. 2, main graph). However, at this stage, no significant difference was observed compared with control discharge rate (106.7 ± 12.8%; 12.4 ± 2.0 Hz; P > 0.05). The following sustained isometric contraction at both muscle lengths was associated with a progressive return to control values, and 1.4 s after the end of the shortening contraction no significant difference in discharge rate was observed.
Recruitment of additional motor units
Eighteen additional motor units were recruited during the shortening contraction. These units, recorded in six of the eight subjects, were recruited at an average isometric threshold of 7.8 ± 4.6% MVC (3.4 ± 2.1 N m) and 13.2 ± 6.1% MVC (5.4 ± 2.6 N m) at short and long muscle lengths, respectively. These additional units were recruited at an average ankle angle of 2.4 ± 4.3 deg dorsiflexion (range: from 4.8 deg plantarflexion to 8.6 deg dorsiflexion) during the shortening contractions, and derecruited during the lengthening phase of the contraction at an average ankle angle of 7.2 ± 2.9° plantarflexion (range: from 2.5 to 10 deg plantarflexion; Fig. 3A). In some inaccurate trials, when the subject overshot the target torque and a greater dorsiflexion torque during the lengthening contraction was produced, these units continued to discharge up to the end of the movement and during the following isometric contraction. When averaged throughout their activation, the discharge rate of these units was 13.1 ± 2.1, 11.4 ± 2.1 and 12.6 ± 2.9 Hz for the shortening, lengthening and isometric contractions, respectively. The average discharge rate was significantly greater (P < 0.001) during shortening compared with lengthening contraction (Fig. 3B) and differed significantly for both shortening and lengthening contractions from isometric contraction (P < 0.01).
Figure 4 displays an example of additional recruitment and derecruitment during dynamic contractions. In this example, the second unit (MU2) was recruited during the shortening contraction at an ankle angle of 1 deg dorsiflexion and derecruited during the lengthening contraction at a more extended ankle joint angle (8.0 deg plantarflexion). The second unit was recruited (shortening contraction) or derecruited (lengthening contraction) although a nearly constant and similar discharge rate of the first unit (15.2 ± 3.2 and 15.5 ± 2.0 Hz for shortening and lengthening contractions, respectively) and shows comparable changes of the general discharge pattern to the first recruited unit (MU1).
Surface EMG activity
The modulation in motor unit activity during movement paralleled the change in the surface EMG activity of the tibialis anterior, as illustrated by Fig. 5. When computed across all contraction intensities, the aEMG (percentage change) decreased to 76.4 ± 13.7% of the initial values (from 0.075 ± 0.031 to 0.058 ± 0.028 mV; P < 0.001) from the beginning of the movement to the first 0.1 s sequence of shortening contraction. Thereafter, the aEMG increased progressively up to the end of the shortening contraction and reached 134.8 ± 21.0% (0.098 ± 0.037 mV; P < 0.001) of the initial values. The aEMG declined progressively during the subsequent isometric contraction at the short muscle length, but stayed above (118.7 ± 16.4%; 0.087 ± 0.035 mV; P < 0.001) the values recorded at the long muscle length. In contrast, the aEMG reached slightly higher values (108.1 ± 6.3%; from 0.084 ± 0.035 to 0.09 ± 0.036 mV; P = 0.06) as soon as the lengthening contraction was initiated, and it remained constant up to the end of the movement. The aEMG regained its control value (96.9 ± 19.1%; 0.078 ± 0.029 mV; P > 0.05) during the subsequent isometric contraction at the long muscle length.
Coactivation of antagonist muscles (soleus and lateral gastrocnemius) paralleled the changes observed for the agonist muscles and augmented progressively with increased dorsiflexion torque during both shortening and lengthening contractions. When computed across subjects and over all contraction intensities, excluding the transition phases (first 0.4 s) at the beginning of each movement, the ratio between antagonist and agonist aEMG activity differed slightly but significantly between contraction types (Table 1). The coactivation ratio was significantly greater (P < 0.001) during lengthening compared with shortening contractions for both the soleus and lateral gastrocnemius, and during isometric contractions at long compared with short muscle length for the soleus.
