In humans, propriospinal neurons located at midcervical levels receive peripheral and corticospinal inputs and probably participate in the control of grip tasks, but their role in reaching movements, as observed in cats and primates, is still an open question. The effect of ulnar nerve stimulation on flexor carpi radialis (FCR) motor evoked potential (MEP) was tested during reaching tasks and tonic wrist flexion. Significant MEP facilitation was observed at the end of reach during reach-to-grasp but not during grasp, reach-to-point, or tonic contractions. MEP facilitation occurred at a longer interstimulus interval than expected for convergence of corticospinal and afferent volleys at motoneuron level and was not paralleled by a change in the H-reflex. These findings suggest convergence of the two volleys at propriospinal level. Ulnar-induced MEP facilitation was observed when conditioning stimuli were at 0.75 motor response threshold (MT), but not 1 MT. This favors an increased excitability of propriospinal neurons rather than depression of their feedback inhibition, as has been observed during tonic power grip tasks. It is suggested that the ulnar-induced facilitation of FCR MEP during reach may be due to descending activation of propriospinal neurons, assisting the early recruitment of large motoneurons for rapid movement. Because the feedback inhibitory control is still open, this excitation can be truncated by cutaneous inputs from the palmar side of the hand during grasp, thus assisting movement termination. It is concluded that the feedforward activation of propriospinal neurons and their feedback control may be involved in the internal model, motor planning, and online adjustments for reach-to-grasp movements in humans.
- propriospinal neurons
the functional role of the propriospinal system in human motor control is still debated. First explored in cats, this interneuronal system is located at midcervical levels (C3–C4) and mediates the command for target-reaching, while the act of food-taking involves interneurons in the same segments as motoneurons (Alstermark and Lundberg 1992). The development of the cerebral cortex and of direct corticomotoneural projections have been paralleled by enhanced manual dexterity, and this has led to the conclusion that the descending motor command for hand movements involves mainly the corticomotoneuronal projections in primates and humans (see Kirkwood et al. 2002). However, these evolutionary changes have been accompanied by greater feedback and feedforward inhibition of propriospinal neurons (Alstermark et al. 1999). Further studies in macaque monkeys have revealed the role of the propriospinal system in reach and grip movements and its possible implication in motor recovery (Alstermark et al. 2011; Isa et al. 2007; Perlmutter 2009). These studies indicate differences in synaptic connectivity at the propriospinal level across species, accompanied by modification of its functional role. The behavioral results of Isa, Alstermark, and colleagues (Isa et al. 2007; Alstermark et al. 2011) were based on lesion experiments, and these do not give any information about the control at the propriospinal level during normal movement. In the awake monkey, Perlmutter (2009) confirmed powerful activation of C3–C4 neurons during movement and their projections to upper limb motoneurons, but in light of the weakness of the functional linkage between propriospinal neurons and muscular activity he raised questions about the functional role of C3–C4 neurons during normal movement.
Although indirect, electrophysiological investigations in humans can provide insight into changes in neural excitability during movement. In previous studies, it was shown that propriospinal excitability is enhanced during a power grip but not during a pointing task. The evidence indicated that this was due to descending control of feedback inhibition to propriospinal neurons. The release of inhibition allowed afferent feedback to assist movement through the propriospinal system (Iglesias et al. 2007). Another study revealed enhanced propriospinal excitation at the end of a lift-grip movement (Roberts et al. 2008). These two studies suggest that the propriospinal system transmits afferent feedback to assist contraction of proximal muscles to control arm position during hand movements in humans. However, there are no data on whether the propriospinal system participates in the transmission of the motor command during movement, and particularly during reaching, the task in which it was first implicated in the cat (Alstermark and Lundberg 1992).
Reaching movements have been extensively investigated in humans, with a particular focus on the coordination of eyes, head, and arm during visually guided target-reaching. Brain imaging and transcranial magnetic stimulation (TMS) studies have revealed task-, condition-, and time-related changes in cortico-cortical interactions during reaching movements (Filimon 2010; Koch and Rothwell 2009). While segregated pathways have been revealed at the cortical level for reaching tasks (Koch et al. 2010), there are no data on the resulting descending motor command and its transmission to spinal motoneurons. Our hypothesis was that the cortico-subcortico-cortical interactions during the planning of reaching movements create different outputs, which in turn activate different spinal neural pathways, leading to changes in motor behaviors. We therefore tested whether changes at the propriospinal level occur during two different reaching tasks.
