941
J Physiol 564.3 (2005) pp 941–951
Phase-dependent and task-dependent modulation of stretch reflexes during rhythmical hand tasks in humans Ruiping Xia1,2 , Brian M. H. Bush1 and Gregory M. Karst2 1 2
Department of Physiology, School of Medical Sciences, University of Bristol, Bristol, UK Division of Physical Therapy, University of Nebraska Medical Center, Omaha, NE 68198, USA
Phase-dependent and task-dependent modulation of reflexes has been extensively demonstrated in leg muscles during locomotory activity. In contrast, the modulation of reflex responses of hand muscles during rhythmic movement is poorly documented. The objective of this study was to determine whether comparable reflex modulation occurs in muscles controlling finger motions during rhythmic, fine-motor tasks akin to handwriting. Twelve healthy subjects performed two rhythmic tasks while reflexes were evoked by mechanical perturbations applied at various phases of each task. Electromyograms (EMGs) were recorded from four hand muscles, and reflexes were averaged during each task relative to the movement phase. Stretch reflexes in all four muscles were found to be modulated in amplitude with respect to the phase of the rhythmic tasks, and also to vary distinctly with the tasks being conducted. The extent and pattern of reflex modulation differed between muscles in the same task, and between tasks for the same muscle. Muscles with a primary role in each task showed a higher correlation between reflex response and background EMG than other muscles. The results suggest that the modulation patterns observed may reflect optimal strategies of central–peripheral interactions in controlling the performance of fine-motor tasks. As with comparable studies on locomotion, the phase-dependency of the stretch reflexes implies a dynamically fluctuating role of proprioceptive feedback in the control of the hand muscles. The clear task-dependency is also consistent with a dynamic interaction of sensory feedback and central programming, presumably adapted to facilitate the successful performance of the different fine-motor tasks. (Received 28 December 2004; accepted after revision 1 March 2005; first published online 3 March 2005) Corresponding author R. Xia: Division of Physical Therapy, University of Nebraska Medical Center, Omaha, NE 68198, USA. Email:
[email protected]
Phase-dependent modulation of reflex responses to perturbation during locomotion, in vivo or ‘fictive’, has been widely reported in both vertebrate and invertebrate animal species, and in humans (Forssberg et al. 1975, 1977; Forssberg, 1979; Duysens & Loeb, 1980; Akazawa et al. 1982; Skorupski & Sillar, 1986; Capaday & Stein, 1986, 1987; Belanger & Patla, 1987; Edamura et al. 1991; Skorupski et al. 1992; Brooke et al. 1995). As first demonstrated by Forssberg et al. (1975), a stimulus applied to the dorsum of the foot of a cat elicited an excitation of the flexors during the swing phase and an excitation of extensors during the stance phase. Subsequently, many investigators have examined the dependency of the cutaneomuscular reflex, stretch reflex or H-reflex on the phase of the step cycle during different forms of locomotion, including walking, running and standing, in animal preparations and increasingly in humans (Duysens et al. 1996; Schillings et al. 1999; Simonsen & Dyhre-Poulsen, 1999). Some recent studies C The Physiological Society 2005
on the modulation pattern of the soleus H-reflex during human walking and running (Simonsen & Dyhre-Poulsen, 1999; Ferris et al. 2001) reflect some controversy with earlier studies (Capaday & Stein, 1987; Edamura et al. 1991). Moreover, the degree of reflex modulation has also been found to change with different motor activities and postures, e.g. walking, running and standing. It appears that the pattern of modulation can be specifically modified for different functional requirements of each motor task. It has been suggested that this task-dependent modulation, like the phase-dependent modulation referred to above, is regulated by ‘central pattern generators’ (CPGs) within the brain stem and spinal cord interacting with sensory feedback from the periphery (Stein, 1995). (For reviews of these and related studies, see Stein & Capaday, 1988; Skorupski, 1991; Brooke et al. 1997; Zehr & Stein, 1999; McCrea, 2001; Dietz, 2002.)
DOI: 10.1113/jphysiol.2004.082271
942
R. Xia and others
Many studies of reflex activity in the human upper limb muscles have explored central–peripheral interactions controlling arm or finger movements (Caccia et al. 1973; Evans et al. 1989; Matthews, 1989, 1993; Chen & Ashby, 1993; Nakazawa et al. 1997; Wallace & Miles, 1998; Brooke et al. 2000; Zehr & Chua, 2000; Zehr & Kido, 2001; Zehr et al. 2003). However, most investigations of stretch reflexes in the upper arm were undertaken during steady state conditions (e.g. Matthews, 1989, 1993; Noth et al. 1991; Doemges & Rack, 1992a,b; Carter et al. 1993). A few studies have been made of modulation of stretch reflexes during arm tracking movements (Dufresne et al. 1980; Johnson et al. 1991), and Zehr and colleagues have recently examined the modulation patterns of cutaneous reflexes and H-reflexes during rhythmic arm movements (Zehr & Chua, 2000; Zehr & Kido, 2001; Zehr et al. 2003). Little attention has been paid to the modulation of stretch reflexes or their phase- and task-dependency during rhythmic, fine-motor tasks involving the hand and fingers. No report has been seen on how sensory feedback interacts with the central motor drive in regulating the hand muscles while performing rhythmic finger tasks related to handwriting. Nevertheless the importance of proprioceptive feedback and its interaction with the CPG or any central motor pattern in the control of smooth, voluntary and involuntary movements cannot be too strongly emphasized (Stuart, 2002). Accordingly, the purpose of this study was to examine the behaviour of reflex responses in selected hand muscles, and to assess the role of proprioceptive feedback in mediating skilled, rhythmical actions related to handwriting. Our hypotheses were: (1) that the amplitude of reflex responses to mechanical perturbation during rhythmic motor performance is modulated with respect to the phase of the movement cycle (phase-dependency), and (2) that the pattern of reflex modulation in each muscle is dependent on the particular fine-motor task performed (task-dependency). Some of the results obtained from a preliminary study have been communicated previously (Xia & Bush, 1996). Methods Subjects
Twelve healthy subjects (7 men and 5 women), ranging in age from 25 to 65 years, participated in this study. No subject had a history of neurological disorders or musculoskeletal disorders affecting their upper limbs. Informed consent was obtained from each subject before participating in the experiment. The protocol was approved by the Ethical Committee of the University of Bristol, UK, and was performed in accordance with the Declaration of Helsinki. All subjects were right-hand dominant and held a pen using the standard dynamic tripod grip described by Wynn-Parry (1966).
