Short-term upper limb disuse induces a hemispheric unbalance between the primary motor cortices (M1s). However, it is still unclear whether these changes are mainly attributable to the absence of voluntary movements or to the reduction of proprioceptive information. The goal of this work was to investigate the role of proprioception in modulating hemispheric balance during a short-term right arm immobilization. We evaluated the 2 M1s excitability and the interhemispheric inhibition (IHI) between M1s in 3 groups of healthy subjects. Two groups received during the immobilization a proprioceptive (P-VIB, 80 Hz) and tactile (T-VIB, 30 Hz) vibration to the right hand, respectively. Another group did not receive any conditioning sensory inputs (No-VIB). We found that in the No-VIB and in the T-VIB groups immobilization induced a decrease of left M1 excitability and IHI from left to right hemisphere and an increase of right M1 excitability and IHI from right to left hemisphere. Differently, only a partial decrease in left M1 excitability, no change in right M1 excitability and in IHI was observed in the P-VIB group. Our findings demonstrate that the maintenance of dynamic proprioceptive inputs in an immobilized arm through muscle vibration can prevent the hemispheric unbalance induced by short-term limb disuse.
Primary motor cortex (M1) can exhibit considerable plasticity following pathological or traumatic changes and in relation to everyday experience (Cohen et al. 1993; Sanes and Donoghue 2000; Chen et al. 2002; Draganski et al. 2004; Hummel and Cohen 2005; Dayan and Cohen 2011). It has also been demonstrated that a short-term upper limb nonuse can induce brain plasticity in M1 (Facchini et al. 2002; Huber et al. 2006). From these works, a question raises and concerns whether the plastic changes in M1s induced by limb nonuse are mainly attributable to the absence of voluntary movements during hand immobilization or to the reduction of dynamic proprioceptive information from the immobilized hand or to the combination of both. Indeed, proprioception might be affected by limb immobilization as already shown for tactile perception (Lissek et al. 2009). A first answer to the previous question has been given by a recent work showing by fMRI that a proprio-tactile stimulation administered during 5-days of right hand immobilization prevents the disruption of sensorimotor networks located in the left hemisphere (Roll et al. 2012).
However, it is worth noting that unilateral limb nonuse is able to induce plastic changes not only in the contralateral M1, but also in the ipsilateral one, because of the transcallosal connection between these areas. We recently demonstrated that limb nonuse reduced the excitability of the contralateral M1 and decreased the inhibitory influence onto the ipsilateral one, while the contrary occurred for the ipsilateral M1 (increased excitability and deeper inhibitory interaction onto the contralateral M1) (Avanzino et al. 2011). This topic is extremely relevant because it has been proposed that balanced interhemispheric interactions within the sensorimotor networks are required for the generation of proper unilateral voluntary movements (Bloom and Hynd 2005; Carson 2005; Beaulé et al. 2012). Also, an abnormally high interhemispheric inhibitory drive from the primary motor cortex of the intact to the lesioned hemisphere was observed in patients with chronic subcortical stroke, and this phenomenon was more prominent in cases with greater motor impairment (Murase et al. 2004), suggesting that interhemispheric balance can influence functional recovery.
Therefore, the aim of our work was to assess the role of proprioception in preventing interhemispheric unbalance induced by short-term limb nonuse. Here, for the first time, we investigated how proprioception can modulate not only the activity of the left primary motor cortex primarily influenced by immobilization (i.e., contralateral to immobilized arm) but rather the activity of 2 motor cortices and their interaction via corpus callosum. Thus, we aim to have a complete picture of how the 2 motor cortices respond to an unilateral limb nonuse in absence or presence of a proprioceptive feedback.
