The role of inhibition from the left dorsal premotor cortex in right-sided focal hand dystonia

Abstract

Background

The left dorsal premotor cortex (PMd) plays an important role in movement selection and is abnormally activated in imaging studies in patients with right-sided focal hand dystonia (FHD).

Objective

The aims of this study were to assess the role of left PMd in patients with focal hand dystonia (FHD) and in the genesis of surround inhibition, which is deficient in FHD.

Methods

Single and paired pulse transcranial magnetic stimulation (TMS) was applied during different phases of an index finger movement using the abductor pollicis brevis muscle (APB), a surrounding, non-synergistic muscle, as target muscle. To look at the effect of PMd on the primary motor cortex (M1), a sub-threshold conditioning pulse was applied to PMd 6 ms before stimulation over M1.

Results

There was surround inhibition during movement initiation in controls, but not in FHD patients. In contrast, FHD patients, but not controls, showed premotor-motor inhibition (PMI) at rest. During movement, PMI was absent in both groups.

Conclusion

We conclude that PMI does not appear to play a key role in the formation of surround inhibition in normal subjects, since it was not enhanced during movement initiation. However, in FHD, inhibition from PMd on M1 was abnormally increased at rest and declined during movement initiation. The behaviour of PMd can therefore partly explain the loss of surround inhibition in the FHD patients. The functional significance of increased PMI at rest is not clear, but might be an attempt of compensation for losses of inhibition from other brain areas.

Introduction

The dorsal premotor cortex (PMd) is thought to play an important role in the selection of movement execution in response to learned associations (1). Functional imaging studies show a side-dominance of the left PMd, which is activated during right- and left-handed movements, while the right PMd is more active during movement of the left hand depending on the task (2), Similarly, the application of transcranial magnetic stimulation (TMS) over the left PMd has been shown to disrupt motor performance bilaterally, while stimulation of the right PMd only impaired the performance of the left hand (3). Facilitatory and inhibitory effects have been identified between each PMd and the contralateral primary motor cortex (M1), in the phase before EMG onset (4). It appears that cued movements are facilitated, while prepared, but not selected movements are suppressed before EMG onset by PMd projections. Both effects are independent from circuits in M1, such as short intracortical inhibition (SICI) and intracortical facilitation (ICF) (4, 5). However, little is known about the intra-hemispheric interaction of left PMd and left M1.

Dystonia is generally regarded as motor execution abnormality, caused by a dysfunction in the cortico-striato-thalamo-cortical motor loop (6). Typical clinical features of focal hand dystonia (FHD) are task- and context-specific abnormal posturing due to sustained muscle contractions interfering with the performance of motor tasks (7). Neurophysiological findings in patients with dystonia are characterized by loss of inhibition on multiple levels of the central nervous system (8–10). In the dystonic M1, increased cortical excitability and deficiencies in intra-cortical inhibition have been found (11–13). Concerning the role of left PMd in dystonia, a recent electrophysiological study showed that transcallosal inhibition to the right M1 may be reduced (14). However, the intra-hemispheric effects towards the left M1, which is most commonly the dystonic M1, has not been assessed.

Previous neuroimaging studies have provided some evidence that the left PMd plays a role in the pathophysiology of dystonia (15 – 17). Although the results of different imaging studies are not consistent concerning other primary and secondary motor areas (15–18), a task-specific decrease of activation during writing was demonstrated for the left PMd (16), while another study found increased activation for other active tasks (15). This increased activation of PMd persisted even after successful treatment with botulinum toxin, which, for example, restored the pre-treatment hypo-activation in the sensorimotor area (15, 17). Therefore, it seems unlikely that altered activation of PMd is secondary to dystonic posturing. The deficient activation of PMd during writing and the decreased correlation between premotor cortical regions and the putamen rather suggest a dysfunction of the premotor-cortical network, which may arise from the basal ganglia (16). Another hint that PMd is in fact involved in the pathophysiology of FHD is derived from an interventional study, in which the application of low-frequency repetitive TMS (rTMS) over left PMd has been shown to result in a transient improvement of motor performance in FHD patients (19). However, the exact nature of premotor-motor interaction is not clear.