Percutaneous electrical stimulation of the ulnar nerve at wrist level (activating group I afferents from intrinsic muscles of the hand) was combined with TMS to study the modulation of motor evoked potential (MEP) in wrist flexors (Iglesias et al. 2007; Lourenço et al. 2006, 2007), during reaching, when the subjects had to grasp a target placed in front of them (reach-to-grasp) or when they had to touch it with the index finger (reach-to-point). Because enhanced propriospinal excitation has been described during power grip (Iglesias et al. 2007), stimulation was also delivered during the grasp of reach-to-grasp to test whether similar effects could be observed during a dynamic grasping task. Enhanced propriospinal excitation was found during the reach of reach-to-grasp but not during the actual grasp. We discuss the role of the propriospinal system in prehension tasks in humans and its role in the internal model for motor planning and online adjustments of the initial motor program.
The experiments were carried out on 13 healthy subjects (8 men, 5 women; 38.5 ± 4.3 yr), all of whom had given informed written consent to the experimental procedures, which had been approved by the ethics committee of the Pitié-Salpêtrière Hospital (CPP Ile de France VI). The study conformed to the standards set by the latest revision of the Declaration of Helsinki.
General Experimental Procedures
Subjects were comfortably seated in a chair with the right arm resting on a table, relaxed with no detectable EMG, with the palmar side of the forearm and the hand facing the table. The position of the subject was checked at the beginning of the experiment so that the target immediately in front could be touched without moving the trunk (Fig. 1, A and B). The subject was seated slightly higher than the target in order to facilitate wrist flexion, and to limit the activation of wrist extensors (Fig. 1C, Fig. 2). Whatever the reaching task, the arm had to be fully extended by the end of the movement.
Electrophysiological investigations were performed during three different tasks: 1) reaching movement to grasp a plastic egg (6-cm diameter) with the full hand (reach-to-grasp; Fig. 1A), 2) reaching movement to touch the top of the plastic egg with the index finger fully extended and the other fingers clenched in a fist (reach-to-point; Fig. 1B), and 3) tonic wrist flexion with closed fist (tonic). The reaching tasks were self-initiated. Two pressure-sensitive sensors were placed under the elbow and on the top of the egg to signal respectively the beginning of the movement (onset movement) and when the target was reached (touch) (Fig. 1, A and B, Fig. 2, C and D). The time difference between the signals from the two sensors therefore reflected the duration of the reaching phase. After training for 5–15 min, the subjects were able to perform each reaching task reproducibly (reach-to-grasp 413 ± 21 ms and reach-to-point 383 ± 19 ms, with a coefficient of variation of ∼10%). The signal generated by the pressure transducer under the elbow was used to trigger the stimulation during the reaching tasks.
EMG activity was recorded by surface electrodes (Delsys, Boston, MA; 20–450 Hz bandwidth; Ag electrodes DE-2.1) secured to the skin over the muscle belly of 1) flexor carpi radialis (FCR), in the upper part of the forearm, at a site where selective wrist flexion produced much more activity than selective flexion of the fingers, 2) deltoid (anterior head), 3) extensor carpi radialis (ECR), and 4) abductor digiti minimi (ADM). EMG signals were amplified (×10,000) and stored (2-kHz sampling rate) on a personal computer for off-line analysis (Notocord-hem 3.4; Notocord, Croissy s/Seine, France). The wrist, elbow, and shoulder positions were measured with goniometers (Biometrics, Cwmfelinfach, UK; 500 Hz) in six subjects. The goniometers were placed on the medial aspect of the dorsal surface of the hand and of the arm, with the two arms on either side of the relevant joint. For the shoulder, one transducer was on the shoulder and the other above the medial head of deltoid. In the resting position (before reaching tasks), the wrist was in slight flexion (∼10–20°), the elbow was semiflexed (∼25–35°), and the shoulder was in slight abduction (∼30°) and forward flexion (∼5–10°). The signals from goniometers were aligned to the onset of movement, and the x-axis was normalized to the duration of the reaching phase (before touch; 100% in Fig. 1, C–E).
Peripheral nerve stimulation.
Electrical stimulation (rectangular pulse, 1-ms duration) was applied percutaneously to the ulnar nerve at wrist level and to the median nerve at elbow level with bipolar electrodes (silver plates, 0.5 cm2 for the ulnar and 1.8 cm2 for the median; cathode proximal). The intensity of ulnar nerve stimulation was expressed in multiples of the threshold of the motor response evoked in ADM (× MT), and this was checked at each delay investigated during each motor task. Stimuli were adjusted to 0.75 and 1 × MT. On average, no significant difference was found in threshold intensity between tasks: 1) reach-to-grasp 6.1 ± 1.2 vs. reach-to-point 7.4 ± 1.5 vs. tonic 6.2 ± 0.6 mA (ANOVA, P = 0.67), 2) reaching 6.0 ± 0.7 vs. grasping 6.0 ± 0.7 mA (paired t-test, P = 0.96). Supramaximal stimulation of median nerve was delivered during each motor task, and at each delay investigated during movement, to measure the maximal motor response (Mmax) in FCR. The intensity of median nerve stimulation was then adjusted to produce a H-reflex in FCR of ∼10% Mmax.
