S.M. and H.R. contributed equally to this study.
Plasticity of cortical inhibition in dystonia is impaired after motor learning and paired-associative stimulation
Article first published online: 19 MAR 2012
© 2012 The Authors. European Journal of Neuroscience © 2012 Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience
Volume 35, Issue 6, pages 975–986, March 2012
How to Cite
Meunier, S., Russmann, H., Shamim, E., Lamy, J.-C. and Hallett, M. (2012), Plasticity of cortical inhibition in dystonia is impaired after motor learning and paired-associative stimulation. European Journal of Neuroscience, 35: 975–986. doi: 10.1111/j.1460-9568.2012.08034.x
- Issue published online: 19 MAR 2012
- Article first published online: 19 MAR 2012
- Received 20 October 2011, revised 26 December 2011, accepted 13 January 2012
- motor learning;
- transcranial magnetic stimulation
Artificial induction of plasticity by paired associative stimulation (PAS) in healthy volunteers (HV) demonstrates Hebbian-like plasticity in selected inhibitory networks as well as excitatory networks. In a group of 17 patients with focal hand dystonia and a group of 19 HV, we evaluated how PAS and the learning of a simple motor task influence the circuits supporting long-interval intracortical inhibition (LICI, reflecting activity of GABAB interneurons) and long-latency afferent inhibition (LAI, reflecting activity of somatosensory inputs to the motor cortex). In HV, PAS and motor learning induced long-term potentiation (LTP)-like plasticity of excitatory networks and a lasting decrease of LAI and LICI in the motor representation of the targeted or trained muscle. The better the motor performance, the larger was the decrease of LAI. Although motor performance in the patient group was similar to that of the control group, LAI did not decrease during the motor learning as it did in the control group. In contrast, LICI was normally modulated. In patients the results after PAS did not match those obtained after motor learning: LAI was paradoxically increased and LICI did not exhibit any change. In the normal situation, decreased excitability in inhibitory circuits after induction of LTP-like plasticity may help to shape the cortical maps according to the new sensorimotor task. In patients, the abnormal or absent modulation of afferent and intracortical long-interval inhibition might indicate maladaptive plasticity that possibly contributes to the difficulty that they have to learn a new sensorimotor task.
Decrease of inhibition and maladaptive brain plasticity are dominant traits in the pathophysiology of dystonia (Hallett, 2006). Aberrant plasticity was suggested by the enhanced response of the patients to plasticity-inducing protocols such as paired associative stimulation (PAS) (Quartarone et al., 2003), theta-burst stimulation (Edwards et al., 2006) and 5-Hz repetitive transcranial magnetic stimulation (Gilio et al., 2007). This is clinically apparent in patients with task-specific focal hand dystonia where repetitive activity is certainly a trigger for the development of dystonia. Enhanced plasticity is not the consequence of the abnormal movements, as it does not exist in psychogenic dystonia (Quartarone et al., 2009). Inhibition is deficient in dystonia at several levels of the central nervous system. Short-interval intracortical inhibition (SICI) was less in both hemispheres of patients with focal dystonia (Ridding et al., 1995; Edwards et al., 2003). The contribution of SICI to surround inhibition is selectively reduced in movement initiation (Beck et al., 2008). At rest, healthy subjects and dystonic patients have similar amounts of long-interval intracortical inhibition (LICI), believed to involve GABAB receptor neurotransmission (Florian et al., 2008), while LICI is reduced during movement in patients (Chen et al., 1997). Long-latency afferent inhibition (LAI) reflects an aspect of somatosensory-motor processing, but the transmitters and pathways involved are unknown. LAI is reduced at rest in task-specific dystonia (Abbruzzese et al., 2001).
To what extent abnormal inhibition participates in enhanced plasticity and/or in abnormal sensorimotor adaptation or motor learning (Ghilardi et al., 2003) in dystonia is unknown.