Fascicle length change
The effect of a change in ankle position on fascicle length during shortening and lengthening contractions at different torque levels (from 5 to 30% MVC) is illustrated for one subject in Fig. 6A. This graph shows that regardless of ankle angles, fascicle length shortened progressively with an increase in dorsiflexion torque. Furthermore, there was no difference in the change in fascicle length during the shortening and lengthening contractions. When averaged across subjects and over the torque levels at which the 63 units were recorded (Fig. 6B), fascicle length decreased by 18.3 ± 1.4% (from 75.9 ± 2.7 to 62.0 ± 2.7 mm; P < 0.001) and increased by 22.1 ± 2.9% (from 62.2 ± 3.6 to 75.8 ± 4.0 mm; P < 0.001) during shortening and lengthening contractions, respectively. When comparing both shortening and lengthening values across all contraction intensities, no significant difference was observed throughout the 20 deg ankle range of motion (Fig. 6B). Within the range of the ankle dorsiflexion torques during the motor unit recordings, a linear change of muscle fascicle length was observed during shortening (y = 0.644x + 69.2; r2 = 0.997) and lengthening (y = 0.650x + 69.4; r2 = 0.998) contractions. When computed across both contraction modalities and contraction intensities, fascicle length changed by 13.7 ± 1.4 mm (20.1 ± 3.1%) when the ankle joint moved over a 20 deg range of motion around the neutral position. The average fascicle velocity did not differ significantly between shortening and lengthening contractions (7.0 ± 0.8 and 6.8 ± 1.2 mm s−1, respectively).
Discussion
The main finding of the current study was a difference in the modulation of motor unit discharge rate in the tibialis anterior, with no change in recruitment order between shortening and lengthening submaximal contractions performed at a constant ankle angular velocity and for a similar change in torque. Furthermore, a similar change in fascicle length at the same speed involved a greater recruitment and modulation of discharge rate of the same motor units during slow shortening contractions compared with lengthening contractions.
A major strength of our study was the measurement of average fascicle length during shortening and lengthening contractions; this provided length information on the portion of the muscle from where the motor units were recorded. The results indicated that fascicle length of the tibialis anterior varied linearly with ankle joint rotation, despite a slight variation in the moment arm of the dorsiflexors during ankle rotation (Maganaris et al. 1999). The average fascicle length changed by ∼20% when the ankle joint moved over the 20 deg range of motion and there was no statistical difference in the change in fascicle length and its average velocity between shortening and lengthening contractions. Although the range of motion examined in the current study might represent a different portion of the active torque–angle curves for the different subjects, it should have minor effect on our results because this part of the torque–angle curve is relatively flat during submaximal activations (Marsh et al. 1981) and our subjects showed a similar increase in maximal isometric torque (range 17–32%) when tested at long compared with short muscle lengths.
It is well known from animal studies on whole muscles and single fibres that force during lengthening contractions increases above isometric force when measured on the plateau and on the descending limb of the length–tension curve (Katz, 1939; Edman et al. 1978; Morgan et al. 2000). The usual explanation for this extra force is the development of sarcomere length non-uniformity in the fibres (Julian & Morgan, 1979), although the contribution of other mechanisms cannot be excluded (see Pinniger et al. 2006). Force enhancement is not observed for all muscle groups in humans during voluntary lengthening contractions, possibly due to a tension-limiting mechanism (Westing et al. 1991; Aagaard et al. 2000; Pinniger et al. 2000). However, the absolute torque produced by the dorsiflexor muscles around the neutral (90 deg) ankle angle during lengthening contraction is usually much greater than during isometric and shortening contractions (Pasquet et al. 2000; Klass et al. 2005). Due to the greater force capacity of muscle during lengthening contractions, neural activation must be augmented during a submaximal shortening contractions performed against a given load or for a similar change in torque.
Although the utility of EMG to infer changes in the voluntary drive to muscle can be misleading (Keenan et al. 2005), our results for a similar change in torque are consistent with this interpretation as a greater aEMG was reached during shortening compared with lengthening contractions. The results included transient changes in surface aEMG at the transition between the sustained isometric contraction and the movement. There was a small increase in aEMG at the onset of the lengthening contraction that was probably due to increased motor unit discharge rate (see Figs 1 and 4) caused by the sudden muscle stretch (Struppler, 1975; Roll & Vedel, 1982; Wise et al. 1999). Thereafter, the aEMG remained nearly constant during muscle lengthening and at a slightly greater level compared with isometric contractions at both lengths. In contrast, aEMG decreased initially during the shortening contraction due to the unloading reflex (Struppler, 1975; Roll & Vedel, 1982; Wilson et al. 1997); this matches the transient reduction in motor unit discharge rate (see Figs 1 and 4). Subsequently, the aEMG increased progressively until the end of the shortening contraction by the recruitment of additional motor units and increased discharge rate (see below). The greater aEMG during the sustained isometric contraction at the short muscle length compared with the long length was mainly due to the recruitment of motor units.
