Mechanical cutaneous stimulation alters Ia presynaptic inhibition in human wrist extensor muscles: a single motor unit study

Abstract

1. Reflex responses were evoked by radial nerve stimulation in 25 single motor units in the extensor carpi radialis muscles of seven subjects during voluntary isometric wrist extension. The responses consisted of narrow peaks in the post-stimulus time histograms with latencies compatible with monosynaptic activation.

2. When the skin of the palm and finger tips was continuously swept using a soft rotating brush, the purely monosynaptic components of the motor unit responses, as assessed from the contents of the first two 0.25 ms bins of the peak, were found to increase. This increase did not affect the motoneurone net excitatory drive, as assessed by measuring the mean duration of the inter-spike intervals. The cutaneous inputs activated by the brush may have reduced the tonic presynaptic inhibition exerted on the Ia afferents homonymous to the extensor motor units tested.

3. To further investigate whether Ia presynaptic inhibition was involved, the responses of the extensor motor units were conditioned by stimulating the median nerve 20 ms earlier, using a protocol which is known to induce Ia extensor presynaptic inhibition originating from flexor Ia afferents. The median nerve stimulation did not affect the motoneurone excitatory drive, but led to a decrease in the responses of the extensor motor units to the radial nerve stimulation, especially in the purely monosynaptic components. This decrease was consistent with the Ia presynaptic inhibition known to occur under these stimulation conditions.

4. The cutaneous inputs activated by the brush were found to reduce the Ia presynaptic inhibition generated by the median nerve stimulation, without affecting the distribution of the Ia presynaptic inhibition among the various types of motor units tested.

5. The present data suggest that cutaneous inputs from the palm and finger tips may relieve the Ia presynaptic inhibition exerted on the wrist extensor motor nuclei, and thus enhance the proprioceptive assistance to fit the specific requirements of the ongoing motor task.

Presynaptic inhibition is itself inhibited by input from cutaneous afferents, as clearly demonstrated in animal experiments (see Rudomín, 1990). In humans, most previous studies dealing with the effects of cutaneous afferents on the magnitude of Ia presynaptic inhibition have involved applying electrical stimulation to the superficial cutaneous nerves. Inhibitory effects of cutaneous afferents on Ia presynaptic inhibition have been observed in both leg (Iles & Roberts, 1987; Iles, 1996) and forearm muscles (Berardelli et al. 1987; Nakashima et al. 1990), based on changes in the amplitude of the H reflex.

Iles (1996) has tested the effects of applying mechanical stimulation to the skin of the plantar sole on the H reflex evoked in the soleus muscle in two subjects. Mechanical stimulation of the low threshold cutaneous receptors was found to decrease the Ia presynaptic inhibition, as assessed by measuring the changes in the H reflex, as efficiently as electrical stimulation.

The aim of the present study was to investigate the possibility that the numerous cutaneous inputs arising on the palmar side of the hand during manipulatory movements may selectively modulate the proprioceptive assistance to the wrist myotatic unit by altering the Ia presynaptic inhibition. In order to activate low threshold cutaneous receptors, the skin of the palm and the fingertips was continuously swept with a soft rotating brush, using a procedure based on microneurographic recordings of human sensory afferents (Ribot-Ciscar et al. 1989). The effects of the cutaneous inputs on the presynaptic inhibition elicited by stimulating the median nerve were assessed from the changes in the reflex responses of single motor units evoked by stimulating the radial nerve. The motor units were identified on the basis of their biomechanical and electrical properties (see Aimonetti et al. 2000)

The main result to emerge from the present study was that the cutaneous inputs activated by the brush were liable to reduce the Ia presynaptic inhibition exerted on the responses of the wrist extensor motor units to the radial nerve stimulation. Without affecting the motoneurone excitatory drive, the cutaneous inputs were consistently found to have inhibitory effects on the Ia presynaptic inhibition in a very similar way in all the types of motor units tested. Part of this study has been previously published in abstract form (Aimonetti et al. 1997).