J Physiol 564.3
Electromyographic (EMG) recording
EMG activity was recorded simultaneously from two intrinsic hand and two extrinsic forearm muscles: first dorsal interosseous (FDI) and flexor pollicis brevis (FPB), and flexor digitorum superficialis (FDS) and extensor pollicis brevis (EPB). The areas of skin chosen for electrode placement were gently abraded using special skin preparation pads (Biolect, UK). The skin was then cleaned with alcohol and allowed to dry completely before the placement of the electrodes. Bipolar surface EMG recordings were obtained using a pair of 9 mm diameter Ag–AgCl bio-adhesive electrodes (Biotrace, UK), 2 cm apart, attached to the skin over the belly of each muscle. A reference electrode was positioned over the sternum. Four pairs of electrodes and the reference electrode were connected to a 4-channel isolated preamplifier, the proximal electrode being linked to the positive terminal and the distal one to the negative terminal, and the reference electrode to the ground. Before the recording session began, the placement of the electrodes over each muscle was checked by monitoring the EMG activity displayed on an oscilloscope during rapid alternating contraction and relaxation of each muscle. For example, when checking FPB electrode placement, the subject was asked to perform alternating flexion and extension of the proximal phalanx of the thumb, while the EMG signal was displayed on the oscilloscope for confirmation. The EMG signals from each of the four muscles were amplified by isolated preamplifiers (× 10) and main amplifiers (× 100), and were band-pass filtered (bandwidth 10 Hz to 5 kHz). The amplified EMG signals were then sampled, along with the force and stimulus signals (see below), at a rate of 2000 Hz per channel by a 16-bit A/D converter CED 1401 (Cambridge Electronic Design, Cambridge, UK), and stored on a desktop computer for subsequent analysis. Force-measuring pen
A pen with strain gauges on three sides was used in one of the two tasks studied, ‘rhythmic pen squeezing’ (RPS; see below and Fig. 1), to measure the forces exerted by the thumb and fingers, and also to identify the onset of each cycle of the rhythmic task. The pen allowed the forces exerted along three axes, 120 deg apart and perpendicular to the long axis of the pen, to be independently measured. A standard ballpoint pen refill was housed in a brass hexagonal tube, machined to produce three flexible beams equally spaced around the circumference. A foil strain gauge bonded onto each beam was electrically connected as one arm of a Wheatstone bridge, and the output signals from each transducer were suitably amplified, level shifted and low-pass filtered (25 Hz), before being sampled by the CED 1401. C The Physiological Society 2005
J Physiol 564.3
Phase-dependent reflex modulation in rhythmic hand tasks
Figure 1. Photograph of the apparatus and experimental setup with a subject participating Illustration of the protocol showing the task ‘rhythmic pen squeezing’ (RPS). The same actuator was used in both rhythmical tasks of this study.
Mechanical stimulation
A linear electromagnetic actuator in series with a force transducer was used to produce mechanical stimulation in this study (Fig. 1). The actuator’s moving coil was located in the air gap between the pole pieces of a permanent magnet by a spring suspension. This provided a restoring force so that the device could be operated without feedback. The force generated by the coil was directly proportional to the current flowing through it. The force transducer comprised four foil strain gauges bonded as two back-to-back pairs onto a brass ring connected to the actuator’s output rod, which was located by a sleeve bearing. The strain gauges were electrically connected as a bridge with four active arms, and the output signal was suitably amplified, level shifted and filtered before being digitized and sampled. Mechanical stimuli (1 mm in amplitude and 100 ms duration) were applied via a 12 mm diameter disc fixed onto the moving end of the actuator, which was driven with a power amplifier by a series of square wave pulses
Figure 2. Original record showing the constancy of the mechanical stimuli, randomly applied at different phases of the movement cycle Recordings of the mechanical stimulus pulses (S), the perturbations applied to the pen-tip (P), the force exerted by the index finger on the force-measuring pen (F), and (lower traces) the concurrent raw EMGs in the four hand muscles as indicated, during the rhythmic pen squeezing task in one subject. The mean velocity of the applied stretches was 5.505 ± 0.282 m s−1 (S.D.), ranging from 4.97 to 5.79 m s−1 . FDI, first dorsal interosseous; FDS, flexor digitorum superficialis; EPB, extensor pollicis brevis; FPB, flexor pollicis brevis. C The Physiological Society 2005
943
from a function generator. Interstimulus intervals varied randomly, between 0.6 and 1.2 s, to prevent anticipation. The minimum interval was selected to include all possible reflex components and to avoid potential adaptation of the response, and the maximum interval was chosen to obtain a sufficient number of stimuli to average responses whilst minimizing fatigue. Stimuli were applied to one of two sites, either the lateral side of the proximal interphalangeal (PIP) joint of the index finger or the tip of the force-measuring pen, depending on the task performed (see below in ‘Experimental protocol’). Figure 2 illustrates a series of randomly delivered stimuli and the resulting force perturbations, along with EMGs of four muscles, during one task in a participating subject. The successive perturbations remained almost uniform in intensity and shape at the different phases of the cycle. The mean velocity of the applied stretches shown in Fig. 2 was 5.505 ± 0.282 m s−1 (mean ± s.d.). In order to examine the phase-dependent modulation of the reflex output, it is critical to ensure that the perturbation is constant at various phases of the cycle. Experimental protocol
Prior to recording, the subject’s maximal voluntary force (MVF) was measured under each condition associated with each task being performed. Subjects sat on a height-adjustable chair and performed maximal voluntary contraction under two static conditions with their dominant hand: (1) abducting the index finger as hard as possible with the lateral side of the index finger PIP joint pushing against a 12 mm diameter firm rubber disc on a cantilever beam with strain gauges bonded to it, and (2) squeezing the force-measuring pen as hard as possible. The MVF from the index finger was recorded from each subject during both task conditions. Three reproducible readings were taken to obtain a consistent maximum in each condition.