Due to technical difficulties, few studies have analysed the behaviour of the same motor units during shortening and lengthening contractions (Nardone et al. 1989; Howell et al. 1995; Søgaard et al. 1996; Stotz & Bawa, 2001; Semmler et al. 2002). Such recordings are more problematic during dynamic than isometric conditions because of electrode movement and the difficulty in performing shortening and lengthening contractions against an inertial load at a constant velocity. To minimize these technical difficulties and because differences in the kinematics of dynamic contractions may change the discharge pattern of motor units, we imposed an identical movement velocity during muscle shortening and lengthening. The protocol required the subject to sustain the discharge of an identified motor unit during isometric contractions at two muscle lengths and during shortening and lengthening contractions of the dorsiflexors. All motor units (n = 63) activated during the initial isometric contraction at long muscle length discharged continuously throughout the task, including the lengthening contraction. Furthermore, the recruitment of an additional motor unit was observed at 18 recording sites during the shortening contraction. These additional motor units had a greater force threshold than the units that were active from the beginning of the task and were recruited to maintain the required torque during muscle shortening. They continued to discharge during the subsequent isometric contraction at the short muscle length and were derecruited during the lengthening or isometric contraction at the long muscle length. The derecruitment during the lengthening contraction was always observed at a more extended ankle joint angle (fascicle length) than during the shortening contraction (Fig. 3A). In all the trials when the recruitment of another motor unit was observed during the shortening contraction, the unit recruited initially remained active beyond the derecruitment of the later-recruited unit. In agreement with previous studies (Garland et al. 1994; Søgaard et al. 1996; Bawa & Jones, 1999; Stotz & Bawa, 2001; Semmler et al. 2002), these observations indicate that recruitment order was preserved during slow shortening and lengthening contractions at a constant velocity.
As for studies that have compared populations of motor units (Søgaard et al. 1996; Linnamo et al. 2003; Del Valle & Thomas, 2005), the average discharge rate of the same motor unit was lower during lengthening compared with shortening contractions in our study. After the transient increase or decrease in discharge rate due to the stretch or unloading reflex, respectively (Fig. 2), the modulation of motor unit discharge differed for the two contraction types when compared for a similar change in fascicle length and torque. The rate of motor unit discharge was nearly constant during the entire lengthening contraction and slightly greater than that recorded during the isometric contractions at short and long muscle lengths. In contrast, motor unit discharge rate increased progressively up to the end of the movement during shortening contractions and reached greater values than during isometric conditions. The greatest difference between shortening and lengthening contractions was observed when the muscle was at a short length due to an increased neural activation that was required to compensate for the reduced muscle force capacity at that length. A similar behaviour was observed for motor units recruited during the course of the task. These findings indicate that rate coding is more important during shortening than lengthening contractions (Del Valle & Thomas, 2005). The greater motor unit recruitment in isometric contractions at short muscle length, although similar discharge frequencies were recorded at the two muscle lengths (Pasquet et al. 2005), further suggests that recruitment is more related to muscle length and that rate coding is critical during shortening contractions. It could be argued that the torque produced during shortening and lengthening contractions relative to their respective MVCs can be slightly less in the latter condition and might have reduced the absolute discharge rate. Although this possibility cannot be excluded, it is unlikely that the contrasting modulation of motor units discharge rate during the two contraction types would be influenced when there was a smooth change in torque between the two targets. Furthermore, in the inaccurate trials recorded during lengthening contractions, during which the subject overshot the target torque and a greater relative dorsiflexion torque was produced, motor units discharge rate did not increase when compared with the accurate trials.
The contribution of the other synergistic muscles (extensor hallucis longus, extensor digitorum longus and peroneus tertius) to the torque developed by the dorsiflexors during ankle movement could have influenced the recruitment–derecruitment and discharge of motor units in the current study. Although the relative contributions of these muscles to the net dorsiflexor torque could change during shortening and lengthening contractions, these muscles are all monoarticular with extensive retinaculum systems surrounding the distal tendons and are likely to experience only a minor relative variation in the moment arms during the small (20 deg) ankle rotation. Also, the similar excursion for these muscles suggests that the length–tension curves probably did not differ greatly (Rassier et al. 1999). Furthermore, when an additional motor unit was derecruited during muscle lengthening, the constant discharge rate of the unit that remained active (see Fig. 4) indicates that the synaptic excitatory drive to the motor neurone pool was not greatly reduced at a time the second unit ceased to discharge.