METHODS

Experiments were performed on seven healthy male right-handed subjects aged 20–30 years, with the approval of the Ethics Committee of the local Medical University (CCPPRB-Marseille I, approval N8 92/74). All the subjects gave their informed written consent to the experimental procedure as required by the Declaration of Helsinki (1964). The basic methods were described in detail in the preceding paper (Aimonetti et al. 2000) and only additional methods are described below.

Stimulation protocol

In order to activate cutaneous mechanoreceptors, the skin of the palm and fingertips of the hand was lightly swept throughout the nerve stimulation sessions, using a soft rotating brush (40 rotations per minute, Fig. 2F), while the subjects were performing wrist extension with their fingers relaxed. The brushing was focused in particular on the fingertip pads, where the cutaneous mechanoreceptors show the highest pattern of density (Johansson & Vållbo, 1979).

Figure 2. Typical motor unit response.

The subjects were asked to performed wrist extension, keeping their fingers relaxed (A). In order to activate the cutaneous mechanoreceptors, the skin of the palm and fingertips was continuously lightly swept using a soft rotating brush (F). The similarity of the macro-potentials (B and G) and of the twitches (C and H) obtained in both cases confirms that the same motor unit was tested during wrist extension both with and without cutaneous stimulation. In both the recordings in which the radial nerve was being stimulated alone (D) and associated with median nerve stimulation (E,) the response probability of the motor unit was smaller than that measured during wrist extension associated with cutaneous stimulation (I and J).

During the brushing, 100 electrical stimuli were applied to the radial nerve. Motor units were recorded: first, while stimulating the radial nerve alone; second, during paired stimulation comprising a pulse applied to the median nerve followed 20 ms later by a pulse applied to the radial nerve; third, while stimulating the radial nerve alone during brushing; and fourth, during paired median/radial nerve stimulation during brushing. In some experiments, this order was reversed and no differences were observed. The data obtained with the fixed and the reversed order of presentation were therefore combined in the figures.

Data analysis

The firing patterns of the motor units were characterized in terms of the mean duration of the inter-spike intervals. The motor unit responses comprised narrow peaks in the post-stimulus time histograms (PSTHs). These responses were assessed either in terms of the probability of firing over the whole of the peak or over the first 0.5 ms part of the peak, the latter corresponding to the purely monosynaptic components (see Hultborn et al. 1987). With paired median/radial nerve stimuli, the responses were smaller than those with radial nerve stimulation alone. We attributed the decreases to presynaptic inhibition and expressed them as percentage changes (change in probability with paired stimuli divided by the probability with radial nerve stimulation alone × 100). Percentage changes were compared in trials with and without concomitant brushing of the skin (brushing reduced the changes, i.e. attenuated presynaptic inhibition).

Each motor unit tested was identified by its force threshold, macro-MUP area and twitch contraction time.

Statistical analyses

Differences in motor unit reflex responses between trials were tested by making paired comparisons using an ANOVA procedure for repeated measures. The significance level was set at P < 0.05. The relationships between the response probability (or the changes in the response probability) and the motor unit's functional parameters were analysed by linear regression. The characteristics of the regression lines corresponding to the data obtained during the various stimulations (variance homogeneity, Y-intercepts, and slopes) were compared by means of Snedecor's F test (Snedecor & Cochran, 1989). Changes in probability of motor unit response to radial nerve stimulation induced by prior stimulation of the median nerve were compared with and without concomitant cutaneous stimulation using a non-parametric procedure for paired measures (Wilcoxon signed-rank test).

RESULTS

A total number of 25 motor units was tested in the extensor carpi radialis muscles of seven subjects during wrist extension with and without cutaneous stimulation. The motor units were characterized in terms of the mean duration of the inter-spike intervals, the level of the force thresholds, the area of the macro-potentials, the rise time of the twitches, and the amplitude, duration and latency of the PSTH peaks occurring in response to the radial nerve electrical stimulation. Table 1 gives the statistics (means ±s.d., with range) on each of these functional parameters. The data given here are the pooled data obtained from all seven subjects.