To this goal, we investigated both the excitability of the 2 M1s and the transcallosal interaction between them, by evaluating the interhemispheric inhibition (IHI), in 3 groups of healthy subjects whose right hand was immobilized. One group received a high-frequency, 80-Hz vibration to the right hand during the immobilization period to reproduce trains of dynamic proprioceptive inputs (P-VIB) (Burke et al. 1976; Matthews 1988) while another one did not receive any conditioning sensory inputs to the right hand (No-VIB). Then, a third group received a low-frequency, 30-Hz vibration (T-VIB) suitable to activate rapidly adapting intradermal tactile receptors (Meissner corpuscles) (Weerakkody et al. 2007) to explore the specificity of the sensory input effect (Rosenkranz and Rothwell 2004). We expect that only proprioceptive and not tactile inputs administered to the right immobilized hand might cope with the downregulation of the contralateral M1 occurring after 10 h of limb immobilization (Avanzino et al. 2011). Further, proprioceptive signals might also influence ipsilateral M1 excitability through a transcallosal sensorimotor integration mechanism (Swayne et al. 2006).
Materials and Methods
Twenty-eight subjects were recruited for the study and were divided in 3 groups (no-vibration group (No-VIB), 10 subjects; proprioceptive vibration group (P-VIB), 10 subjects and tactile vibration group (T-VIB), 8 subjects). Participants in the No-VIB group did not receive any conditioning sensory inputs to the right hand during the 10h of hand immobilization, while participants in the P-VIB and T-VIB group received, during the immobilization period, a high-frequency, 80-Hz vibration to the right hand and a low-frequency, 30-Hz vibration to the right hand, respectively. The 3 groups were matched for age and gender (No-VIB, mean age, 25.4 ± 3 years, 6 females; P-VIB, mean age, 23.8 ± 4 years, 5 females; T-VIB, mean age, 24.4 ± 5 years, 4 females). All participants were right handed, as determined by the Edinburgh Handedness Inventory (Oldfield 1971). All subjects were naive to the purpose of the experiment. They reported no previous history of neurological disorders or orthopedic problems. Subjects had no contraindication to transcranial magnetic stimulation (TMS), and they participated in this study after giving an informed consent. The study was conducted in accordance with the Declaration of Helsinki.
We used the same immobilization procedure previously described in a recent paper by our group (Avanzino et al. 2011). All subjects were instructed not to move their right hand for 10 h from the morning (8:00 AM) to the evening (6:00 PM). To prevent any right hand movement, subjects wore a dorsal wrist-hand-fingers orthosis, typically used in clinical practice, to immobilize the joints (fingers and wrist). In addition, a soft shoulder and elbow splint was used to support the forearm in a comfortable way during the 10 h of hand immobilization.
All participants received no instructions concerning the left arm use, which was completely free to move. We quantitatively monitored the physical activity duration of the left arm by means of an accelerometer set up in a multisensor actigraph (InnerView Professional, SenseWear PRO Armband). Subjects wore this device on their left forearm for 2 days, 1 day before, and 1 day during the immobilization period, from ∼8:00 AM to 6:00 PM in each day. The multisensor actigraph records the cumulative amount of time spent (in minutes) during physical activity at a certain level of energy expenditure by means of mathematical algorithms. Data were sampled at 32 Hz. The energy cost of an activity can be measured in units called metabolic equivalent (METs). Using the definition for 1 MET as the ratio of work metabolic rate to a standard resting metabolic rate, 1 MET is considered a resting metabolic rate obtained during quiet sitting. Here, to record physical activity duration, we set a threshold level of 1.5 METs that is usually the energy expenditure during deskwork (Ainsworth et al. 2000). Figure 1 summarizes the experimental protocol.
Electromyography (EMG) was recorded with silver disc surface electrodes placed in a tendon belly arrangement over the bulk of the first dorsal interosseus (FDI) muscle and the first metacarpophalangeal joint bilaterally. The ground electrode was placed at the elbow. EMG signals were amplified and filtered (20 Hz to 1 kHz) with a D360 amplifier (Digitimer). The signals were sampled at 5000 Hz, digitized using a laboratory interface (Power1401; Cambridge Electronics Design), and stored on a personal computer for display and later offline data analysis. Each recording epoch lasted 400 ms, of which 100 ms preceded the TMS. Trials with background EMG activity were excluded from analysis.