TMS was applied over the primary motor cortex with a Magstim Rapid (Magstim, Whitland, UK) through a figure-of-eight coil (70 mm) held at the optimal position to evoke a MEP in FCR. The coil position was first determined during tonic wrist flexion, and it was ensured that during reaching movements the MEP was larger in wrist flexors than in extensors. The coil position was then marked on a cap worn by the subject. The coil was held by one of the investigators to ensure its stability during movement. The size and shape of the FCR MEP were monitored on an oscilloscope throughout the experiments, together with the EMG activity of deltoid, ECR, and ADM. TMS intensity was adjusted to evoke MEPs in FCR of ∼5–15% Mmax.
Protocol 1 (12 subjects).
EMG activity and signals from pressure transducers were collected during reach-to-grasp and reach-to-point (Fig. 2, A and B). They were then rectified and averaged (10 sweeps) and aligned to the onset of movement (Fig. 2, C and D) to determine the reaching pattern. The moment when the target was reached (touch) was also determined (Fig. 2, C and D). The stimuli were delivered at the beginning of FCR activity (Fig. 2, C and D), i.e., at the end of the reaching phase, just before touch (see also Fig. 1, C–E). On average, stimulation was delivered at later delays during reach-to-grasp (389 ± 28 ms, range 250–520 ms after onset of movement, depending on subject) than during reach-to-point (339 ± 30 ms, range 150–500 ms; ANOVA P < 0.01) because the latter had a shorter duration than the former: the subjects touched the egg with the tip of the index finger but grasped it with the full hand. However, whatever the reaching task, the stimulation was delivered during the phase of increasing EMG in FCR, at a delay corresponding to ∼85% of the total duration of the reaching phase (reach-to-grasp 87.3 ± 2.7 vs. reach-to-point 82.4 ± 4.3%; t-test, P = 0.12). The subjects were then asked to perform a tonic wrist flexion with closed fist at an EMG level within the same range as during reaching tasks; the intensity of the contraction was ∼40–60% of the maximal voluntary tonic contraction as estimated from the EMG. During movement, the ulnar nerve stimulation preceded TMS with an interstimulus interval (ISI) of between 6 and 12 ms (ISI was changed in 1-ms steps), depending on subject. Twenty test MEPs (TMS by itself) and 20 conditioned MEPs (ulnar + TMS) were randomly alternated at ∼0.5 Hz for each ISI and for the two intensities of ulnar nerve stimulation investigated (0.75 and 1 × MT).
Protocol 2 (6 subjects).
Ulnar nerve stimulation at 0.75 × MT was used to condition the MEP and the H-reflex of FCR during reach-to-grasp. The reaching pattern was determined as in protocol 1, and stimulation was delivered at the onset of FCR activity, i.e., at the end of the reaching phase (before touch; ∼85% of the reaching phase duration). The intensities of TMS and median nerve stimulation were adjusted to evoke a MEP and a H-reflex of similar size to ensure that the same motoneurons were studied within the pool (% Mmax; Crone et al. 1990; Lackmy and Marchand-Pauvert 2010; Morita et al. 1999). ISIs between 6 and 11 ms were investigated, dependent on subject, and 20 test (MEP or H-reflex alone) and 20 conditioned (+ ulnar nerve stimulation) responses were randomly alternated at 0.3 Hz.
Protocol 3 (13 subjects).
Protocol 1 was repeated at two delays during reach-to-grasp, corresponding to the end of the reaching phase (as in protocol 1) and the grasping phase (after touch). In the latter, the stimuli were triggered on average 548 ± 29 ms after the onset of movement, at a moment when the subjects were flexing the fingers around the egg. On average, this delay corresponded to 125 ± 4% of the total duration of the reaching phase.
MEP recruitment curves (6 subjects).
In the same experiment TMS was applied over the primary motor cortex to evoke a MEP in FCR during reach-to-grasp (at the end of the reaching phase, as in protocol 1) and then during tonic wrist flexion in three subjects and the other way round in the three remaining subjects. The intensity of TMS was varied (2–5% maximal stimulator output) between MEP threshold and MEP maximal size. MEP size was plotted against TMS intensity, and regression analysis was performed to determine the equation for the sigmoid curve best fitting the experimental measurements. MEP size was normalized to Mmax for interindividual comparisons.