Using the PAS technique, we showed (Russmann et al., 2009), in healthy volunteers (HV), that the development of long-term potentiation (LTP)-like plasticity in the M1 cortex is accompanied by a lasting decrease of LICI and LAI. Changes of SICI were less robust, in agreement with previous results (Stefan et al., 2002; Quartarone et al., 2003; Rosenkranz & Rothwell, 2006). According to the interaction between the effects of PAS and motor activity performed before, during or after the PAS, it is thought that PAS and motor practice-induced plasticity develop in a shared set of synapses or, at least, involve circuits interacting with each other (Stefan et al., 2006; Rosenkranz et al., 2007). Effects of motor learning on LICI and LAI have not been studied in humans so far, although it has been shown that motor practice affects SICI (Liepert et al., 1998; Nordstrom & Butler, 2002; Perez et al., 2004; Rosenkranz & Rothwell, 2006; Rosenkranz et al., 2007).
We conducted two sets of experiments to answer: (i) whether the ‘physiological’ plasticity that develops during motor learning is accompanied by similar plastic changes in inhibitory circuits as those observed after the ‘artificial’ process of PAS, and (ii) whether plasticity of inhibitory circuits is similar, in extent, direction and type of circuit involved, in patients with task-specific dystonia and in HV. We measured LICI and LAI before and after a PAS intervention and during and after the learning of a simple motor task involving the same muscles like PAS (Muellbacher et al., 2001, 2002).
Subjects and methods
Seventeen patients (mean age 51.3 ± 2.2 years) with an occupational dystonia [nine writer’s cramp (WC), five musician’s dystonia (MD) and three with both WC and MD] were enrolled in the study. They were recruited through the Movement Disorders Outpatient Clinic of the NINDS. Clinical characteristics of the patients are presented in Table 1. A subset of subjects participated in each experiment (Table 1).
|Patient no.||Dystonia||Age (years)||Dominant Hand/affected hand||Sex||Duration of disease (years)||Dystonic pattern*||Experiment†|
|1||WC||56||R/R||M||15||Wrist flexion, finger extension (2–4), ulnar deviation||1,2,3,4|
|2||MD + WC (guitar)||72||R/R||M||16||Wrist flexion, ulnar deviation, with tremor||1|
|3||MC (guitar)||36||R/R||M||3||Wrist flexion, ulnar deviation||1.4|
|4||WC||49||R/R||M||6||Fingers extension, wrist flexion, ulnar deviation||1,3,4|
|5||WC||58||R/R||M||6||Finger extension (2–4)||1,3,4|
|6||WC + MD (Banjo)||54||R/R||M||WC: 8 MD: 7||Wrist extension, finger flexion (1–3)||1,2,3,4|
|8||WC||53||R/R+L||F||22||Wrist flexion, ulnar deviation||1.4|
|9||WC||42||R/R+L||M||Right hand: 19 left: 2||Wrist extension, ulnar deviation||1.4|
|10||MD + WC||40||R/R||M||5||Wrist extension, finger flexion (5)||1,2,3|
|11||MC (guitar and Bass)||47||R/R||M||3||Thumb and wrist extension, finger flexion (2–3)||1,3,4|
|12||WC||63||R/R||M||Wrist flexion, fingers flexion||1.3|
|13||WC||40||R/R||M||18||Thumb shaking while writing||1|
|14||WC||58||R/R||M||5||Wrist extension, finger flexion (4)||2|
|15||MD (bagpipe)||44||L/L||M||4||Finger flexion (4)||2|
|16||MD (piano)||45||L/R||M||14||Finger flexion (4–5)||2|
|17||MD (piano)||55||R/R||M||2||Finger flexion (4)||2|
Their results were compared with those of 19 HV (mean age 46.3 ± 3.1 years) with no history of either neurological or psychiatric disease and a normal neurological examination. Part of the results obtained in HV has been presented in a previous paper (Russmann et al., 2009).
The experimental protocol was approved by the NINDS Institutional Review Board, and all subjects gave written informed consent.