944
R. Xia and others
Subjects were then instructed to perform a standard protocol incorporating two rhythmic motor tasks. In each task, they followed a metronome pulse at approximately 0.8 Hz for 3 min, while monitoring on an oscilloscope the force exerted by the index finger of the dominant hand. The two tasks were: (1) rhythmic finger abduction (RFA) – abducting the index finger at a frequency of 0.8 Hz isometrically between two force limits (10% and 20% of MVF), with the lateral side of the PIP joint pressing against a 12 mm disc mounted on the force transducer in series with the mechanical actuator; and (2) rhythmic pen squeezing (RPS) – squeezing the force transducer pen with a dynamic tripod grip, so as to vary the force exerted by the index finger cyclically between 10% and 20% of MVF, while holding the pen-tip on the mechanical actuator with a constant force of about 10% of maximum downward pressure (Fig. 1). In both tasks, subjects produced finger forces isometrically, while constant perturbations were applied at varying phases of each task cycle, as illustrated in the example of Fig. 2. During the performance of each task, a sufficient period of rest was taken between recordings to avoid the incidence of muscle fatigue. Recordings were also made for 30 s while subjects performed the same manoeuvres without any mechanical stimulation. These were recorded in order to subtract the baseline EMG activity from the stimulated recording in subsequent analyses. Data analyses
EMG data were stored on a computer for off-line analysis. The digitized EMGs were first rectified, low-pass filtered (250 Hz) and then averaged between cycles with respect to phase, or averaged time-locked to the stimulus. Each rhythmic cycle was separated by the troughs of the sinusoidal force signal produced by the index finger on the mechanical actuator for the RFA task or on the force-measuring pen for the RPS task. To examine the dependency of the reflex responses on the phase of these rhythmic tasks, each task cycle was divided into 8 or 16 parts (phases) of equal duration. All responses occurring within the same phase of all the cycles in a trial were then grouped and averaged together, for comparison with the averages from other phases (Akazawa et al. 1982; Yang & Stein, 1990). Typically, each trial consisted of a total of approximately 200 stimuli, and the number of stimuli occurring in each phase of the task cycle was approximately equal. Stimulus onset was taken to be the rising phase of the stimulus pulse. Latency of the reflex response was defined as the duration from the stimulus onset to the time at which the reflex EMG response, which was usually clearly identifiable, exceeded the pre-stimulus EMG level, defined as the mean EMG amplitude for the 50 ms prior to the
J Physiol 564.3
stimulus onset. The reflex latency normally did not vary over the task cycle in each muscle. Therefore the same latency measurement, based on the computed average of all responses in a given trial, was used to calculate the magnitude of the responses in all phases of the task cycle. To obtain a measurement of the net reflex component, the unperturbed recording was subtracted from the stimulated recording. The magnitude of each reflex response, calculated from the subtracted response, was defined as the difference between the averaged EMG amplitude over a specified time window and the mean prestimulus EMG level. The time window was specified as the period between the intersection points of the EMG response and the pre-stimulus EMG level. The size of the response was thus measured as the EMG average within the above window period. This process was repeated for all phases of the task cycle. Thus, the magnitude of the net reflex as a function of the cycle phase was obtained. Reflex magnitude was also examined in parallel with background EMG, as represented by the ‘unperturbed recording’, within each phase of the rhythmic task. To examine the feature of task-dependency, the responses occurring at various phases of the cycle were plotted against the mean level of normalized background EMG activity in each phase of the task cycle. Reflex magnitude was not normalized as the background level had been removed from the measurement. In addition, linear regression fit was applied to evaluate the correlation (r 2 ) between the reflex amplitude and background EMG, and to assess the task-dependency and sensitivity of the reflex response (slope of the regression line). Analysis of variance (ANOVA) and paired t tests were applied to examine the significance of the correlation and the difference in slope of the regression, at a statistically significant level P < 0.05. Results Reflex responses elicited during rhythmic tasks
Reflex responses were elicited in each muscle by mechanical stimuli while subjects performed the RPS task. Figure 3 shows for one subject averaged EMG responses from all stimuli in each of the four hand muscles studied, irrespective of the phase in which the stimuli occurred. Short-latency responses were evident in both intrinsic (FDI, FPB) and extrinsic muscles (FDS, EPB), whereas long-latency responses appeared only in the two intrinsic muscles. In all 12 subjects, the short-latency component was consistently observed in each of the four muscles, while the long-latency component was only recorded in the two intrinsic hand muscles in 11 subjects. Table 1 lists the mean latency of the short-latency EMG component for each of the four muscles in the C The Physiological Society 2005
Phase-dependent reflex modulation in rhythmic hand tasks
J Physiol 564.3
945
Table 1. Mean latencies of the short-latency stretch reflexes evoked in the four muscles during the RPS (rhythmic pen squeezing) task in all 12 subjects Muscle
Mean (ms)
S.D.
n
30.7 21.2 22.8 30.5
1.9 2.4 2.6 3.2
12 12 12 12
FDI FDS EPB FPB n, number of subjects.