The area of the MEP was measured by computer from the rectified EMG; the analysis window was determined by the MEP latency and its duration, both of which were measured from the raw EMG (Fig. 3A). H-reflexes were analyzed as the peak-to-peak amplitude. The difference between conditioned and test responses was expressed as a percentage of the mean test response, to quantify the level of facilitation (ulnar facilitation) produced by the convergence of peripheral (ulnar nerve) and corticospinal volleys (TMS). In each individual, the sizes of test and conditioned responses were compared with a paired t-test. The relationship between the background FCR EMG and the MEP size, and between the background ADM EMG and ulnar facilitation, was assessed with Pearson's correlation with repeated measures (Poon's treatment to take into account within- and between-subject variances; Poon 1988). Differences between motor tasks were tested with ANOVA (background EMG activity, test response size, TMS output, trigger delay, ulnar nerve intensity). ANCOVA was performed to compare the ulnar facilitation between motor tasks, using the background FCR or ADM EMG as a covariate. Two-way ANOVA was performed to test the influence of the conditioning intensity and motor task. If these tests provided significant P values, post hoc Fisher least significant difference (LSD) tests were performed for comparisons of two means. Paired t-tests were used to compare the background ADM EMG during reaching tasks, MEP threshold, TMS intensity evoking a MEP of half-maximal size (I50) and maximal MEP size in the recruitment curves, the effect of ulnar nerve stimulation on FCR MEP and H-reflex, and the ulnar facilitation during reaching and grasping phases. For all tests, the significance level was set at P < 0.05. Data are given as means ± standard error (SE).
Influence of Motor Task on Ulnar Facilitation of MEP of FCR
Figure 1, C–E, show the angle joint variations at wrist (Fig. 1C), elbow (Fig. 1D), and shoulder (Fig. 1E) levels during reach-to-gasp and reach-to-point in one subject. At elbow and shoulder, the movements were quite similar during the reaching tasks and stimuli were delivered at similar angles. During reach-to-grasp the subject progressively flexed the wrist to grasp the egg, whereas there was limited change at the wrist during reach-to-point. However, the stimuli were delivered at similar angles because of the weak wrist extension required to position the hand for grasping; see also the weak ECR EMG activity before FCR burst (Fig. 2C). Wrist flexion was greater after the touch as the subject grasped the plastic egg (>100% reaching phase). Similar changes were observed in the five other subjects so investigated.
Figure 2, A–D, show the EMG activity in ADM, FCR, ECR and deltoid for two subjects (Fig. 2, A and B, and Fig. 2, C and D), during reach-to-grasp (A and C) and reach-to-point (B and D). The relative intensity of EMG activity was similar during the two reaching movements except in ADM, which was weakly activated during reach-to-point. In both cases, FCR was activated at the end of the reaching phase, ∼150–200 ms before touching the target.
Figure 3A shows the raw and rectified MEP evoked in FCR EMG activity during reach-to-grasp, and Fig. 3B shows the MEPs on an expanded x-axis during the three motor tasks. As indicated in the dotted trace in Fig. 3B, left, the size of the mean rectified MEP in FCR was significantly enhanced (t-test, P < 0.05) by ulnar nerve stimulation (0.75 × MT; ISI 8 ms) at the end of the reaching phase (before touch) of reach-to-grasp but not during reach-to-point or during the tonic contraction (Fig. 3B, center and right). Given the similar latency of FCR MEP and H-reflex (17 ms) and the afferent conduction time between wrist (ulnar nerve) and elbow (median nerve) stimulation sites (4–5 ms; Marchand-Pauvert et al. 2000), convergence of the ulnar afferent volley and the corticospinal input at motoneuron level should have occurred at an ISI between 4 and 5 ms. However, ulnar facilitation of the MEP was observed 3–4 ms later, greatest at ISI 8 ms (the “optimal” ISI for this subject; Iglesias et al. 2007).
Figure 3C shows the time course of the effects produced by the combination of ulnar nerve stimulation and TMS on FCR EMG in another subject. The ulnar nerve stimulation did not change the size of the FCR MEP at ISIs of 7 and 9 ms. At 8 ms, the MEP was significantly larger after ulnar nerve stimulation (0.75 × MT; t-test, P < 0.05) during reach-to-grasp but not during the other motor tasks. Similar facilitation was observed at the end of the reaching phase of reach-to-grasp in all 12 subjects so tested, at a mean ISI of 8.8 ± 0.3 ms (range 7–11 ms). Figure 3D shows the group data (for 12 subjects) obtained at the optimal ISI for ulnar-induced MEP facilitation, at the end of the reaching phase of reach-to-grasp (28.4 ± 4.8% test MEP), of reach-to-point (1.2 ± 4.4%), and during tonic contractions (−2.2 ± 4.3%). No facilitation was observed 1 ms earlier than the optimal ISI in each task.