The similar recruitment order during shortening and lengthening contractions in our study implies that the nervous system employs a single size-related strategy to activate the involved motor neurones in the different types of contractions. The contrasting modulation of motor unit discharge rate in the two contraction types, however, indicates a difference in the distribution of the sensory inputs to the motor neurone pool together with a possible change in the supraspinal command (Nielsen, 2004). Spinal networks do appear capable of controlling afferent input for specific tasks (Rudomin, 1999). Accordingly, the greater amplitude of the motor evoked potential induced by transcranial magnetic and electrical stimulation during shortening contractions suggests that excitation of the motoneurone pool is reduced during lengthening contractions (Abbruzzese et al. 1994; Sekiguchi et al. 2003). Furthermore, the amplitude of the Hoffmann (H) reflex appears to be modulated in a similar manner as for transcranial stimulations (Romano & Schieppati, 1987; Abbruzzese et al. 1994; Nordlund et al. 2002). Despite a slightly greater level of coactivation during lengthening contractions in the present study, the similar modulation of the H reflex at rest and during contraction (Abbruzzese et al. 1994) discounts a key role for reciprocal Ia inhibition and autogenic Ib inhibition in the excitation of the motoneurone pool. Rather, the comparable variation in both magnetically and electrically evoked motor responses and in the H reflex suggests that the effect was mediated by mechanisms located in the spinal cord, presumably presynaptic in origin (Abbruzzese et al. 1994). Because muscle spindle activity is increased to a greater extent during lengthening than shortening contractions (Burke et al. 1978), centrally and peripherally regulated presynaptic mechanisms (Hultborn et al. 1987; Morita et al. 1998) might explain the different modulation of motor unit discharge rate during the two contraction types. This possibility is consistent with the observation that, regardless the level of voluntary drive, the mean discharge rate of motor neurones is lower when deprived of muscle afferent feedback (Macefield et al. 1993). Because lengthening contractions are more difficult to control than shortening contractions (Nordlund et al. 2002; Semmler et al. 2002), depression of Ia excitation from muscle spindles may facilitate an accurate performance of the task.
In conclusion, the current study demonstrated that submaximal lengthening contractions at constant velocity involved a specific modulation of motor unit discharge rate with no change in motor unit recruitment order. This behaviour contrasted with that observed during shortening contractions, despite a similar linear change in fascicle length and torque during the two tasks. These observations indicate that recruitment order is preserved during submaximal lengthening contractions at a slow constant velocity, but that motor unit discharge is modulated less compared with shortening contractions.
Acknowledgements
The authors are particularly grateful to Professor R. Enoka and Dr K. Keenan for useful comments on this paper. We would also like to thank Professor M. Narici for his helpful advice regarding ultrasonography, and Ms A. Deisser for assistance in the preparation of the manuscript. This study was supported by the Université Libre de Bruxelles and the Fonds National de la Recherche Scientifique of Belgium.
Footnotes
-
(Resubmitted 26 July 2006; accepted after revision 6 September 2006; first published online 6 September 2006)
- 2006 The Authors. Journal compilation © 2006 The Physiological Society
References
Figure 1. Behaviour of a single motor unit during isometric and dynamic contractions A typical discharge pattern of the same motor unit recorded in the tibialis anterior is illustrated during a sustained isometric contraction and during shortening (A) and lengthening (B) contractions of the dorsiflexor muscles. For both conditions, angular ankle displacement (a), the torque produced during dorsiflexion (b), rectified surface (c) and intramuscular (d) EMG of the tibialis anterior, and instantaneous discharge rate (e) of the motor unit are illustrated. F, the action potentials of the identified unit are superimposed with an expanded display. The recruitment threshold of the unit was 3.9 N m (10.9% maximal voluntary contraction (MVC)) and 6.6 N m (18.0% MVC) in short and long positions, respectively. The discharge rate was first decreased from 9.3 to 5.3 Hz at the transition between the isometric and shortening contractions before increasing progressively up to 12.1 Hz at the end of the movement. Conversely, for the lengthening contraction, the discharge rate was first increased from 10.0 to 15.0 Hz before slowly returning to its initial value (9.4 Hz) after the end of the movement. The vertical dashed lines indicate the beginning and the end of the movement in each condition.