RPS task. The short-latency component is generally considered as the monosynaptic stretch reflex, as its latency is compatible with monosynaptic excitation involving Ia muscle afferents. The reflex circuits mediating the longer-latency component have been under debate since it was first reported (Hammond, 1954). Results presented in the remainder of this paper are based solely on the short-latency component, as it is consistently seen in all four muscles in all 12 subjects. Two of the four muscles studied, FDI and FDS, are prime movers for the tasks RFA and RPS, respectively, and only data from these two muscles are presented in the following results. Phase-dependent modulation of reflex response during rhythmic finger tasks
In order to examine whether the reflex responses were dependent on the cycle phase, the cycle period was divided into 16 phases of equal duration for the RPS task. For the RFA task the cycle period was divided into eight phases, as no stimuli occurred in one phase when it was divided into 16 phases in two subjects. Responses were averaged separately for each of the 8 or 16 phases of the cycle period. In general, approximately 10–15 stimuli occurred in each of the 16 phases, and about 20–30 stimuli occurred in each of the 8 phases. Reflex responses were strongly modulated as a function of the phase of the task cycle. Figure 4 shows an example of the reflex response in FDI (index-finger abductor) within each phase of the task cycle for the RFA task in one subject. The first phase covered the start of abduction. The reflex response of FDI during the RFA task increased progressively during the period of rising force, reaching its largest amplitude shortly before the background EMG activity reached its highest level. The reflex then gradually decreased with the reduced EMG activity during the falling phase of abduction force. In this example, the reflex response is clearly modulated in a phase-dependent way. In addition, the depth and pattern of reflex modulation varied with both the muscle and the task, and will be described later. To eliminate the contribution of the background EMG activity, the net reflex size was measured from the subtracted recordings, i.e. the unperturbed recordings C The Physiological Society 2005
Figure 3. Reflex responses evoked by mechanical stimuli during a rhythmic task in a typical subject Reflex EMG responses elicited in four hand muscles during the rhythmic pen squeezing task in one subject. Each of the four traces represents an average of approximately 200 sweeps, irrespective of the phase in which the stimuli occurred. Stimuli were applied at time zero as indicated by arrow (S).
946
R. Xia and others
subtracted from the perturbed ones within each corresponding phase. This process was applied to all cycle phases in muscles FDI and FDS for both tasks, RFA and RPS. Figure 5 plots the average values of the reflex EMG responses, and the mean background EMG levels (the latter normalized to the peak EMG value), as a function of the phase of the rhythmic task cycle. Reflex magnitude for each phase appeared to change roughly in parallel with the background EMG activity within the same phase. Correlation analyses quantified the degree of these parallel changes. Results indicated that the response of FDI muscle was significantly correlated with the background EMG level in both tasks, with a stronger correlation observed in the RFA task (r 2 = 0.896, P < 0.001 in Fig. 5A), compared to the RPS task (r 2 = 0.708, P < 0.001 in Fig. 5C). The response of FDS muscle was significantly correlated with background activity in the RPS task (r 2 = 0.525, P < 0.01 in Fig. 5D), but insignificantly correlated with background activity in the RFA task (r 2 = 0.273, P > 0.05 in Fig. 5B). The foregoing observations clearly show that the magnitude of the reflex responses varied with the phase
Figure 4. Phase-dependency of reflex responses at various phases of a cyclical finger movement A, modulation of FDI reflex responses evoked in 8 successive phases of the full task cycle during rhythmic finger abduction in one subject. Each trace represents the average of the individual responses to approximately 25 stimuli occurring in the phase indicated on the left. B, average of background EMG over the full cycle, and resultant force exerted by the index finger, aligned vertically in relation to the 8 phases shown in A. Force: arbitrary units (increasing towards the right).
J Physiol 564.3
of the task cycle, indicating that the reflex response is modulated in a phase-dependent manner. It is also evident that the extent and pattern of this modulation varied according to the muscle being examined and the task being investigated.
Task-dependent modulation of reflex response during rhythmic finger tasks
In order to examine (1) whether modulation of the reflex is dependent on the task performed, and (2) whether reflex modulation is solely determined by the α-motoneurone excitation level during the rhythmic finger tasks, reflex magnitude was plotted as a function of normalized background EMG within the same phase in which the reflex was elicited. Figure 6 illustrates the relationships between the reflex response and the corresponding background EMG activity in muscles FDI and FDS for both tasks in all 12 subjects. Figure 6 indicates that, at the same background activity level, FDI reflex magnitude evoked during the RFA task was higher than that during the RPS task (Fig. 6A). It is obvious that the ratio of the FDI reflex response to background EMG activity is greater for RFA than for RPS. As described above and illustrated in Fig. 5, the FDI reflex elicited during the RFA task had a stronger correlation with background activity level than that during the RPS task (Fig. 5A cf. Fig. 5C). In the FDS muscle, reflex magnitude was significantly correlated with background activity in the RPS task (P < 0.01, Fig. 5D), but not in the RFA task (P > 0.05, Fig. 5B). As shown in Fig. 6B, the reflex magnitude of FDS was higher during RPS than during RFA at the same level of background EMG. Apparently, the ratio of the FDS reflex response to background EMG was higher for RPS than for RFA. These results suggest that the reflexes examined in this study are also modulated in a task-dependent manner, and that the