The background level of FCR EMG was similar between tasks (185 ± 30 vs. 217 ± 39 vs. 154 ± 32 mV during reach-to-grasp, reach-to-point, and tonic, respectively; ANOVA, P = 0.34). Care was taken to adjust the TMS intensity to evoke a MEP of the same size in each task, but the mean test MEP was still significantly smaller during tonic than during reaching movements (tonic 5.0 ± 0.9 vs. reach-to-grasp 7.6 ± 1.2 vs. reach-to-point 8.0 ± 1.6% Mmax; ANOVA, P < 0.05; Fisher's LSD, P < 0.05 for tonic vs. reaching tasks). On the other hand, TMS output was within the same range in each task (53.3 ± 2.9 vs. 55.6 ± 2.8 vs. 55.7% maximal stimulator output; ANOVA, P = 0.23). Correlation analysis revealed a significant relationship between the test MEP and the level of ulnar facilitation (Pearson's analysis using Poon's treatment, P < 0.001, r = 0.62). ANCOVA was thus performed to compare the level of ulnar facilitation between motor tasks, using the test MEP as covariate. The facilitation was significantly greater during reach-to-grasp (Fig. 3D), taking into account the difference in the sizes of the test MEP between tasks (ANCOVA, P < 0.001; Fisher's LSD, P < 0.001 when comparing reach-to-grasp to reach-to-point or tonic).
The mean level of EMG in ADM was significantly less during reach-to-point (76.5 ± 12.2 mV) than during reach-to-grasp (157.7 ± 37.1 mV; t-test, P < 0.05; Fig. 2, A–D). A significant linear correlation was found between the background EMG of ADM and the level of ulnar facilitation (Pearson's analysis with Poon's treatment, P < 0.001, r = 0.99). However, there was still a significant difference in ulnar facilitation between reaching movements when this difference in EMG level was taken into account (ANCOVA, P < 0.01).
Given the significant difference in test size of FCR MEP during reach-to-grasp and tonic contraction despite a similar TMS intensity, MEP recruitment curves were studied in six subjects. Figure 4 shows the recruitment curves obtained at the end of the reaching phase of reach-to-grasp (Fig. 4A) and during tonic wrist flexion (Fig. 4B) in one subject. In both cases, the regression analysis revealed a sigmoid curve with a highly significant coefficient of regression (R2 was ∼97%; P < 0.0001). The difference between the two curves lies in the size of the MEP reached as stimulus intensity increased, the MEP being larger for reach-to-grasp. However, the normalized stimulus-response curves were virtually identical. This suggests that the sequence of motoneuron recruitment was comparable in the two tasks. Similar findings were seen in the other five subjects. On average, the threshold for the MEP was lower, but not significantly so (48.2 ± 5.0 vs. 53.2 ± 5.6% maximal stimulator output; t-test, P = 0.51), and its maximal size greater (18.0 ± 5.1 vs. 8.3 ± 0.9% Mmax; P < 0.05) during reach-to-grasp than during tonic contractions. However, as implied by the similarity of the normalized stimulus-response curves, the mean I50 was identical (58.7 ± 6.6 vs. 58.5 ± 6.6% maximal stimulator output; P = 0.94). For protocol 1, the size of the test MEP in the subject illustrated in Fig. 4 was about the I50, i.e., 5.7% Mmax during reach-to-grasp and 2.9% during tonic contraction, and this was so in the other five subjects. Importantly, this indicates that the effects of ulnar nerve stimulation were studied on the approximately linear part of the sigmoid curve during both forms of contraction.
Absence of Ulnar Facilitation of FCR H-Reflex
Additional experiments were undertaken using the H-reflex of FCR as the test response to characterize further the origin of ulnar facilitation during reaching. Figure 5A shows the time course of the effect of ulnar nerve stimulation on the MEP and H-reflex of FCR at the end of the reaching phase of a reach-to-grasp movement in one subject. The MEP was significantly greater after ulnar nerve stimulation at the 7-ms ISI, but there was no change in the H-reflex. The test MEP was of similar size as the test H-reflex (6.1 ± 0.6 vs. 7.2 ± 0.8% Mmax; t-test, P = 0.55). The group data in Fig. 5B show that the ulnar-induced change in the MEP was significantly greater at the optimal ISI than that of the H-reflex (28.3 ± 2.8 vs. −2.1 ± 0.8% test response; t-test, P = 0.01).
Difference Between Reaching and Grasping During Reach-to-Grasp
Figure 6A shows the time course of the effects of ulnar nerve stimulation on the MEP of FCR during reach-to-grasp in one subject, when stimuli were delivered at the end of reach and during grasping. Significant ulnar facilitation (t-test, P < 0.05) was observed at ISIs of 8 and 9 ms at the end of the reaching phase but not during the grasping phase. Similar ulnar facilitation was observed in all 13 subjects at the end of reach, and in 5 of 13 subjects during grasping, at a mean ISI 8.8 ± 0.3 ms, and the facilitation was greater at the end of reach in 12 of the 13 subjects. Figure 6B shows the group data obtained at the optimal ISI for ulnar facilitation, at the end of reach and during grasping. There was significantly greater ulnar facilitation at the end of reach (P < 0.001; paired t-test).