Transcranial Magnetic Stimulation
We tested cortical excitability of left and right M1s by means of input/output (IO) recruitment curve and interhemispheric communication between the 2 M1s assessing the IHI. TMS procedure was performed 1 day before (PRE) and immediately after 10 h of hand nonuse (POST), always at ∼6:00 PM.
For IO curves study, TMS was performed with a single Magstim 200 magnetic stimulator (Magstim Company) connected with a figure-of-8 coil with wing diameters of 70 mm. For IHI study, TMS was given through 2 Magstim 200 stimulators, one connected to a figure-of 8 coil with wing diameters of 70 mm (test stimulus, TS) and the other connected to a small figure-of-8 coil with wing diameters of 50 mm (conditioning stimulus, CS) (Udupa et al. 2010).
The coils were placed tangentially to the scalp with the handle pointing backward and laterally at a 45° angle to the sagittal plane inducing a posteroanterior current in the brain. We determined the optimal position for activation of the left and right FDI muscles by moving the coil in 0.5 cm steps around the presumed motor hand area. Resting motor threshold (RMT), defined as the minimum stimulus intensity that produced a motor evoked potential (MEP) of at least 0.05 mV in 5 of 10 consecutive trials, was found and expressed as a percentage of maximum stimulator output (MSO).
Input Output Curve
IO curve was examined by measuring peak-to-peak amplitude (expressed in mV) of MEPs elicited at stimulus intensities of 5%, 10%, 15%, 20%, and 25% of MSO above RMT (calculated on the values of the individual RMT obtained each day). Ten trials were recorded at each stimulus intensity, and the average MEP amplitude was taken as MEP size.
Further, the data from each subject were pooled together and a linear regression of MEP values as a function of stimulus intensities in the linear part of the IO curve (between 5% and 20% of MSO) was performed in all groups before and after immobilization.
IHI both from left to right (LtoR) and from right to left (RtoL) M1s were tested by a randomized conditioning test design reported previously (Ferbert et al. 1992). A suprathreshold CS set at 130% of the RMT was given 10 ms before a TS delivered to the contralateral side. The TS was adjusted to produce a MEP of 1 mV peak-to-peak amplitude. Stimuli were randomly delivered in one set of 30 trials: 15 conditioned and 15 unconditioned. IHI was expressed as the ratio between the mean peak-to-peak MEP amplitude in conditioned versus unconditioned trials. In each day, CS was adjusted on the basis of the individual RMT found in that session, and TS was adjusted to obtain a 1-mV MEP amplitude.
In proprioceptive vibration group (P-VIB) and in the tactile vibration group (T-VIB), during the immobilization, the right immobilized hand was vibrated through a vibrator with a head of about 5 mm diameter (Vibralgic model, Electronic Conseil, Alès, France). The vibrator was fixed to adjustable devices so that it could be easily positioned perpendicularly to the right FDI muscle through an appropriate small hole in the bandage. In both groups, we vibrated the FDI muscle of the right totally relaxed hand but with different frequency and amplitude parameters.
In the P-VIB group, the vibrator was driven at 80 Hz with an amplitude of 80% of the maximum amplitude (corresponding to 5-mm axial displacement) to activate muscle spindles (Burke et al. 1976; Matthews 1988); whereas in the T-VIB group, the vibrator was driven at 30 Hz with an amplitude of 15% of the maximum amplitude in order to activate tactile receptors (i.e., Meissner corpuscles) (Weerakkody et al. 2007).
Because it is well known that visual feedback of the vibrated limb facilitates the occurrence of a tonic vibration reflex (TVR) (Roll et al. 1980), subjects were blindfolded and EMG activity was recorded during the first 2 sessions of vibration (i.e., during the first hour of immobilization).
The FDI muscle of the right immobilized hand was vibrated twice every hour from 9:00 AM to 5:00 PM (17 times in total). In particular, each time the vibration procedure lasted 7.5 min consisting of 1-min vibration ON followed by 30-s vibration OFF, repeated 5 times. The effect of vibration on M1s cortical excitability and IHI was evaluated only at the end of the 10 h of immobilization and not during.