Another neural mechanism involved in skilled motor behaviour may be surround inhibition. Surround inhibition is a neural mechanism observed in M1 (20, 21), which is thought to enhance contrast between neural signals and has been shown to be deficient in FHD patients (22 – 24). Motor evoked potentials (MEPs) to un-involved, neighboring muscles are inhibited before and during the first peak of EMG in the synergistic muscle (22– 24). It is not clear, how surround inhibition is generated in healthy subjects. SICI may contribute (23, 24), but there might be inhibition from other cortical and sub-cortical areas as well. Since it has been shown that left PMd exerts inhibitory effects on the left M1, it might be possible that this inhibitory pathway contributes to the genesis of surround inhibition.

In the present study, we used a single and paired pulse TMS to assess surround inhibition and ipsilateral dorsal premotor-motor inhibition (idPMI) from the left PMd to the left M1 using a paradigm introduced by Civardi et al.(25). Civardi et al. (25) studied sites 4, 6 and 8 cm anterior to M1. As the PMd is most likely best localized with a TMS coil positioned about 3 cm anterior to M1, we studied that site and assumed that we might find similar results to what Civardi and collegues found at 4 cm anterior to M1. The aim of this study was to assess the modulation of idPMI in healthy volunteers and patients with FHD at rest and during movement in a surrounding muscle. First, we hypothesized that idPMI would contribute to surround inhibition. Second, we hypothesized that idPMI would be abnormally reduced in FHD patients during the phase of surround inhibition.

Methods

1. Subjects

Ten healthy volunteers (age 39–69 years, mean 51.7 ± 3.1 years; 3 females) and ten FHD patients (age 46–58 years, mean 53.1± 1.5 years; 3 females) participated in the study. All subjects were right-handed according to the Edinburgh handedness inventory (26). In all FHD patients, the right, dominant hand was affected. None of the participants had ever been exposed to neuroleptic drugs and had no history of other neuropsychiatric disorders, neurosurgery, metal or electronic implants. All participants gave their informed consent prior to the experiments, which were approved by the Institutional Review Board (IRB) of the National Institute of Neurological Disorders and Stroke (NINDS). Most of the FHD patients had been treated with local injections of botulinum toxin type A in the affected muscles. The last injection had been given at least three months before testing (see Table 1).

Table 1.

Patient demographics

Sex	Age [years]	Type	Duration [years]	Botulinum toxin/ last injection
F	46	WC	24	Yes / 4 years
F	48	WC	11	Yes / 3 months
F	52	WC	11	Yes / 3 months
M	46	MC	3	No
M	52	MC	9	Yes / 3 months
M	57	MC	11	Yes / 11 years
M	57	MC	7	No
M	57	MC	17	Yes / 5 months
M	57	MC	22	No
M	58	MC	6	Yes / 4 months

WC = writer’s cramp; MC = musician’s cramp

2. Recording

Subjects were seated in a chair with their right arm resting on a board beside them, to which a force transducer was attached (Strain Measurement Devices, Inc, Meriden, CT, model S215 load cell). In some subjects, the wrist was supported by a towel to help the subject keep the hand and arm muscles relaxed. Acoustic feedback from the EMG machine (Nicolet Viking, Skovlunde, Denmark) was used to ensure a relaxed state throughout the experiment.

Disposable surface silver-silver chloride EMG electrodes were placed on the abductor pollicis brevis muscle (APB) and first dorsal interosseous muscle (FDI) of the right hand in a belly-tendon montage. Impedance was reduced below 5 kΩ. The EMG signal was amplified using a conventional EMG machine (Nicolet Viking, Skovlunde, Denmark) and bandpass filtered (20–2000 Hz). The signal was digitized at a frequency of 5 kHz and fed into a computer for off-line analysis.