reflex is not simply a passive reflection of the pre-existing voluntary contraction. Figure 6 shows that the pattern of reflex modulation varied in terms of the slope of the regression line, according to the muscle in combination with the task, thus indicating a difference in the sensitivity of the reflex response. The regression line for the FDI muscle was steeper for the RFA task than for the RPS task (paired t test, P < 0.05), whereas the reverse was true for the FDS muscle (t test, P < 0.01). Among all four plots, the highest slope was observed in FDS for the RPS task, while the smallest slope was obtained in the same muscle for the RFA task. Anatomically, FDI plays a primary role in abducting the index finger, the active component of the RFA task, whereas the primary function of FDS is to perform finger flexion, the key element in the RPS task. Both muscles are also intimately involved in handwriting. Thus, the pattern of reflex modulation differs C The Physiological Society 2005
Phase-dependent reflex modulation in rhythmic hand tasks
J Physiol 564.3
between tasks for the same muscle, and between muscles in a given task. Discussion The findings obtained in this study show that stretch reflexes elicited in both intrinsic and extrinsic hand muscles acting on the index finger are strongly modulated in magnitude with respect to both phase and task during rhythmic manual tasks. The dependency of modulation of reflex magnitude on the cycle phase within a rhythmic task is referred to as ‘phase-dependent’, and the different reflex response pattern observed as a result of different task conditions is referred to as ‘task-dependent’. Determinants of phase-dependent modulation during rhythmic finger tasks
As presented in Results, the modulation patterns of the reflexes studied here are highly phase-dependent (see Fig. 4), in both the rhythmic finger abduction (RFA) and the rhythmic pen squeezing (RPS) tasks (Fig. 5). These reflexes appear to be modulated to adapt
A FDI response (µV)
Reflex Background EMG
140 120 100 80 60 40 20 0
0.8 0.6 0.4 0.2 0 1
2
3
4
5
6
7
8
1.2 1 0.8 0.6 0.4 0.2 0
0
1
2
3
4
5
6
Phase of movement cycle
7
8
1.2 1 0.8 0.6 0.4 0.2
2
4
6
8
10
12
14
0 16
FDS in RPS
D
Reflex Background EMG
140 120 100 80 60 40 20 0
Reflex Background EMG
0
FDS in RFA
B
FDI in RPS 140 120 100 80 60 40 20 0
1.2 1
0
FDS response (µV)
motor programmes to match the changing conditions. Suppose that a stimulus occurred during the rising phase of abduction force in the RFA task. In order to continue index finger abduction, the finger has to overcome the interruption from mechanical perturbation. The large stretch reflex elicited in first dorsal interosseous (FDI) muscle at that moment is well adapted to compensate for the perturbation. During the falling phase of abduction force, however, the FDI muscle reacts much less to the same stimulus: a strong stretch reflex would now be inappropriate. The reflex response elicited in flexor digitorum superficialis (FDS) compensated in an analogous way in successfully accomplishing the RPS task. Similar patterns of phase-dependent reflex modulation have been reported in leg muscles during locomotion, such as walking and running in man (Capaday & Stein, 1986, 1987; Dietz et al. 1990; Edamura et al. 1991). It was found that the H-reflex in the human soleus muscle was strongly modulated in amplitude as a function of the phase of the gait cycle. The reflex increased progressively during the stance phase of the step cycle and reached its peak value in the latter half of the stance phase. This large
C
FDI in RFA
947
Reflex Background EMG
140 120 100 80
1.2 1 0.8 0.6
60 40 20 0
0.4 0.2 0 0
2
4 6 8 10 12 14 Phase of movement cycle
16
Figure 5. Patterns of phase-dependency of reflex modulation compared with background EMG cycle Modulation of reflex EMG responses elicited in FDI and FDS as a function of the phase of two rhythmic hand tasks, compared with background EMG activity, obtained in all 12 subjects. A and B, average of reflex response and background EMG activity in 8 equal phases for the RFA (rhythmic finger abduction) task in FDI and FDS. C and D, average of reflex response and background EMG activity in 16 equal phases for the RPS (rhythmic pen squeezing) task in FDI and FDS.Y-axis on the left of each graph shows mean reflex amplitude for each phase; secondary Y-axis on the right represents background EMG (normalized). Error bars (S.E.M.) are for reflex responses (omitted from background EMG for clarity). C The Physiological Society 2005
948
R. Xia and others
reflex response would serve to assist with the propulsive phase. In contrast, the reflex was small or absent during the swing phase of the step cycle, and this would avoid ankle plantar-flexion which could result in dragging the toe on the ground. Thus, the increase in the reflex during the stance phase and the decrease or absence of a reflex during the swing phase are closely matched to the functional requirements of locomotion. The same pattern of reflex modulation is observed in running as in walking, but the slope of the relationship between reflex and voluntary EMG activity is smaller for running than for walking: that is, the ‘sensitivity’ of the reflex is higher for walking than for running.
A
120
Response of FDI (µV)
110 FDI in RFA
100 90 80 70
FDI in RPS
60 50 40 30 20 0.3
0.5
0.7
0.9
1.1
Normalized background EMG
B
120
Response of FDS (µV)
110 100
FDS in RPS
90 80 70 FDS in RFA
60 50 40 30 20 0.3
0.5
0.7
0.9
1.1
Normalized background EMG Figure 6. Task-dependency of reflex modulation as a function of background EMG activity Relationship between the reflex EMG response and normalized background EMG for the tasks RFA (rhythmic finger abduction) and RPS (rhythmic pen squeezing) in muscles FDI and FDS, averaged across 12 subjects. A, correlations and regression lines for FDI in tasks RFA ( e, r 2 = 0.896, slope k = 100.14 ± 17.88) and RPS (, r 2 = 0.708, slope k = 59.76 ± 6.99). The difference in slope is statistically significant between the tasks (P < 0.05). B, correlations and regression lines for FDS in tasks RFA (•, r 2 = 0.273, slope k = 22.55 ± 3.49) and RPS (, r 2 = 0.525, slope k = 123.27 ± 8.43). The difference in slope is also significant (P < 0.01). Standard errors are omitted for clarity (see Fig. 5).