The mean background level of FCR EMG, TMS output, and test MEP size were similar across tasks [at the end of reach and during grasping, respectively: EMG 294 ± 95 vs. 257 ± 74 mV (ANOVA, P = 0.38); TMS 52.1 ± 3.1 vs. 52.7 ± 3.0% maximal output (P = 0.9); MEP 14.0 ± 3.9 vs. 17.4 ± 3.7% Mmax (P = 0.34)].
Influence of Conditioning Intensity
To determine the origin of the enhanced ulnar facilitation during reach-to-grasp, and particularly at the end of the reach, the effect of ulnar nerve stimulation on FCR MEP was tested at 0.75 and 1 × MT. It has been shown previously that, with stronger stimuli, activation of inhibitory interneurons dominates transmission in the human propriospinal system (Nicolas et al. 2001).
Figure 7, A–D, show the FCR MEP conditioned by ulnar nerve stimulation at 0.75 and 1 × MT, at the end of reach of reach-to-grasp (Fig. 7A) and of reach-to-point (Fig. 7B), during tonic contraction (Fig. 7C), and during the grasping phase of reach-to-grasp (Fig. 7D) in one subject. The conditioned MEP was significantly larger at 0.75 × MT than at 1 MT (P < 0.01) at the end of the reach phase of reach-to-grasp, but there was no significant difference during the other motor tasks.
Figure 7E shows the mean ulnar facilitation obtained at 0.75 and 1 × MT at the end of the reach phase of reach-to-grasp and of reach-to-point and during tonic contraction (protocol 1). Two-way ANOVA revealed a significant influence of the motor task (P < 0.001) and of ulnar nerve stimulation intensity (P < 0.001), with significant interaction between the two factors (P < 0.001). This result indicates that the influence of stimulation intensity on the level of ulnar facilitation changed according to the motor task. Indeed, post hoc analyses revealed a larger facilitation at 0.75 × MT, especially at the end of the reaching phase of reach-to-grasp (Fisher's LSD, P < 0.01).
Figure 7F illustrates the group data from protocol 3 in which the effect of ulnar nerve stimulation intensity was tested during the reaching and grasping phases of reach-to-grasp. Here again, the effect of ulnar facilitation was significantly influenced by the movement phase (P < 0.001) and by the ulnar nerve stimulation intensity (P < 0.001), with a significant interaction between factors (P < 0.001). This result again indicates that the ulnar facilitation was significantly larger at 0.75 × MT, at the end of reach of reach-to-grasp.
This study has shown that the ulnar-induced facilitation of FCR MEP was not accompanied by facilitation of the H-reflex and was particularly enhanced at the end of a reaching movement, only when associated with grasping. Increasing the intensity of ulnar nerve stimulation depressed the MEP facilitation, and no difference was then found between motor tasks and movement phases.
Origin of Ulnar Facilitation of FCR MEP
Ulnar nerve stimulation facilitated the MEP of FCR at a longer ISI than expected if the peripheral and corticospinal inputs converged at motoneuron level. Moreover, if the convergence occurred at the spinal motoneuron, the H-reflex of FCR would have been similarly facilitated. These findings indicate that the site of convergence was at a spinal interneuron, probably some 3–4 ms remote from the motoneuron pool of FCR. Increasing the intensity of the peripheral nerve stimulus caused the facilitation to disappear. Ulnar facilitation with similar characteristics has been documented in previous studies (Iglesias et al. 2007; Lourenço et al. 2006, 2007), and, given the similarity to the spinal excitation transmitted through propriospinal-like interneurons (Pierrot-Deseilligny and Burke 2005), it has been proposed that muscle spindle group I afferents from the intrinsic muscles of the hand activate both propriospinal neurons and inhibitory interneurons projecting to them, to control the excitability of the motoneuron pool of FCR (Fig. 8). With conditioning stimuli of low intensity, the transmission across inhibitory interneurons is less effective than with more intense inputs, and this favors the appearance of propriospinal excitation. Increasing stimulus intensity strengthens the output of feedback inhibitory interneurons to propriospinal neurons, and this would then truncate any propriospinal excitation transmitted to spinal motoneurons (Iglesias et al. 2007; Nicolas et al. 2001; Pauvert et al. 1998).