The temporal feature of the vibration protocol was designed following the data of the physical activity duration of the right arm for 10 h, recorded by means of an actigraph on 12 subjects some weeks before the experiment. During 10 h of normal activity, the physical activity duration ranged between 80 and 90 min, corresponding to about 10 min of activity per hour. On the basis of this result, in this work, we decided to administer the vibration stimulus, usually inducing the illusion of movement, uniformly distributed during the day. Further, the choice to alternate vibration (ON) and resting periods (OFF) was done to avoid the elicitation of strong synaptic depression phenomena both during the vibration and in the postvibration period (Hultborn et al. 1987; Bove et al. 2003) that could affect the stimulation.
The left-arm physical activity durations measured the day before and during immobilization and the RMTs of left and right hemisphere were separately entered in a repeated-measures (RM) ANOVA with the factor GROUP (No-VIB, P-VIB and T-VIB) as between subjects factor and TIME (PRE and POST immobilization) as within subjects factor.
TMS IO curve data were analyzed, separately for each hemisphere by means of an RM-ANOVA with GROUP (No-VIB, P-VIB, and T-VIB) as between-subjects factor and TIME (PRE and POST immobilization) and intensity (5%, 10%, 15%, 20%, and 25% of MSO above RMT) as within-subjects factors. Then, we evaluated the differences between the slopes of the 6 regression lines obtained by pooling together data in the different experimental conditions (PRE: No-VIB, P-VIB and T-VIB; POST: No-VIB, P-VIB, and T-VIB), separately for each hemisphere. In particular, we performed a multiple regression analysis and a Student's t-test on these data (Armitage 1971) to determine whether the linear part of the IO curve was significantly influenced by limb immobilization and, proprioceptive and tactile vibration.
Paired t-tests were performed to compare the conditioned versus the unconditioned trials of LtoR IHI and RtoL IHI at baseline (first day, 6:00 PM) in each group separately. Then, LtoR IHI and RtoL IHI of vibration and no-vibration groups were subjected separately to a RM-ANOVA with GROUP (No-VIB, P-VIB, and T-VIB) as between-subjects factors and TIME (PRE and POST immobilization) as within-subjects factors.
Significance threshold was set at P < 0.05. If ANOVA showed a significant interaction effect, we performed post hoc comparisons using the least significance difference (Fisher's) test to directly compare the experimental factors. All statistical analyses were performed by using SPSS 13.0. Data are presented as mean ± SE.
All subjects tolerated TMS procedure. Physical activity duration of the left arm increased during immobilization in all subjects. Accordingly, RM-ANOVA showed a significant effect of TIME (F1,25 = 50.00, P < 0.0001) and post hoc analysis revealed that the duration of physical activity of the left arm recorded for 10 h in the day before the immobilization procedure was shorter than that recorded during the 10 h of immobilization (P < 0.0001). No difference among the 3 groups was observed (GROUP and TIME*GROUP; P > 0.05).
All subjects of the P-VIB group reported sensations of a continuous and slow illusory movement throughout vibration, while experiencing only relatively transient sensations of hand displacements after the end of the vibration procedure (i.e., after-effects). No overt limb movements were observed during vibration and EMG activity of FDI did not show any tonic vibration reflex during the first 2 sessions of vibration.
Left M1 (Contralateral to the Immobilized Hand)
For RMT of the left M1, RM-ANOVA showed a significant interaction TIME*GROUP (F2,25 = 5.71, P = 0.009). Post hoc analysis revealed that RMT significantly increased in the No-VIB (P < 0.001) and T-VIB (P < 0.001) groups whereas it did not change in the P-VIB group (P = 0.26) (Table 1).