Table 1.

Motor unit characteristics

	Response probability (imp/trigger)	Monosynaptic component (imp/trigger)	Peak latency (ms)	Peak duration (ms)	Inter-spike interval (ms)	Force threshold (N)	Macro-MUP area (mV ms)	Twitch contraction time (ms)
Radial nerve stimulation
Without cutaneous stimulation	0.62 ± 0.11	0.19 ± 0.07	21.7 ± 1.9	1.8 ± 0.4	99.7 ± 11.9	1.58 ± 1.38	0.35 ± 0.24	51.0 ± 21.7
	0.36–0.85	0.10–0.57	18.25–27.00	1.25–2.75	82.0–126.9	0.05–4.29	0.10–0.86	19.5–89.3
With cutaneous stimulation	0.71 ± 0.08	0.23 ± 0.07	22.8 ± 2.0	1.7 ± 0.4	97.4 ± 11.7	1.57 ± 1.37	0.34 ± 0.21	52.6 ± 22.1
	0.57–0.84	0.13–0.65	18.50–27.00	1.25–2.75	82.1–127.0	0.04–4.26	0.08–0.84	19.9–90.3
Radial and median nerve stimulation
Without cutaneous stimulation	0.32 ± 0.10	0.09 ± 0.05	21.7 ± 2.0	1.7 ± 0.4	100.0 ± 12.1	1.64 ± 1.35	0.40 ± 0.19	54.0 ± 20.6
	0.11–0.51	0.03–0.25	18.25–27.00	1.00–2.75	79.2–131.4	0.05–4.29	0.09–0.85	19.6–89.3
With cutaneous stimulation	0.41 ± 0.10	0.16 ± 0.07	22.8 ± 2.0	1.6 ± 0.3	99.5 ± 12.3	1.62 ± 1.36	0.42 ± 0.21	52.0 ± 21.6
With cutaneous stimulation	0.29–0.63	0.07–0.45	18.25–27.00	1.00–2.25	82.7–129.1	0.05–4.31	0.12–0.87	19.6–88.6

Motoneurone excitatory drive

The subjects were required to adjust their contraction force and consequently the EMG activity of the wrist extensor muscles to the minimum level required to keep the motor units tested firing continuously. The overall drive of the motoneurone pool is reflected in the background integrated EMG activity. As shown in Fig. 1a, no consistent differences in the mean integrated EMG activity were observed in the various experimental conditions. The mean values of the integrated EMG activity recorded when the radial nerve was being stimulated alone without (0.47 ± 0.26 mV) and with (0.49 ± 0.27 mV) cutaneous stimulation were not found to differ from those recorded when the median nerve was being stimulated without (0.48 ± 0.24 mV) and with (0.48 ± 0.27 mV) cutaneous stimulation.

Figure 1. Motoneurone excitatory drive.

A, no consistent differences in the mean integrated EMG activity associated with each motor unit tested with and without median nerve stimulation were observed between the recordings with or without cutaneous stimulation. B, likewise, no consistent differences in the mean duration of the inter-spike intervals observed with each motor unit tested were detected during the various stimulations.

The motoneurone net excitatory drive is reflected in the mean duration of the inter-spike intervals. As shown in Fig. 1b, no consistent differences in the mean duration of the inter-spike intervals recorded with each motor unit tested were observed with and without median nerve stimulation, whether or not cutaneous stimulation was being applied (see Table 1).

Motor unit response characteristics

All the motor units tested produced extraneous firing in response to the radial nerve stimulation, yielding very narrow peaks in the post-stimulus time histograms with latencies compatible with monosynaptic activation. The latencies and duration of the peaks were not found to differ significantly during the test and conditioning stimulation without and with cutaneous stimulation (see Table 1).