Figure 2. Change in average discharge rate in motor units during isometric and dynamic contractions The average discharge rate is reported during shortening (•) and lengthening (ˆ) contractions (from 0 to 2 s), and during isometric contractions (from 2 to 4 s) at short (▪) and long (□) muscle lengths. Each value, expressed as percentage of the discharge rate recorded during the initial isometric contractions, represents the average (±s.e.m.) over 0.2 s bins for all motor unit (n = 63) computed across all contraction intensities. The inset illustrates the average (±s.e.m.) changes in discharge rate for the first 10 discharges at the transition between the sustained isometric contraction and the onset of movement. The horizontal dashed line represents the average discharge rate during the previous isometric contraction. Significant difference from initial value in each conditions: *P < 0.05, **P < 0.01, ***P < 0.001. Significant difference between the two conditions: +P < 0.05, ‡P < 0.01, †P < 0.001.
Figure 3. Characteristics of the additional motor units recruited during shortening contractions A, each line shows the recruitment threshold of 1 of 18 motor units recruited during a shortening contraction (•) and its derecruitment threshold during the subsequent lengthening contraction (ˆ). Each threshold is expressed as a function of ankle angle (°) around the neutral position (horizontal dashed line), and the negative and positive values correspond to dorsi- and plantarflexion, respectively. B, comparison of the average discharge rate of these units during shortening (•) and lengthening (ˆ) contractions. For each unit, the average discharge rate was computed from its recruitment to the end of the shortening contraction (shortening) and from the beginning of lengthening contraction to its derecruitment (lengthening)
Figure 4. Behaviour of an additional motor unit recruited during shortening contraction A typical discharge and recruitment pattern of two motor units (MU1 and MU2) recorded in the tibialis anterior during a sustained isometric contraction and during shortening (A) and lengthening (B) contractions. In both conditions, angular ankle displacement (a), rectified surface (b) and intramuscular (c) EMG of the tibialis anterior, and instantaneous discharge rate (d and e) of MU1 (d) and MU2 (e) are illustrated. Action potentials of MU1 (f) and MU2 (g) are superimposed with an expanded display. MU2 was recruited during the shortening phase at an ankle angle of 1 deg dorsiflexion and derecruited during the lengthening contraction at a more extended ankle joint angle (8.0 deg plantarflexion). Note that the discharge rate of the first motor unit remained relatively constant when the second unit was recruited (shortening) and derecruited (lengthening contraction). The vertical dashed lines indicate the beginning and the end of the movement in each condition.
Figure 5. Change in aEMG during isometric and dynamic contractions The surface average full-wave rectified EMG (aEMG) of the tibialis anterior is reported during shortening (•) and lengthening (ˆ) contractions (from 0 to 2 s), and during isometric contractions (from 2 to 4 s) at short (▪) and long (□) muscle lengths. Each value represents the average (±s.e.m.) over 0.2 s bins for all subjects and trials computed across all contraction intensities. The horizontal dashed line represents the average aEMG during the previous isometric contraction. Significant difference from initial value in each conditions: **P < 0.01, ***P < 0.001. Significant difference between the two conditions: +P < 0.05; †P < 0.001.
Figure 6. Change in muscle fascicle length during dynamic contractions The average fascicle length of the tibialis anterior when the ankle joint moved over a 20 deg range of motion around the neutral position (0 deg) is reported during shortening (•; from 10 deg plantarflexion to 10 deg dorsiflexion) and lengthening contractions (ˆ; from 10 deg dorsiflexion to 10 deg plantarflexion). A, the effect of a change in ankle position on fascicle length during shortening and lengthening contractions at different torque levels (from 5 to 30% MVC by steps of 5% MVC) is illustrated for one subject. B, average (±s.e.m.; n = 63) fascicle length computed across subjects and over contraction intensities corresponding to the torque recorded during motor unit recordings. A linear change between muscle fascicle length and ankle joint ankle is obtained for shortening (y = 0.644x + 69.2; r2 = 0.997) and lengthening (y = 0.650x + 69.4; r2 = 0.998) contractions. Note the similar change in fascicle length between shortening and lengthening contractions for different torque levels in a single subject and when averaged across subjects and over contraction intensities.
Table 1. Coactivation ratio for the soleus and the lateral gastrocnemius during different contraction types with the tibialis anterior