|Left M1||PRE||35.80 ± 2.04||35.50 ± 3.62||36.45 ± 4.26|
|POST||37.70 ± 1.82||35.90 ± 3.87||38.20 ± 4.25|
|Right M1||PRE||36.5 ± 1.35||35.5 ± 2.63||37.16 ± 1.14|
|POST||36.8 ± 1.39||35.3 ± 2.71||37.41 ± 1.16|
|Left M1||PRE||35.80 ± 2.04||35.50 ± 3.62||36.45 ± 4.26|
|POST||37.70 ± 1.82||35.90 ± 3.87||38.20 ± 4.25|
|Right M1||PRE||36.5 ± 1.35||35.5 ± 2.63||37.16 ± 1.14|
|POST||36.8 ± 1.39||35.3 ± 2.71||37.41 ± 1.16|
The excitability of the left hemisphere, when tested with IO curve, was significantly reduced by immobilization in all the groups of subjects (Fig. 2A–C). Accordingly, RM-ANOVA showed a main effect of TIME (F1,25 = 52.51, P < 0.001) and no main effect of GROUP or interaction TIME*GROUP (P > 0.05). Further, a main effect of INTENSITY (F4,100 = 39.49, P < 0.001) was found, suggesting that MEP amplitude increased with the increasing of the stimulator output.
The amplitude of MEP, trial by trial, was positively and linearly correlated with the TMS stimulus intensity, in the linear part of the IO curve between 5% and 20% of MSO, under all the 6 conditions for all subjects data collapsed. No difference was observed in the slope of the 2 vibration groups before immobilization with respect to the control group (Fig. 2D) (No-VIB PRE vs. P-VIB PRE: P = 0.43; No-VIB PRE vs. T-VIB PRE: P = 0.24). Further, immobilization reduced the slope of these relationships in all the groups of subject (Fig. 2D) (No-VIB PRE vs. No-VIB POST: P < 0.00001; P-VIB PRE vs. P-VIB POST: P = 0.016; T-VIB PRE vs. T-VIB POST: P = 0.0028). However, only proprioceptive vibration was able to cope with immobilization effect on corticospinal excitability allowing only a partial reduction of the slope. In fact, after immobilization, the slope of the left M1 RC was significantly greater in the P-VIB group with respect to the No-VIB group (No-VIB POST vs. P-VIB POST: P < 0.00001), whereas there was no difference between the T-VIB group and the No-VIB group (No-VIB POST vs. T-VIB POST: P = 0.1) (Fig. 2D).
Right M1 (Ipsilateral to the Immobilized Hand)
RMT of right M1did not change after immobilization in all groups. Accordingly RM-ANOVA showed no significant effect of TIME and GROUP and no interaction TIME*GROUP (P always > 0.05). All RMT data are reported in Table 1.
The excitability of the right M1, when tested with IO curve, was differently influenced by immobilization in the 3 groups of subjects (Fig. 2E–G). Indeed, it significantly increased in the No-VIB (Fig. 2E) and in the T-VIB (Fig. 2G) groups, but remained unchanged in the P-VIB (Fig. 2F) group. RM-ANOVA showed a significant interaction TIME*GROUP (F2,25 = 4.04, P = 0.030). Post hoc analysis showed that at baseline (first day, 6:00 PM) no difference among the 3 groups was observed. Further, only in the No-VIB and T-VIB groups, the immobilization procedure induced an increase of excitability of the right M1 (Fig. 2E,G) (No-VIB: P = 0.006; T-VIB, P = 0.005), while when immobilization was associated with proprioceptive vibration (P-VIB), the excitability of the right M1 did not change (Fig. 2F) (P = 0.74). Further, a main effect of INTENSITY (F4,100 = 26.34, P < 0.001) was found, suggesting that MEP amplitude increased with the increasing of the stimulator output.
Again, also for the right hemisphere, the amplitude of MEP, trial by trial, was positively and linearly correlated with the TMS stimulus intensity, in the linear part of the IO curve between 5% and 20% of MSO, under all the 6 conditions for all subjects' data collapsed. No difference was observed in the slope of the 2 vibration groups before immobilization with respect to the control group (Fig. 2H) (No-VIB PRE vs. P-VIB PRE: P = 0.17; No-VIB PRE vs. T-VIB PRE: P = 0.12). Immobilization affected the slope of the regression lines obtained from the IO of the right M1 only in absence of the muscle proprioceptive vibration (No-VIB PRE vs. No-VIB POST: P < 0.00001; T-VIB PRE vs. T-VIB POST: P = 0.006), while left the slope of these relationships completely unchanged in the presence of the proprioceptive muscle vibration (P-VIB PRE vs. P-VIB POST: P = 0.77) (Fig. 2H).