Figure 2 illustrates the response pattern of a single motor unit to radial nerve stimulation recorded during wrist extension alone (Fig. 2a) and while cutaneous stimulation was being applied (Fig. 2F). In the control trials, the response probability of the motor unit tested during wrist extension alone (Fig. 2D, 0.57 impulses per trigger) was smaller than that measured during wrist extension performed while cutaneous stimulation was being applied (Fig. 2I, 0.69 impulses per trigger). The similarity of the macro-potentials (Fig. 2b and G) and the twitches (Fig. 2C and H) obtained in both cases confirms that the same motor unit was tested during wrist extension with and without cutaneous stimulation. The median nerve stimulation decreased the response probability of the motor unit tested both during wrist extension alone (Fig. 2E, 0.37 impulses per trigger) and when cutaneous stimulation was being applied (Fig. 2J, 0.53 impulses per trigger). The decrease was greater when the motor unit was tested during wrist extension alone (-35.1 %) than when cutaneous stimulation was being applied (-23.2 %).

These findings were obtained in all the 25 motor units tested. Figure 3a and B illustrates the effects of cutaneous stimulation of the hand on the presumed presynaptic inhibition analysed over the whole reflex peak (Fig. 3a) and over the interval corresponding to the monosynaptic components (Fig. 3b) by performing regression analysis. For each motor unit, the probability of response (abscissa, impulses per trigger) measured during wrist extension alone while the radial nerve was being stimulated was plotted against the change in response probability (ordinate, %) induced by prior stimulation of the median nerve (without concomitant cutaneous stimulation: ▪, dotted line; with concomitant cutaneous stimulation: ▵, continuous line). Consistent but not significant trends were observed between each motor unit response probability and its changes, whenever the whole reflex peak (without cutaneous stimulation: r =−0.32, P = 0.12; with cutaneous stimulation: r =−0.51, P = 0.08) or only the purely monosynaptic components (without cutaneous stimulation: r =−0.29, P = 0.12; with cutaneous stimulation: r =−0.56, P = 0.06) was taken into account. The higher the motor unit response probability measured while the radial nerve was being stimulated alone, the stronger its response probability was found to be depressed while applying stimulation to the median nerve. In the recordings with concomitant cutaneous stimulation, similar trends were observed between each motor unit response probability and its changes. The higher Y-intercept suggests that the presynaptic inhibition presumed to be evoked by the median nerve stimulation may be relieved by the tonic cutaneous inputs.

Figure 3. The effects of cutaneous inputs on the Ia presynaptic inhibition induced by median nerve stimulation.

The higher the motor unit response probability measured while the radial nerve was being stimulated alone, the stronger its response probability assessed from the whole peak (A) or from the first 0.5 ms (B) was found to be depressed while applying stimulation to the median nerve, both during wrist extension without (▪, dotted line) and with (▵, continuous line) cutaneous stimulation. The higher Y-intercept suggests that the presynaptic inhibition presumed to be evoked by the median nerve stimulation may be relieved by the tonic cutaneous inputs. C and D, mean changes in the response probability in trials during which the median nerve was stimulated and cutaneous stimulation was applied () or not (□). Applying cutaneous stimulation was found to reduce the presumed presynaptic inhibition estimated both from the whole peak (C) and from the monosynaptic components (D).

The effects of the brushing on presumed presynaptic inhibition were tested by comparing for each motor unit its changes in response probability to radial nerve stimulation induced by prior stimuli to the median nerve without (□) and with concomitant brushing (). Cutaneous stimulation reduced the presumed presynaptic inhibition estimated from the whole peak by 24.49 % (P < 0.001; Fig. 3C) and the presumed presynaptic inhibition estimated from the monosynaptic components by 35.18 % (P < 0.001; Fig. 3D).