3. Motor task

A simple acoustic reaction time task was applied that had previously been introduced to assess surround inhibition in APB (24). In brief, subjects performed an index finger flexion in the metacarpal-phalangeal joint (MCP joint) in response to a tone. FDI participates as synergist rather than as prime mover in this motion, but it has been shown that the modulation of cortical excitability is similar to prime movers (21).

Subjects exerted 10% of their maximum force (F_max) of index finger flexion as fast as possible after the onset of an acoustic signal. The acoustic signal lasted 200ms. Subjects maintained the contraction for approximately 500 ms. The force level was individually adjusted and displayed online as a target line on an oscilloscope in front of them. The output of the force transducer was also displayed on the oscilloscope as direct online feedback. Subjects practiced the task at the beginning of the experiment to attain a consistent motor performance.

At the same time, EMG from APB was monitored. APB is not involved in the task and therefore remained relaxed throughout all experiments. Trials in which there was background EMG in APB, assessed as root mean square over 50ms prior to MEP onset in each phase, were rejected. Once the subjects showed consistent motor performance, three different phases of the movement were assessed: rest, pre-EMG (75ms before EMG onset in FDI) and phasic (first peak of EMG in FDI). In order to be able to randomize the single pulse trials, rest stimulation was given 100ms before the onset of the acoustic signal. EMG onset and first peak was measured individually in an average of FDI EMG in ten consecutive trials. For the pre-EMG phase, we stimulated 75ms before EMG onset in FDI, which corresponds approximately to the time interval used in a previous report, where single pulse TMS over PMd has been shown to interfere with motor performance (3).

4. TMS

For TMS, two high-power Magstim 200 machines (Magstim Co., Whitland, Dyfed, UK) were connected to two custom-made figure-of-8 coils with an inner loop diameter of 35mm. At the beginning of each experiment, the “motor hot spot” for eliciting MEPs in APB was determined. This position was marked on a tight fitting EEG cap to ensure proper coil placement throughout the experiment. Coil orientation was tangential to the scalp with the handle pointing backwards and laterally at a 45-degree angle away from the midline inducing a posterior-directed current in the brain to activate the cortico-spinal system preferentially trans-synaptically (27). Resting (MT) and active motor threshold (aMT) were determined for each subject. MT was defined as the minimal stimulation intensity needed to induce a MEP of 50 μV in at least five out of ten consecutive trials at rest, whereas aMT was defined as the minimal stimulation intensity needed to induce a MEP of 200μV in at least five out of ten consecutive trials during a contraction of the target muscle of 10%F_max.

In experiment 1, single pulse TMS with an intensity of 140%MT was applied over the APB motor hotspot. MEPs were recorded from the right APB and FDI. Two similar sessions were performed consisting of 45 trials each. Four different conditions were tested in each session in a randomized order: rest (stimulation 100ms before the onset of the acoustic signal), pre-EMG (stimulation 75ms before EMG onset in FDI), phasic (stimulation during the first EMG peak in FDI) and one condition, in which no stimulation was given. Therefore, 22 trials were recorded per condition overall. Trials, in which the subject reacted more than 30mg earlier or later than the average reaction time, recorded as start of force production, were rejected. On average, this resulted in 18–22 MEPs that were analyzed for each subject in each condition. MEP size was determined by averaging peak-to-peak amplitudes. Background EMG was calculated by assessing the root mean square over 50ms prior to MEP onset in each phase.