At baseline (first day, 6:00PM) when the TS was preceded by a conditioning one delivered 10 ms earlier in the contralateral hemisphere, a significant reduction of the MEP size was observed either in the right and left hemisphere with respect to unconditioned trials (TS alone) in all groups of subjects (No-VIB: LtoR IHI: P = 0.0012; RtoL IHI: P = 0.002; P-VIB: LtoR IHI: P = 0.00001; RtoL IHI: P = 0.00004; T-VIB: LtoR IHI: P = 0.0006; RtoL IHI: P = 0.00008).
After immobilization, LtoR IHI was reduced in the No-VIB and in the T-VIB groups, while it remained unchanged in the group of subjects who underwent proprioceptive vibration (P-VIB) during the 10 h of immobilization (Fig. 3A). Accordingly, RM-ANOVA revealed a main interaction TIME*GROUP (F2,25 = 7.51; P = 0.003) and post hoc analysis showed that in the No-VIB and in the T-VIB groups LtoR IHI decreased after immobilization No-VIB, P = 0.002, T-VIB, P < 0.0001), while in the P-VIB group no change was detected (P = 0.84) (Fig. 3A).
After immobilization, RtoL IHI increased in the No-VIB and in the T-VIB groups, while it remained unchanged in the group of subjects who underwent proprioceptive vibration (P-VIB) during the 10 h of immobilization (Fig. 3B). Accordingly, RM-ANOVA a main interaction TIME*GROUP (F2,25 = 4.29; P = 0.025) was found and post hoc analysis showed that in the No-VIB in the T-VIB groups RtoL IHI increased after immobilization (No-VIB, P = 0.002, T-VIB, P = 0.045), while in the P-VIB group, no change was detected (P = 0.57) (Fig. 3B).
In this work, we demonstrated that the maintenance of dynamic proprioceptive inputs in an immobilized arm through muscle vibration can prevent the hemispheric unbalance induced by short-term limb nonuse. Here, the knowledge on the interaction between proprioception and upper limb immobilization and the resulting effects on controlateral M1 activity was inserted in a more general framework including the transcallosal interaction between the 2 hemispheres and the ipsilateral M1 activity.
Effect of Right FDI Proprioceptive Vibration on the Left M1
First, we found that right hand proprioceptive vibration was able to significantly reduce the effect induced by limb immobilization on corticospinal excitability of the hemisphere contralateral to the immobilized hand (i.e., left M1). After immobilization, the slope of the IO curve significantly decreased in all groups, but in P-VIB group, it was significantly steeper than that of the no-vibration and tactile vibration groups indicating that the administration of proprioceptive inputs during immobilization can partially cope with the downregulation effect on left M1. Our neurophysiological findings are complementary to recent neuroimaging data by Roll et al. (2012) reporting that the sensorimotor network of the left “immobilized” hemisphere was preserved when subjects received proprio-tactile stimulation during hand immobilization, while the sensorimotor network of the no-vibrated subjects was significantly altered. More, we observed that even if participants in P-VIB and T-VIB groups received the same temporal pattern of sensory stimulation during the immobilization period, the effect of sensory feedback on corticospinal excitability was clearly stimulus dependent. First, the specificity of the effects related to proprioceptive vibration and not to tactile vibration fits finely with a large piece of evidence in the literature showing that muscle proprioceptive vibration exerts its influence on M1 neurons (Naito et al. 1999, 2002, 2007; Romaiguère et al. 2003; Rosenkranz and Rothwell 2003; Casini et al. 2006, 2008; Kito et al. 2006; Avanzino et al. 2012). Indeed, among somatic senses, proprioception is the sensory feedback mechanism for motor control. A large body of evidence in the literature supports this assumption; even if tactile and proprioceptive inputs project to the primary somatosensory cortex (although to different subcomponents), imaging and neurophysiological studies demonstrated that proprioceptive feedback, together with tactile feedback or alone, is largely stronger than tactile alone in activating or modulating M1 excitability (Radovanovic et al. 2002; Devanne et al. 2009; Mizuguchi et al. 2011). A direct access of proprioceptive input into M1 neurons and/or transcortical loops including sensorimotor associative areas have been called into question to explain the modulating effect of proprioceptive inputs on M1 excitability (Colebatch et al. 1990; Naito and Ehrsson 2001; Radovanovic et al. 2002; Naito et al. 2007).