Dependence of the responses on the functional parameters of the motor units

The changes caused by prior median nerve stimuli in the monosynaptic components of the motor unit responses to radial nerve stimuli were significantly correlated with motor unit functional parameters as shown by the regression analysis of Fig. 4. Figure 4a shows that the higher the wrist extension force threshold, the smaller the change in the monosynaptic responses, i.e. the weaker the presumed Ia presynaptic inhibition, whether the motor units were tested during extension alone (r = 0.68, P = 0.004) or during brushing (r = 0.80, P = 0.003). Likewise, the larger the macro-potential area (Fig. 4b), the smaller the presumed Ia presynaptic inhibition during wrist extension alone (r = 0.58, P = 0.01) or during brushing (r = 0.70, P = 0.005); whereas the longer the twitch contraction time (Fig. 4C), the larger the presumed Ia presynaptic inhibition during wrist extension alone (r =−0.61, P = 0.006) or during brushing (r =−0.69, P = 0.001). These data confirm the existence of a downward gradient of Ia presynaptic inhibition, working from the slow to the fast contracting motor units, as previously suggested (Aimonetti et al. 2000).

Figure 4. Ia presynaptic inhibition depended on the functional parameters of the motor units.

The strength of the Ia presynaptic inhibition, as assessed from the changes in the purely monosynaptic components of each motor unit response, was significantly correlated with the functional parameters of the unit. The Ia presynaptic inhibition decreased when the recruitment threshold (A) and the macro-potential area (B) increased, and when the twitch contraction time (C) decreased. The similarity of the variances and the slopes suggests that the cutaneous stimulation may not have affected the distribution of Ia presynaptic inhibition among the various types of motor units tested. Note the marked differences between the Y-intercepts of the regression lines computed with and without cutaneous stimulation. ▪, wrist extension; ▵, wrist extension with cutaneous stimulation.

In each analysis, the regression lines computed during wrist extension with and without cutaneous stimulation yielded similar variances and slopes, which suggests that the cutaneous stimulation did not affect the distribution of the Ia presynaptic inhibition among the various types of motor units. The marked difference (P < 0.001) observed between the Y-intercepts (degrees of freedom = 1–46; force threshold (F) = 52.92; macro-potential area, F = 48.28; twitch contraction time, F = 54.41) clearly shows that in all the types of motor unit tested, the Ia presynaptic inhibition was weaker during wrist extension performed while cutaneous stimulation was being applied than during wrist extension alone.

The effects of the size of the unconditioned response on the strength of Ia presynaptic inhibition

In the recordings in which the radial nerve was stimulated alone, applying cutaneous stimulation increased the response probability of all the types of motor units tested. It was thus necessary to determine if the sensitivity of the motor units tested with cutaneous stimulation to presynaptic inhibition may depend on the size of the unconditioned responses (see Crone et al. 1990) by changing the intensity of the radial nerve stimulation, as described in the preceding paired paper (Aimonetti et al. 2000). This was performed on six low threshold motor units. While keeping the intensity of the median nerve constant at 0.8 × motor threshold (MT), the motor units were first tested using radial nerve stimulation (stimulation 1). Second, the same motor units were tested while decreasing the intensity of the radial nerve stimulation (stimulation 2) in order to obtain comparable peak sizes with and without cutaneous stimulation.

Figure 5 illustrates the responses of a low threshold motor unit tested under these conditions. The response probability of this motor unit to radial nerve stimuli without concomitant brushing was 0.58 impulses per trigger (test stimulation 1 at 1.1 MT). Prior median nerve stimulation induced a decrease of 58 % of its response probability. Figure 5a shows that the response probability of the same motor unit tested during brushing was 0.75 impulses per trigger with radial nerve stimulation alone (test stimulation 1). Additional prior median nerve stimulation decreased its response probability by 41.3 % (Fig. 5b). When the radial nerve stimulation intensity was set at 0.9 MT (test stimulation 2), the response probability of this motor unit to radial stimuli during brushing was 0.55 impulses per trigger (Fig. 5C). With prior median nerve stimulation this decreased by 38.9 % (Fig. 5D). In the case of the monosynaptic components, the median nerve stimulation decreased the response probability of the motor unit by 41.3 % while using test stimulation 1 and its response probability by 38.9 % while using test stimulation 2. The intensity of the radial nerve stimulation, i.e. the size of the unconditioned motor unit response, did not affect its sensitivity to Ia presynaptic inhibition. Similar data were obtained in the five other motor units tested under these conditions, as summarized in Table 2. These observations support the idea that the brushing cutaneous input attenuated the presumed presynaptic inhibition induced by the median nerve, whatever the size of the unconditioned response.