J Physiol 564.3
Factors of task-dependent modulation during rhythmic finger tasks
The foregoing discussion has shown that the reflex modulation is strongly dependent on the phase of the task cycle. To what extent does this modulation vary with different motor tasks? The evidence presented here clearly shows that the finger muscle reflexes are indeed task-dependent, and are well adapted to the particular needs of each task. First, the cyclical pattern of reflex modulation, i.e. the shape of the curve showing the variation in reflex amplitude with phase of the task cycle (Fig. 5), differed between tasks for the same muscle, as well as between muscles in a given task. Second, comparing reflex magnitude for each muscle in the two tasks, the FDI reflex during the RFA task was generally larger than during the RPS task, whilst the FDS reflex was greater in RPS than in RFA. Third, the difference in slope of the four regression lines representing the two muscles in each task (Fig. 6) indicates that the reflex sensitivity varied both with the tasks performed and with the muscles investigated. The steepest slopes for the relationship between reflex magnitude and background EMG activity were for FDS in the RPS task and FDI in RFA, and the correlation of reflex amplitude with background EMG was strongest for FDI in RFA. In contrast, the slope for FDS during RFA was the smallest, and the correlation of reflex amplitude with background EMG the weakest. In functional terms, FDI is primarily an abductor of the index finger, although it also acts synergistically in flexing the proximal phalanx. Thus, FDI plays a more important role in the RFA task than in the RPS task, whereas FDS serves as a prime mover in the RPS task but has only a minor role in the RFA task. Relationship between reflex amplitude and background EMG activity
Our findings on stretch reflexes during rhythmic finger tasks demonstrate that these reflexes are in some cases highly dependent on the state of the muscular contraction, as indicated by the relatively high correlation values associated with those tasks in which the investigated muscle was the prime mover. This is reminiscent of the phenomenon of ‘automatic gain compensation’, as defined for the stretch reflexes in the human thumb (Marsden et al. 1972, 1976; Matthews, 1986). The EMG response elicited by a perturbation increases progressively with the prevailing amount of voluntary contraction, and so remains an approximately constant proportion of the pre-existing level of activation. The above authors suggested that such an automatic gain compensation mechanism, reflecting the excitation level of the motoneurone pool, would ensure that reflexes remained appropriate for fine-motor tasks of the hand, and also compensated for fatigue. C The Physiological Society 2005
J Physiol 564.3
Phase-dependent reflex modulation in rhythmic hand tasks
In the present study, if automatic gain compensation was the only determinant of the reflex gain, one would expect reflex amplitude with the same background EMG (or contraction) level to be equal, irrespective of the form of the task. However, as noted above, this was clearly not the case for either FDI or FDS. The reflexes observed were modulated in quite different ways in the two tasks, in accordance with their recognized anatomical functions, suggesting that dynamic functional demand was a more potent determinant of reflex strength than any inherent automatic gain compensation mechanism. Recent studies on locomotory tasks have demonstrated that reflex gain in leg muscles is relatively independent of the background muscle activity in both H-reflexes and cutaneous reflexes (Brooke et al. 1997; Komiyama et al. 2000). In upper limb muscles, reflex responses to cutaneous stimulation during rhythmic, cyclical arm movements were found to be strongly correlated with the background activity (Zehr & Chua, 2000). However, when studying a larger sample of muscles and stimulating more cutaneous nerves, cutaneous reflexes were frequently dissociated from background muscle activity levels (Zehr & Kido, 2001). Possible neural mechanisms to account for either phaseor task-dependent modulation include presynaptic and/or postsynaptic inhibition (see reviews by Skorupski, 1991; Stein, 1995; Brooke et al. 1997; McCrea, 2001). Presynaptic inhibition of the group Ia muscle spindle afferents of the muscles actively involved would reduce the size of the reflex only, whereas postsynaptic inhibition would change both reflex size and voluntary EMG activity simultaneously. The latter effect is consistent with automatic gain compensation, while presynaptic inhibition could underlie, at least in part, phase and task-dependent reflex modulation. Presynaptic excitation, acting like presynaptic inhibition via Ia spindle afferent terminals on motoneurones, may also be involved in the control of the modulation pattern. The variable summation of each component mechanism could thus regulate the reflex modulation in order to obtain an optimal motor performance. Effect of motor task on short- and long-latency reflex components in upper limb muscles
In the present study, the short-latency component of the stretch reflex was the primary focus with respect to the modulation patterns of phase-dependency and task-dependency. It will be of interest to examine the modulation patterns of the long-latency reflex component of hand muscles during similar rhythmic, fine-motor actions. Task-dependent changes of short- and long-latency cutaneomuscular reflex components have been studied in the hand muscles, FDI and abductor digiti minimi C The Physiological Society 2005
949