Also, our findings support the idea that the mere fact of being involved in a stimulation paradigm during the immobilization period is not responsible of cortical excitability changes. However, the precise contribution of attention in modulating cortical excitability during immobilization might be object of future studies.
Effect of Right FDI Proprioceptive Vibration on Transcallosal Pathway and Right M1
The relevance of the present study is that we went beyond the previous observation on the influence of proprioceptive feedback on the sensorimotor network located in the hemisphere contralateral to the immobilized arm. Indeed, we demonstrated that also hemispheric balance can be modulated by the administration of proprioceptive inputs during limb immobilization.
First, the reduction of LtoR IHI induced by immobilization was completely abolished by right hand proprioceptive muscle vibration re-establishing the IHI value observed at baseline. This result is in line with a model in which increased proprioceptive inputs from the hand increases the inhibition onto muscles of the opposite hand (Swayne et al. 2006). IHI refers to the neurophysiological mechanism in which the M1 of one hemisphere inhibits the opposite M1. Therefore, since inputs raising from muscle spindle have direct access to contralateral sensory and motor cortical areas, the observed effect of proprioceptive vibration on IHI would suggest a direct influence of the vibration-induced proprioceptive inputs from the right hand onto the transcallosal neurons projecting from left to right M1.
Further, we observed that even if all subjects showed a similar and significant increase in the quantity of movement of the left arm during the immobilization period (i.e., overuse), the left arm overuse effects on the right M1 excitability, on either the corticospinal and transcallosal pathways, were different between the groups. In the No-VIB and T-VIB groups, we observed an increased excitability of the corticospinal and transcallosal neurons of right M1 while these changes completely vanished when proprioceptive muscle vibration was administered. This finding suggests that during immobilization left arm over-activity alone is not sufficient to influence right M1 excitability. In Avanzino et al. (2011), we stated that the increased excitability in right M1, observed after right hand immobilization, was likely to be not only dependent on the modification of the LtoR IHI but also on the left arm overuse. Here, we demonstrate that this dependence is bidirectional and that the left arm overuse by itself cannot modulate right M1 excitability. Therefore, the interaction between left arm overuse and transcallosal control from the left to the right M1 is needed to induce plasticity in right M1. Transcallosal control from left to right M1 arises from pyramidal neurons of cortical layer III in left M1 projecting, via corpus callosum, onto GABAergic inhibitory interneurons on right M1 which in turn modulate the excitability of corticospinal pyramidal neurons of cortical layer V in right M1 (Ferbert et al. 1992). Here, we demonstrated that when there is no increase in this GABAergic mediated inhibition onto right corticospinal neurons, i.e., no change in LR IHI (P-VIB condition), there is also no increased excitability of right corticospinal neurons, even in presence of left arm overuse during the 10 h of immobilization. Experimental evidence, based on animal studies, suggests that the activity of GABA-related cortical inhibition is important in controlling the extent to which plasticity may occur. In humans, it has been shown that plasticity is facilitated by a decrease of GABA-related inhibition (Ziemann et al. 1998) and that motor practice performed without a previous downregulation of the inhibitory GABAergic activity induced only mild or no changes in cortical excitability (Ziemann et al. 2001).