Figure 5. The size of the unconditioned response with cutaneous stimulation did not alter the strength of the Ia presynaptic inhibition.

The response probability of a low threshold motor unit tested without cutaneous stimulation was 0.58 impulses per trigger when the radial nerve stimulation was being applied alone (test stimulation 1 at 1.1 MT) and the median nerve stimulation induced a decrease of its response probability of 58 %. The response probability of the same motor unit tested with cutaneous stimulation was 0.75 impulses per trigger when the radial nerve stimulation was being applied alone (A, test stimulation 1). Applying median nerve stimulation decreased its response probability by 41.3 % (B). When the radial nerve stimulation intensity was set at 0.9 MT (test stimulation 2, C), applying median nerve stimulation decreased its response probability by 38.9 % (D). These observations suggest that the cutaneous inputs may relieve the presynaptic inhibition induced by the median nerve stimulation, whatever the intensity of the radial nerve stimulation, i.e. the size of the unconditioned motor unit response.

Table 2.

The effects of changing the intensity of the radial nerve stimulation

	Test stimulation 1 (1.1 MT)	Test stimulation 2 (0.9 MT)
Radial nerve stimulation with cutaneous stimulation: response probability (impulses per trigger)	0.84 ± 0.12	0.49 ± 1.09
Radial and median nerve stimulation with cutaneous stimulation: response probability (impulses per trigger)	0.51 ± 0.17	0.31 ± 2.21
Changes in response probability (%)	−39.28 ± 9.18	−36.73 ± 5.8

Six low force treshold MUs: force threshold, 0.85 ± 0.33 N; macro-MUP area, 0.42 ± 0.12 mV ms; CT, 51.17 ± 14.33 ms.

DISCUSSION

The effects of mechanical cutaneous stimulation on Ia presynaptic inhibition were studied in identified single motor units tested during voluntary isometric contractions. The cutaneous inputs generated by sweeping the skin of the palm and fingertips decreased the Ia presynaptic inhibition exerted on the monosynaptic responses of the extensor motor units to radial nerve stimulation. The cutaneous inputs probably reduced both the background Ia presynaptic inhibition due to continuously firing Ia afferents and the extra presynaptic inhibition induced by stimulating the median nerve. The cutaneous inputs did not affect the distribution of the presumed Ia presynaptic inhibition to the various types of motor units tested.

Cutaneous inputs and the motoneurone excitatory drive

The effects of the cutaneous inputs activated by brushing on the reflex responses of the extensor motor units may have resulted from changes in post-synaptic events. Cutaneous inputs have been reported previously to exert opposite synaptic actions, depending on the type of motor units tested in anaesthetized cats (Burke et al. 1970; Kanda et al. 1977; but see Clark et al. 1993). Likewise in humans, electrical stimulation of superficial cutaneous nerves was found to alter the recruitment gain among the various types of motor units tested by defacilitating the slow contracting motor units while facilitating the fast contracting motor units tested in the first dorsal interosseous muscle (Stephens et al. 1978) and the tibialis anterior muscle (Nielsen & Kagamihara, 1993).

If any such changes in the post-synaptic events had occurred, they would have led to differences between the motoneuronal firing rates. In the present study, no differences were observed, however, in the mean duration of the inter-spike intervals between the recordings performed with and without cutaneous stimulation and with and without median nerve stimulation. The present data therefore do not provide any evidence that changes occurred in the post-synaptic events. It is worth noting, however, that in most previous studies on human motor unit activity and in the present one, the subjects were provided with auditory and visual feedback about the motor unit activity to help them keep the motor units firing steadily. The hypothesis cannot be ruled out that the cutaneous stimulation may have actually altered the post-synaptic events, but the subjects were able to compensate for the effects of the cutaneous stimulation by adjusting their level of voluntary contraction.