In experiment 2, a previously established paired pulse TMS paradigm was used, in which a sub-threshold conditioning pulse over the ipsilateral (left) PMd is followed by a supra-threshold TMS pulse over M1 (25). The coil for the conditioning pulse was placed over a region that has been identified by PET studies as the location of PMd (15, 28) and has previously been used as the point of PMd stimulation (3, 17 and 29). The coordinates for the localization are 8% of the individual distance between nasion and inion (i.e., approx. 3cm) anterior and 1cm medial to the “motor hotspot” for FDI (28; see Fig. 1). The intensity for the conditioning pulse over PMd was set at 90%aMT, the induced current direction was anterior-posterior and the interstimulus interval was 6ms, as previously used to induce idPMI (25). The test stimulus was applied over the motor hot spot. Coil orientation was tangential to the scalp with the handle pointing backwards and laterally at a 45-degree angle away from the midline. Due to spatial interference of the two coils, the conditioning coil was placed directly on the skull, while the test pulse coil over M1 was slightly elevated. The stimulus intensity was then adjusted to induce a test MEP of 1mV. Three separate paired pulse experiments were conducted at rest, during the pre-EMG phase and during the phasic phase. In the rest condition, no acoustic signal was given. In each experiment, 30 MEPs were recorded (15 conditioned and 15 un-conditioned stimuli). MEP size was determined by averaging peak-to-peak amplitudes. Trials with a background EMG of more than 0.02mV (assessed as root mean square) over 50ms before the onset of the MEP were rejected.

Figure 1.

Coil placement. The conditioning coil was placed over left PMd in antero-posterior direction with the handle pointing forward, as described in the previous study (Civardi et al. 2001). The test coil over the left primary motor cortex was positioned perpednicular to the central sulcus. First, the motor hot spot for APB was determined without the second coil on the head. When both coils were used together, the test coil over M1 was slighlty elevated, if needed, and stimulation intensity was adjusted to evoke a test MEP of 1mV in all subjects. A tightly fitting cap was used to mark and control coil placement throughout the experiment.

5. Statistics

MTs of both groups were compared using an independent-samples t-test. Since aMT values were not Gaussian, the non parametric Mann-Whitney test was used. MEP sizes obtained at rest in both groups were compared using an independent-samples t-test for APB and FDI. To assess the presence of surround inhibition (Experiment 1) in each population, the MEP sizes obtained at the different phases were compared, in each group, using a one-way analysis of variance (ANOVA) (factor PHASE, 3 levels: rest, pre-EMG and phasic). If the PHASE effect was significant at the 0.05 level, then post hoc tests were performed, using a Bonferroni correction.

To assess the significance of idPMI (Experiment 2), we used a non-parametric Wilcoxon test to compare the unconditioned MEP size to the conditioned MEP size, within each population, to test for significant differences between conditioned and un-conditioned MEP size, since the data were not Gaussian. idPMI was expressed as ratio between conditioned and unconditioned MEP, in percent (idPMI = (MEPcond / MEPtest)*100[%]). In order to compare these ratios between populations and, our statistical analysis used Conover’s free distribution method, a non-parametric ANOVA based on ranks (29). Two factors were used: GROUP (2 levels: FHd patients, healthy controls) and PHASE (3 levels: rest, pre-EMG, phasic). We first tested for main effects. If such effects were found to be significant at the 0.05 level, we performed a contrast analysis. Lastly, if an interaction was observed, the Mann-Whitney test was applied.

Background EMG activity of the APB muscle was also analysed using a CONOVER test, in order to test for any co-contraction in APB in FHD patients or controls.

Statistical analysis was performed using SPSS software v.15 (SPSS Inc., Chicago, USA). Results are presented as means and standard error of means.

Results

In Experiment 1 looking at surround inhibition in APB, the ANOVA analysis revealed that there was a significant main effect of phase in the control group (F = 4.2, p = 0.026) but not in the FHD group (F = 0.02, p = 0.98). In the control group, post-hoc analyses showed that the MEP sizes during the phasic phase were smaller than at rest (p = 0.04). These results indicate that single pulse TMS demonstrated surround inhibition, reflected by a decrease in MEP size during the phasic phase of the movement, only for the healthy subjects. This effect was not found for the pre-EMG phase (75ms before EMG onset in FDI, see Fig. 2).