Possible Mechanisms Underlying Task-Related Changes in Ulnar Facilitation
Because the initial position of the hand and the position and orientation of the target strongly influence goal-directed movements (Desmurget et al. 1998; Roby-Brami et al. 2000), the subjects were asked to perform reaching tasks from a similar starting position and the target did not change. Under these conditions, the motor sequence is reproducible, as supported by kinematic recordings showing similar variations in angle joint position at shoulder and elbow levels during reach-to-grasp and pointing movements (Fig. 1, D and E; Corradini et al. 1992). This suggests that different distal motor strategies do not require different proximal movements. However, wrist position, hand preshaping, and contact with the target are likely to have produced differences in peripheral inputs between movements and movement phases, influencing spinal cord excitability. Other factors such as descending inputs and motoneuron recruitment could also change with task and task phase.
Ulnar nerve stimulation was comparable between tasks (motor threshold, stimulus intensity), and the electrically induced afferent input was therefore probably similar. However, the naturally evoked afferent inputs from the moving limb were probably different, and they may have participated in the online correction of hand orientation (Gosselin-Kessiby et al. 2008; Jones et al. 2001a). ADM EMG activity was weak during pointing and tonic contraction, but there was probably activation of other hand muscles because the fist was clenched. However, the ulnar-induced facilitation of FCR MEP was larger only during the reaching phase of the reach-to-grasp movement, suggesting that additional inputs participate in the task-related enhanced propriospinal excitation; presumably these factors include task-specific differences in the descending command. Visual inputs are also involved in the control of goal-directed movement (Jones et al. 2001b), but not when a stationary target is reached (Gosselin-Kessiby et al. 2008).
The ulnar-induced facilitation of FCR MEP could not be evoked during the subsequent grasp. At this stage, powerful cutaneous inputs coming from the palmar side of the hand may have reduced propriospinal excitability. Indeed, cutaneous inputs from the palmar side project to inhibitory interneurons controlling propriospinal neurons and can truncate propriospinal excitation of FCR motoneurons (Fig. 8; Nielsen and Pierrot-Deseilligny 1991). Such inhibitory control from cutaneous afferents is stronger at the end of elbow extension, suggesting a role in movement termination (Pierrot-Deseilligny 1996; D'Aponte et al. 1999).
Brain imaging has revealed increased activity during both pointing and grasping tasks, in various common cortical areas including the contralateral primary motor cortex (M1), premotor cortex (PM), ventral supplementary motor area, cingulate, and superior parietal and dorsal occipital cortex, with some task-related differences (Grafton et al. 1996; Filimon 2010). Accordingly, studies using TMS have revealed task-, condition- and time-related changes in cortico-cortical interactions between PM or posterior parietal cortex and M1 during reaching movements (see Koch and Rothwell 2009).
Less is known about the resulting corticospinal outputs and the transmission of the motor command to spinal motoneurons. However, the results of TMS and brain imaging are consistent with behavioral studies showing that the nature of the target strongly influences the reaching movement (Desmurget et al. 1998; Roby-Brami et al. 2000), and these indicate that, not surprisingly, the descending command is different according to the executed movement. In this way, the plateau level of MEP recruitment curves of wrist extensors is higher during pointing than during isolated wrist extension, much as we observed here in wrist flexors. This modulation was observed with TMS but not transcranial electrical stimulation, indicating a cortical origin (Devanne et al. 2002). This is in agreement with results from brain imaging and paired-pulse TMS studies, but it is not yet possible to explain how similar descending inputs targeting spinal neurons at the same level of excitability lead to different motor behaviors. It is likely that cortico-subcortico-cortical interactions influence the descending inputs to spinal cord and thus the recruitment of spinal neurons.
Spinal neuron recruitment.
TMS was adjusted during each motor task to evoke test MEPs of similar size. This was not possible for tonic contractions, but recruitment curves revealed differences in plateau level between tasks, as has been previously reported for wrist extensors (Devanne et al. 2002). Importantly, the ulnar nerve stimuli were delivered with a test MEP that was on the equivalent position on the recruitment curve (the more linear part of the sigmoid), suggesting that the excitability levels of the FCR motoneuron pool were comparable. Accordingly, the ulnar nerve stimulation did not produce any change in FCR H-reflex during reaching, and that in wrist extensors was not influenced by the motor task (Devanne et al. 2002).
The recruitment curves suggest that TMS activated a greater proportion of the spinal motoneuron pool during reaching tasks than during tonic contractions, i.e., that the stimuli activate higher-threshold motoneurons more readily (Lackmy and Marchand-Pauvert 2010). The C3–C4 propriospinal system in the cat is known to depart from the Henneman size principle, allowing early recruitment of large motoneurons for visually guided movements (Alstermark and Sazaki 1986). In humans, the nonmonosynaptic excitation mediated by propriospinal neurons is evenly distributed to slow and fast motor units (Marchand-Pauvert et al. 2000). Although the maximal MEP was larger during pointing (Devanne et al. 2002), enhanced propriospinal excitation was observed only during reach-to-grasp in the present study. During pointing, the propriospinal system is not involved in motoneuron recruitment, but during reach-to-grasp it participates in the recruitment of higher-threshold motoneurons.