Cutaneous inputs and Ia presynaptic inhibition

In the recordings in which the radial nerve was being stimulated alone, the motor unit response probability during wrist extension with cutaneous stimulation was greater than during wrist extension alone. This effect of the cutaneous stimulation was consistently observed in the first two 0.25 ms bins of the peak, which can be taken to reflect its purely monosynaptic components (see Hultborn et al. 1987). The present data suggest that the cutaneous inputs activated by the brush may have decreased the tonic Ia presynaptic inhibition acting on the responses of the extensor motor units. These data are in keeping with the results of previous human experiments on the wrist flexor muscles during anaesthesia of the hand (Nakashima et al. 1990). There may exist a tonic cutaneous influence on the activity of the presynaptic inhibitory pathways acting on the wrist muscles’ Ia afferents.

To investigate the possible involvement of Ia presynaptic inhibition further, the responses of the extensor motor units were conditioned by stimulating the median nerve 20 ms earlier in order to induce Ia presynaptic inhibition (see Burke et al. 1994). Applying median nerve stimulation decreased the responses of the extensor motor units tested without cutaneous stimulation, as was to be expected. This decrease was consistently observed in the purely monosynaptic components of the motor units’ responses, which confirms that the median nerve stimulation did induce presynaptic changes. Concomitant mechanical cutaneous stimulation attenuated the decreases in the purely monosynaptic responses of the motor units. The experiments performed while the intensity of the radial nerve stimulation was adjusted in order to obtain comparable responses with and without cutaneous stimulation confirm that the strength of the Ia presynaptic inhibition did not depend on the size of the unconditioned responses. Cutaneous inputs may therefore relieve presynaptic inhibition from the median nerve, whatever the size of the unconditioned response.

The results obtained for paired median/radial nerve stimulation with concomitant brushing are in good agreement with previous findings on the wrist flexor muscles, which showed that electrical stimulation applied to the cutaneous branches of the radial nerve reduced the presynaptic inhibition by 20 % (Berardelli et al. 1987; Nakashima et al. 1990). It is worth noting that in the present study, the mechanical stimulation decreased the Ia presynaptic inhibition by 15 %. It has been previously reported that the Ia presynaptic inhibition was depressed by 7.7 % when mechanical stimulation was applied, as against 10.1 % in the case of electrical stimulation (see Fig. 4a and B in Iles, 1996). The slightly weaker efficiency of the mechanical stimulation as a means of reducing the Ia presynaptic inhibition may be due to the fact that this type of stimulation induced tonic and dispersed cutaneous afferent activity, whereas highly synchronous cutaneous afferent volleys are induced when electrical stimulation is applied (see Burke et al. 1984). These considerations aside, the evidence all supports the notion that the cutaneous inputs interact with other sensory afferents to modulate the efficiency of proprioceptive assistive reflexes, which regulate the voluntary command of the wrist myotatic unit (see Macefield et al. 1996; Macefield & Johansson, 1996).

Functional considerations

The present data confirm the idea that cutaneous inputs may tonically alter the activity of the wrist Ia presynaptic inhibitory pathways. In cats, this tonic cutaneous influence is known to be controlled by supraspinal afferents (Rudomín et al. 1998). The cutaneous modulation of Ia presynaptic inhibition observed here might interact with the other effects exerted by the cutaneous inputs at the segmental level (Burke et al. 1970; Kanda et al. 1977; Stephens et al. 1978; Nielsen & Kagamihara, 1993), at the propriospinal level (see Pierrot-Deseilligny, 1996), and at the supraspinal level via transcortical reflex pathways (see Christersen et al. 1999). These cutaneous effects may ensure that the gain of monosynaptic proprioceptive assistance is adjusted to fit the specific requirements of the ongoing motor task.