Figure 2.

Shown are means and standard errors for MEP size in APB at the three phases that were tested (rest, pre-EMG and phasic) for FHD patients and controls. While there is no difference in MEP size at rest, controls show significant surround inhibition during the phasic phase, which is absent in the FHD patient group.

In FDI, MEP sizes evoked in each phase were not statistically different in the control group (rest 4.5 ± 0.5 mV, pre-EMG 4.5 ± 0.5 mV, phasic 5.1 ± 0.4 mV ; F=0.6, p=0.56), nor in the FHD patient group (rest 4.8 ± 0.4 mV, pre-EMG 5.1 ± 0.3 mV, phasic 5.6 ± 0.4 mV) (F=1.1, p=0.35).

In Experiment 2, idPMI was observed only at rest in the FHD patients (p = 0.017), as indicated by the Wilcoxon test, while there was no significant inhibition for any phase in the control group. The CONOVER analysis showed that there was a significant main effect for GROUP (F = 10.3, p = 0.005), but not for PHASE (F = 0.8, p = 0.44). Moreover, the PHASE-by-GROUP interaction was significant (F = 3.7, p = 0.035). The Mann-Whitney test revealed that FHD patients showed more idPMI at rest compared to controls (Z = −2.72, p = 0.005). No significant differences were observed during the pre-EMG (Z < 0.001, p = 1) or phasic phase (Z = −.832, p = 0.436) phases (see Fig. 3).

Figure 3.

Shown is PICI, expressed as ration between the conditioned and the unconditioned MEP size in percent of the test MEP, as means and standard errors for the three conditions (rest, pre-EMG and phasic) in FHD patients and controls. The only singificant inhibition occurs in the FHD group and at rest indicating that premotor-motor inhibition is stronger in the patients compared to controls for this condition. There was no up-regulation of PICI in the controls group for any condition.

APB MT values were similar in the control group (47.2 ± 1.8%) and in the FHD patient group (47.6 ± 3.6%; p = 0.74). Analogously, aMT values were not different between groups (control group 38.2 ± 2.1%; FHD patients 37.3 ± 2.5%; p = 0.92). MEPs sizes at rest were not different between FHD patients and controls (p = 0.41). Background EMG in APB was not different between PHASEs (F = 17, p = 0.1) or between GROUPs (F = 0.2, p = 0.66) and there was no significant PHASE-by-GROUP interaction (F = 17, p = 0.4).

Discussion

An unanticipated result was that the inhibition from ipsilateral (left) PMd onto left, dystonic M1 was enhanced in patients with FHD at rest. Contrary to the hypothesis, idPMI was not induced in controls, nor did it show an up-regulation during movement initiation, which indicates that it does not make a major contribution to the genesis of surround inhibition in healthy subjects. However, the decrease of idPMI was greater in patients than normals suggesting that it might play a role in the loss of surround inhibition in patients.

Localization of ipsilateral, left PMd

In a previous mapping study, sub-threshold TMS was applied to areas anterior and medial to M1 in order to explore the resulting inhibitory influence (25). The strongest inhibitory effects were reported at an interstimulus interval of 6 ms. This inhibition was obtainable over a quite large area centered approximately 4 – 6 cm anterior from the motor hot spot of FDI, which correlates to left PMd and prefrontal cortex. A second study, in which the extensor carpi radialis muscle (ECR) was the target muscle, found facilitation during dorsoflexion of the wrist, but no inhibition at rest, when TMS was applied 5 cm anterior to the motor hot spot (33). Using a more PMd-specific localization approach for the placement of the conditioning coil (on average 3 cm anterior to the motor hot spot for FDI), as has been described before in FHD patients and healthy controls (3, 19 and 29) and the 6 ms interstimulus interval, we were not able to induce significant inhibition in the control group in the current study. In contrast, FHD patients did show significant inhibition at rest. This indicates that there was strong inhibition from PMd to M1 in the FHD patients compared to healthy subjects and this might correlate with the previously described over-activity in this area (15 – 17). While we were anatomically interested in PMd and stimulated 3cm anterior to the motor hot spot, a TMS position that has been used in previous studies (3), our findings do not contradict previous findings about more anterior, eventually prefrontal stimulation sites (25).