Propriospinal excitation vs. inhibition.
The ulnar-induced propriospinal facilitation of FCR motoneurons was particularly increased at the end of the reach but not during the subsequent grasp. This may appear inconsistent with our previous study showing enhanced propriospinal excitation during grasping tasks (Iglesias et al. 2007). However, in this study, the enhanced propriospinal excitation during power grip was probably due to release of inhibition by the descending command, because the difference between tasks was observed only when ulnar nerve stimulation was above the threshold for feedback inhibitory interneurons projecting to propriospinal neurons (Fig. 8). A specific increase in ulnar-induced facilitation of shoulder muscle MEP has also been described with the use of weak conditioning stimulus intensities, at the end of a grip-lift task (Roberts et al. 2008).
Modulation of the propriospinal excitation is likely to be different during static and dynamic movements. During rapid reach-to-grasp movements, the excitability of propriospinal neurons appears to be increased because of a focused descending command favoring early recruitment of large motoneurons; the convergence of proprioceptive afferents from hand muscles would assist the wrist position to optimize hand preshaping. During the grasping phase, the present results suggest that the propriospinal excitation is truncated by cutaneous inhibition of propriospinal neurons, terminating the reaching movement (D'Aponte et al. 1999; Pierrot-Deseilligny 1996).
There is considerable literature, based on kinematic analyses, visual and proprioceptive processing, brain imaging, and TMS, supporting an internal model for fast reaching movements, with a feedforward component (motor planning) and a feedback component (sensory processing; Desmurget and Grafton 2000). The propriospinal system allows rapid adjustments during movements at spinal level, and given its interface between motor cortex and afferent feedback, its inhibitory control, and its ascending collaterals (see Alstermark et al. 2007), it could send an efference copy to supraspinal structures to update the initial motor command. In humans an analogous system may be similarly involved in the internal model for reaching movements. Indeed, the propriospinal excitation is enhanced during rapid movements, and its inhibition is opened at the end of the movement but depressed during power grip. These mechanisms are of functional interest depending on the tasks. 1) During rapid movements, propriospinal activation by the descending command (feedforward concept) would assist the early recruitment of large motoneurons, without depressing the inhibitory control necessary to avoid an obstacle during movement (feedback concept). Disfacilitation of spinal motoneurons would leave them capable of responding to an updated motor command, unlike the situation that would occur with direct inhibition of the motoneuron pool. 2) Cutaneous disfacilitation at the end of reach would then assist the termination of movement. 3) During power grip there is no need to activate large motoneurons, and a lesser inhibition would allow sensory feedback to assist motoneuron discharge for tonic contraction.
The propriospinal contribution to FCR contraction was less during pointing, whether dynamic or static (Iglesias et al. 2007). Two mechanisms probably contribute to this: both the afferent feedback to the propriospinal neurons and the centrally generated command would have differed. In relation to the latter, pointing tasks are particularly used in communication, whereas grasping is a motivated motor action, both being controlled by different cortical areas (Edwards and Humphreys 1999). This supports the role of the propriospinal system in prehension in primates and humans (Alstermark et al. 2011; Isa et al. 2007).
The present study suggests specific modulation of transmission through propriospinal neurons during reaching only when associated with grasping. This implies that the propriospinal system is particularly involved in the transmission of the descending command for prehension, much as has been observed in primates but not the cat (Alstermark et al. 2007). We presume that this system is involved in the internal model for reach-to-grasp movement involving feedforward activation of the relevant spinal motoneurons and the feedback modification necessary to update the descending command and to terminate movement.
The study was supported by UPMC Université Paris 6, Assistance Publique-Hôpitaux de Paris (AP-HP), Institut pour la Recherche sur la Moelle Epinière (IRME), Institut National de la Santé et de la Recherche Médicale (INSERM), and the National Health and Medical Research Council of Australia (NHMRC). L.-S. Giboin was supported by a grant from UPMC Université Paris 6 (Ministère de l'Enseignement Supérieur et de la Recherche).
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: L.-S.G., A.L.-V., D.B., and V.M.-P. performed experiments; L.-S.G., A.L.-V., and V.M.-P. analyzed data; L.-S.G., A.L.-V., D.B., and V.M.-P. interpreted results of experiments; L.-S.G., D.B., and V.M.-P. prepared figures; L.-S.G., D.B., and V.M.-P. drafted manuscript; L.-S.G., A.L.-V., D.B., and V.M.-P. edited and revised manuscript; L.-S.G., A.L.-V., D.B., and V.M.-P. approved final version of manuscript; D.B. and V.M.-P. conception and design of research.
- Copyright © 2012 the American Physiological Society