during isometric finger abduction and sustained hand power grip (Evans et al. 1989, 1990). Both short-latency (E1) and longer-latency (E2) excitatory components of the cutaneous reflex response were recorded during finger abduction and power grip. However, the E1 component was generally significantly larger during power grips than finger abduction, while E2 was predominant during the isolated finger manoeuvres but decreased significantly during the power grips. Evidence indicates that the E1 component is mediated via a spinal pathway, whereas the E2 component is cortically mediated requiring the integrity of sensory-motor cortex, dorsal columns and cortico-spinal tract (Jenner & Stephens, 1982; Carr et al. 1993). Dietz et al. (1994) proposed a neural mechanism to explain task-dependent modulation of short- and long-latency EMG responses. They investigated the function of compensatory mechanisms in upper arm muscles under two motor task conditions: control of elbow position and control of elbow joint torque. Significant differences in the behaviour of the long-latency reflex component between position and torque control were observed, as reflected in the shape of the EMG responses. It was proposed that the task-dependent modulation of the long-latency component may be due to the muscle spindles responding in a more dynamic fashion during a position-controlling task than during a torque-controlling task. Such a change could be achieved by appropriate central regulation of γ -motoneurone activity. Alternatively, different proprioceptors may be activated during the two different tasks. Similar mechanisms could contribute to the modulation of stretch reflexes in hand muscles observed in the present study, particularly the longer-latency components, under isometric and isotonic conditions. Phase- and task-dependent modulation considered in a broader context
This discussion has highlighted similarities in the modulation characteristics of ‘stretch reflexes’ of two hand muscles described in this study with those of various proprioceptive and cutaneous reflexes in both upper and lower limbs reported in other studies. The functional significance of phase- and task-dependent modulation of the kind observed here has been argued in the context of simple, intrasegmental reflex feedback to the homonomous muscles in the hand (FDI and FDS), just as in previous explanations of the modulation of the soleus H-reflex during locomotion. Comparable cases based on functional need may be made for other types of reflexes, both intra- and intersegmental, including the cutaneous and other reflexes considered in this Discussion. Two interesting examples of interlimb reflexes linking the upper and lower limbs in man (see review by
950
R. Xia and others
Dietz, 2002) illustrate how both phase- and taskdependent reflexes may constitute elements of a more general phenomenon. First, rhythmic voluntary flexion–extension movements of one foot induce phase-related modulation of an ipsilateral wrist flexor H-reflex. In the second example, mechanical or electrical (tibial nerve) stimulation of one leg evokes cyclically modulated bilateral EMG responses in several upper arm muscles, but only when applied during the stance phase of walking, and not at all during standing with voluntary arm swinging, or sitting while writing (even with similar background EMG activity in the three conditions). The existence of these interlimb reflexes, and their inherent phase-dependency, may be manifestations of the phylogenetically older, ‘quadrupedal coordination’ still evident in primate locomotory neural circuitry. In contrast, the lack of a lower-to-upper limb reflex response during standing or sitting may reflect a presumed inhibitory control via the phylogenetically more recent, direct cortico-motoneuronal pathway in humans. In this context, it is perhaps noteworthy that the stretch reflexes described in this paper, evoked during voluntary, rhythmic actions of the hand mediated primarily by this direct cortico-motoneuronal pathway, nevertheless still do exhibit phase- and task-dependent modulation of a similar nature to that seen in more basic motor functions such as locomotion. This observation indicates that even finely controlled, complex manipulations (e.g. handwriting) still appear to rely for their efficient performance upon the integrity of the ‘primitive’, involuntary, reflex circuits which function largely outside our conscious perception. In conclusion, our results indicate that the phasedependent modulation of stretch reflexes examined here in two muscles of the hand, FDI and FDS, probably depends partly on the excitation level of the motoneurone pool, as defined by the ‘automatic gain compensation’ hypothesis. However, the extent and cyclical pattern of modulation, and the degree of correlation with fluctuating levels of excitation of the α-motoneurones, differ markedly between the muscles studied and the motor tasks performed. It is concluded therefore that the stretch reflexes reported in this study, during rhythmic finger tasks akin to handwriting, are modulated by central–peripheral interactions involving several neural mechanisms, and reflecting specific motor strategies adapted to the optimal performance of skilled, fine-motor tasks. References Akazawa K, Aldridge JW, Steeves JD & Stein RB (1982). Modulation of stretch reflexes during locomotion in the mesencephalic cat. J Physiol 329, 553–567. Belanger M & Patla AE (1987). Phase-dependent compensatory responses to perturbation applied during walking in humans. J Mot Behav 19, 434–453.
J Physiol 564.3
Brooke JD, Cheng J, Collins DF, McIlroy WE, Misiaszek JE & Staines WR (1997). Sensori-sensory afferent conditioning with leg movement: gain control in spinal reflex and ascending paths. Prog Neurobiol 51, 393–421. Brooke JD, Cheng J, Misiaszek JE & Lafferty K (1995). Amplitude modulation of the soleus H reflex in the human during active and passive stepping movements. J Neurophysiol 73, 102–111. Brooke JD, Peritore G, Staines WR, McIlroy WE & Nelson A (2000). Upper limb H reflexes and somatosensory evoked potentials modulated by movement. J Electromyogr Kinesiol 10, 211–215. Caccia MR, McComas AJ, Upton ARM & Blogg T (1973). Cutaneous reflexes in small muscles of the hand. J Neurol Neurosurg Psychiatr 36, 960–977. Capaday C & Stein RB (1986). Amplitude modulation of the soleus H reflex in the human during walking and standing. J Neurosci 6, 1308–1313. Capaday C & Stein RB (1987). Difference in the amplitude of the human soleus H reflex during walking and running. J Physiol 392, 513–522. Carr LJ, Harrison LM, Evans AL & Stephens JA (1993). Patterns of central motor reorganization in hemiplegic cerebral palsy. Brain 116, 1223–1247. Carter RR, Crago PE & Gorman PH (1993). Nonlinear stretch reflex interaction during cocontraction. J Neurophysiol 69, 943–952. Chen R & Ashby P (1993). Reflex responses in upper limb muscles to cutaneous stimuli. Can J Neurol Sci 20, 271–278. Dietz V (2002). Do human bipeds use quadrupedal coordination? Trends Neurosci 25, 462–467. Dietz V, Discher M & Trippel M (1994). Task-dependent modulation of short- and long-latency electromyographic responses in upper limb muscles. Electroencephalogr Clin Neurophysiol 93, 49–56. Dietz V, Faist M & Pierrot-Deseilligny E (1990). Amplitude modulation of the