Mechanism of idPMI

From a functional point of view, the exact mechanism of idPMI is not clear. Despite of an average distance between PMd and M1 of about 3 cm, the optimal interstimulus interval is 6 ms (25) which may reflect the transmission via two or more synapses. It would be possible that, as has been shown for inter-hemispheric inhibition (31, 32) the inhibitory effect is generated via projections onto the local inhibitory network of interneurons in M1. Further studies to better describe the underlying mechanism will be needed.

Since there is a variety of studies demonstrating increased motor cortical excitability and deficiencies in intra-cortical inhibition in FHD (12, 14, 21 and 24), at first glance, the current findings of increased inhibition from PMd do not seem to fit into this pattern. From imaging studies (15–17) and electrophysiological reports (14, 19), there is increasing evidence that PMd is in fact involved in the pathophysiology of FHD. Previously, it has also been demonstrated that low-frequency rTMS can improve symptoms in FHD (19). This intervention, which is known to decrease excitability if given to M1 (34), may help to enhance or restore intra-cortical inhibition from PMd to M1 and thereby prevent motor-overflow. A recent TMS study assessing the transcallosal interaction of left PMd and right M1 found reduced inhibition from left PMd in FHD patients compared to controls that was similar to the effect induced by high-frequency rTMS treatment in healthy controls (14). The increased inhibition described in the current study may be compensatory in order to prevent motor overflow and other dystonic symptoms, for example dystonic posturing, at rest. It is most likely that the observed differences between the FHD patients and controls are due to changes in PMd and not M1, as supported by similar test MEP sizes in both groups, similar resting and active motor threshold and similar MEP sizes at rest, when single pulse TMS was given to M1 only reflecting a similar level of M1 excitability. In this study, we excluded FHD patients with spontaneous posturing in order to ensure that APB was at rest and not in tonic pre-activation during the assessment. Therefore, this study does not allow identifying the effect of posturing.

Beyond alterations in cortical interactions that can be assessed in healthy volunteers, the current findings do not exclude the possibility that abnormal circuits are present in patients with FHD and responsible for the increase in inhibition observed.

Contribution of idPMI to surround inhibition

The second question we addressed in this study was whether idPMI contributes to the genesis of surround inhibition. In this case, we would have expected an up-regulation of idPMI synchronous to the occurrence of surround inhibition during movement initiation (24). To our knowledge, there has not been a study confirming ipsilateral idPMI in healthy volunteers, nor assessing ipsilateral idPMI during motor activation. The current results do not show any idPMI in the control group, neither at rest, nor before or during movement initiation, which may indicate that PMd is not the main generator of surround inhibition.

In conclusion, this study confirms that inhibition is abnormal in FHD patients. The results showed increased idPMI in FHD patients selectively at rest, while it is abolished before and during movement initiation. The findings suggest that increased idPMI may be compensatory to avoid dystonic posturing at rest. The finding that idPMI was not enhanced during movement initiation in healthy volunteers suggests that it is not the main source of surround inhibition in normal subjects. However, the loss of idPMI in patients with FHD during motor activation may contribute to their loss of surround inhibition. The current study provides preliminary evidence for abnormal premotor-motor interaction in patients with FHDs. We cannot exclude the fact that different intensities of the conditioning pulse may exert different effects on M1 excitability and provide more information on the role played by PMI on surround inhibition. Future studies will be needed to better identify the role of the over-activity of PMd and the underlying mechanisms of idPMI.