Here, we can assume that muscle vibration increasing proprioceptive inputs from right hand, re-establishes the correct transcallosal control from left to right M1 influencing the cortical plasticity due to limb overuse in right M1.
Limitations of the Present Study
Some limitations of this study deserve discussion. First, we acknowledge that our experimental paradigm does not allow to fully discard that spinal after effects following immobilization and/or muscle vibration could partially influence the results of the IO curve (Gillies et al. 1969; Hultborn et al. 1987; Lundbye-Jensen and Nielsen 2008a, b; Clark et al. 2010). Even if we did not examine whether our findings could have been due to alterations in spinal excitability, many others have provided convincing evidence that vibration applied as in the current study results in modulations at the cortical level (Kossev et al. 1999; Rosenkranz and Rothwell 2003). For example, Kossev et al. (1999) showed that MEP facilitation induced by muscle vibration was evident by using TMS but not transcranial electrical stimulation (TES). As TES activates corticomotoneurons at the axon hillock, whereas TMS activates cortical cells transynaptically (Edgley et al. 1990; Rothwell 1991; Rothwell et al. 1994), the vibration-induced augmentation of MEPs following TMS only indicates a cortical origin. Moreover, when spinal excitability was examined after immobilization in humans, controversial data have been reported, with a significant increase (Lundbye-Jensen and Nielsen 2008a, b; Clark et al. 2010) or no change (Facchini et al. 2002; Kaneko et al. 2003; Zhao et al. 2011).
Second, it is noteworthy that corticospinal and transcallosal output neurons receive inputs from several sets of intracortical interneuron networks which can be explored with paired-pulse TMS protocols (Kujirai et al. 1993; Trompetto et al. 2004; Avanzino et al. 2007). The activity of these facilitatory and inhibitory networks can be modulated by immobilization and vibration (Rosenkranz and Rothwell 2003; Clark et al. 2010), thus suggesting that changes in the intracortical interneurons activity are likely to be responsible for the excitability changes of corticospinal and transcallosal output neurons observed in this work.
In general, we think that all these issues do not negate the relevant findings of the present study; however, we also think that they should be addressed in future studies to expand the understanding of the physiological mechanisms dealing with the interaction between proprioception and motor systems.
Finally, we did not evaluate whether proprioceptive vibration could prevent the deterioration of motor performance induced by a short-term immobilization (Moisello et al. 2008; Bassolino et al. 2012). However, we recently demonstrated that, related to motor behavior, the initial immobilization effects decreased quickly trial by trial, following an exponential function till reaching values equal to those observed in the control condition (Bassolino et al. 2012). Therefore, longer immobilization periods (e.g., days) inducing greater and long-lasting immobilization effects could be more suitable to test the effect of vibration in preventing the deterioration of motor performance.
Relevance of the Present Findings
All these findings might have important clinical implications. In fact, it has been demonstrated that stroke patients show an unbalance activity between the two M1s supported by an abnormal stronger interhemispheric communication from the intact to the lesioned hemisphere (Hummel and Cohen 2006). A current approach to stroke rehabilitation is based on the constraint-induced movement therapy (CIMT) that implies the massed practice of the affected arm by re-training, through immobilization, the unaffected limb (Langhorne et al. 2009; Sirtori et al. 2009).
Here, we propose that the bases of our experimental paradigm could be used in rehabilitation to avoid hemispheric unbalance. Indeed, muscle vibration could be administered on the affected limb muscles with the aim to counteract the stroke-induced hemispheric unbalance, increasing the cortical excitability of the affected M1 and re-establishing a normal transcallosal interaction. At difference of CIMT, with the muscle vibration approach patients would be free to move the intact limb without increasing the intact M1 excitability which could induce an unwanted hemispheric unbalance from the intact to the affected M1.
Towards this goal, the design of ad hoc muscle vibration systems to induce specific proprioceptive patterns to muscle groups, able to create the illusion of performing goal-directed movements, may provide useful tools for sensorimotor rehabilitation purposes (Thyrion and Roll 2010).
No specific funding was received for this study.
Conflict of Interest: None declared.