quadriceps H-reflex in the human during the early stance phase of gait. Exp Brain Res 79, 221–224. Doemges F & Rack PMH (1992a). Changes in the stretch reflex of the human first dorsal interosseous muscle during different tasks. J Physiol 447, 563–573. Doemges F & Rack PMH (1992b). Task-dependent changes in the response of human wrist joints to mechanical disturbance. J Physiol 447, 575–585. Dufresne JR, Soechting JF & Terzuolo CA (1980). Modulation of the myotatic reflex gain in man during intentional movements. Brain Res 193, 67–84. Duysens J & Loeb GE (1980). Modulation of ipsi- and contralateral reflex responses in unrestrained walking cats. J Neurophysiol 44, 1024–1037. Duysens J, Tax AA, Murrer L & Dietz V (1996). Backward and forward walking use different patterns of phase-dependent modulation of cutaneous reflexes in humans. J Neurophysiol 96, 301–310. Edamura M, Yang JF & Stein RB (1991). Factors that determine the magnitude and time course of human H-reflexes in locomotion. J Neurosci 11, 420–427. Evans AL, Harrison LM & Stephens JA (1989). Task-dependent changes in cutaneous reflexes recorded from various muscles controlling finger movement in man. J Physiol 418, 1–12. C The Physiological Society 2005
J Physiol 564.3
Phase-dependent reflex modulation in rhythmic hand tasks
Evans AL, Harrison LM & Stephens JA (1990). Maturation of the cutaneomuscular reflex recorded from the first dorsal interosseous muscle in man. J Physiol 428, 425–440. Ferris DP, Aagaard P, Simonsen EB, Farley CT & Dyhre-Poulsen P (2001). Soleus H-reflex gain in humans walking and running under simulated reduced gravity. J Physiol 530, 167–180. Forssberg H (1979). Stumbling corrective reaction: a phase-dependent compensatory reaction during locomotion. J Neurophysiol 42, 936–953. Forssberg H, Grillner S & Rossignol S (1975). Phase dependent reflex reversal during walking in chronic spinal cats. Brain Res 85, 103–107. Forssberg H, Grillner S & Rossignol S (1977). Phasic gain control of reflexes from the dorsum of the paw during spinal locomotion. Brain Res 132, 121–139. Hammond PH (1954). Involuntary activity in biceps following the sudden application of velocity to the abducted forearm. J Physiol 127, 23–25. Jenner JR & Stephens JA (1982). Cutaneous reflex responses and their central nervous pathways studied in man. J Physiol 333, 405–419. Johnson MTV, Kipnis AN, Lee MC, Loewenson RB & Ebner TJ (1991). Modulation of the stretch reflex during volitional sinusoidal tracking in Parkinson’s disease. Brain 114, 443–460. Komiyama T, Zehr EP & Stein RB (2000). Absence of nerve specificity in human cutaneous reflexes during standing. Exp Brain Res 133, 267–272. McCrea DA (2001). Spinal circuitry of sensorimotor control of locomotion. J Physiol 533, 41–50. Marsden CD, Merton PA & Morton HB (1972). Servo action in human voluntary movement. Nature 238, 140–143. Marsden CD, Merton PA & Morton HB (1976). Servo action in the human thumb. J Physiol 257, 1–44. Matthews PBC (1986). Observations on the automatic compensation of reflex gain on varying the pre-existing level of motor discharge in man. J Physiol 374, 73–90. Matthews PBC (1989). Long-latency stretch reflexes of two intrinsic muscles of the human hand analysed by cooling the arm. J Physiol 419, 519–538. Matthews PBC (1993). Interaction between short- and longlatency components of the human stretch reflex during sinusoidal stretching. J Physiol 462, 503–527. Nakazawa K, Yamamoto SI & Yano H (1997). Short- and longlatency reflex responses during different motor tasks in elbow flexor muscles. Exp Brain Res 116, 20–28. Noth J, Schwarz M, Podoll K & Motamedi F (1991). Evidence that low-threshold muscle afferents evoke long-latency stretch reflexes in human hand muscles. J Neurophysiol 65, 1089–1097. Schillings AM, Van Wezel BMH, Mulder T & Duysens J (1999). Widespread short-latency stretch reflexes and their modulation during stumbling over obstacles. Brain Res 816, 480–486.
C The Physiological Society 2005
951
Simonsen EB & Dyhre-Poulsen P (1999). Amplitude of the human soleus H reflex during walking and running. J Physiol 515, 929–939. Skorupski P (1991). Central modulation of crayfish walking leg reflexes. In Locomotor Neural Mechanisms in Arthropods and Vertebrates, ed. Armstrong DM and Bush BMH, pp. 160–172. Manchester University Press, Manchester. Skorupski P, Rawat BM & Bush BMH (1992). Heterogeneity and central modulation of feedback reflexes in crayfish motor pool. J Neurophysiol 67, 648–663. Skorupski P & Sillar KT (1986). Phase-dependent reversal of reflexes mediated by the thoraco-coxal muscle receptor organ in the crayfish, Pacifastacus leniusculus. J Neurophysiol 55, 689–695. Stein RB (1995). Presynaptic inhibition in humans. Prog Neurobiol 47, 533–544. Stein RB & Capaday C (1988). The modulation of human reflexes during functional motor tasks. Trends Neurosci 11, 328–332. Stuart DG (2002). Reflections on spinal reflexes. Adv Exp Med Biol 508, 249–257. Wallace CJ & Miles TS (1998). Movements modulate the reflex responses of human flexor pollicis longus to stretch. Exp Brain Res 118, 105–110. Wynn-Parry C (1966). Rehabilitation of the Hand, 2nd edn. Butterworth, London. Xia R & Bush BMH (1996). Phase-dependent modulation of reflex gain during repetitive finger actions in man. J Physiol 497.P, 108P. Yang JF & Stein RB (1990). Phase-dependent reflex reversal in human leg muscles during walking. J Neurophysiol 63, 1109–1117. Zehr EP & Chua R (2000). Modulation of human cutaneous reflexes during rhythmic cyclical arm movement. Exp Brain Res 135, 241–250. Zehr EP, Collins DF, Frigon A & Hoogenboom N (2003). Neural control of rhythmic human arm movement: phase dependence and task modulation of Hoffmann reflexes in forearm muscles. J Neurophysiol 89, 12–21. Zehr EP & Kido A (2001). Neural control of rhythmic, cyclical human arm movement: task dependency, nerve specificity and phase modulation of cutaneous reflexes. J Physiol 537, 1033–1045. Zehr EP & Stein RB (1999). What functions do reflexes serve during human locomotion? Prog Neurobiol 58, 185–205. Acknowledgements The awards to R. Xia of an Overseas Research Scholarship by the Committee of Vice-Chancellors and Principals of the Universities of the UK, and of a University of Bristol Postgraduate Fellowship, are gratefully acknowledged. We would like to thank Tony Macdonald in the Physiology Department, University of Bristol, UK, for expert technical assistance in designing and building some of the electronic equipment.