Surface electromyographic (EMG) activity (band-pass 10 Hz to 2 kHz) was recorded from the flexor pollicis brevis (FPB, the target muscle) and the abductor digiti minimi (ADM, an uninvolved muscle), in bipolar belly-tendon arrangements, using a Nicolet Viking electromyograph (Skovlunde, Denmark). The data acquisition system was built with the Labview graphical programming language (sampling rate 5 kHz). During recordings, EMG activity was continuously monitored with visual and auditory feedback to ensure complete relaxation.
Transcranial magnetic stimulation (TMS)
Subjects were seated in a comfortable reclining chair. A figure-of-eight-shaped coil (7 cm inner diameter for each half) connected to a Bistim-module and two Magstim 200 magnetic stimulators (The Magstim Company, Dyfed, UK) was positioned on the scalp over M1. The hot spot for the FPB muscle was defined as the lowest threshold site evoking a motor-evoked potential (MEP) response in FPB accompanied by a clear thumb flexion movement with no palpable contraction in ADM and an ADM MEP smaller than the FPB MEP.
The coil was positioned with the handle pointing backwards at an angle of 45° to the midline (Brasil- Neto et al., 1992).
The right hand was tested in HV and the affected or more affected hand in patients. This was the right hand in all except one patient (Table 1; no. 15).
Resting motor threshold (rMT)
Long-interval intracortical inhibition
To evoke LAI, a conditioning stimulation was applied to the median nerve 150 ms before a test TMS stimulation adjusted to 1.2 × rMT (Devanne et al., 2009). The median nerve was stimulated at the wrist through bipolar surface electrodes (rectangular pulses of 0.2-ms duration) using a Grass S88 stimulator (Grass Instruments Co., Quincy, MA, USA). Intensity was adjusted to 2.5 × the perceptual threshold (PT), which was always below the thumb twitch threshold.
Intensity curve of SICI
To evoke SICI, a subthreshold conditioning TMS stimulation (CS2.5) was delivered 2.5 ms before a test TMS stimulation (Fisher et al., 2002). The intensity curve for SICI was created by using CS2.5 intensities 0.5, 0.6 and 0.7 × rMT (Orth et al., 2003; Stinear & Byblow, 2004), while keeping the test TMS stimulation at 1.2 × rMT. For LICI, LAI and SICI measurements, the intensity of the test stimulation was adjusted, if necessary, to evoke a test MEP of similar size at baseline and after PAS or motor practice.
We conducted two different sets of experiments (Fig. 1).
Set 1: PAS-induced plasticity in inhibitory circuits of M1
Methods have been described in detail in Russmann et al. (2009) and will be only summarized here. Three different experiments were performed.
Experiment 1: assessment of overall cortical excitability and of plasticity of LICI and LAI circuits after PAS25/0.5 mV in HV, compared with patients with dystonia
Thirteen patients and 13 HV were enrolled (HV: 47.2 ± 3 years, patients: 51.5 ± 8.4; P = 0.3). The size of a control test MEP, rMT and measures of cortical inhibition (LAI and LICI) were compared at baseline (T0) and 5–10 min (T1) after the end of a PAS intervention. Measures of inhibition were repeated 30–40 min (T2) after the end of the PAS.
PAS aimed to induce a LTP-like plasticity in the M1 cortex and to promote the development of plasticity in the inhibitory circuits. Accordingly, we used a 25-ms interval (see below) (Wolters et al., 2003) and low TMS intensities (McAllister et al., 2009; Russmann et al., 2009) (MEPs evoked during PAS were around 0.5 mV). This intervention will be referred as PAS25/0.5 mV in the following.
Experiment 2/control experiment 1 (CE1): assessment of plasticity of SICI circuits after PAS25/0.5 mV in patients with dystonia
Seven patients were enrolled (mean age: 50.3 ± 6.2 years). Three patients had participated in experiment 1. The intensity curve of the SICI was compared at baseline and at 5–10 and 30–40 min after the same PAS25/0.5 mV intervention as in experiment 1.
Experiment 3/Control experiment 2 (CE2): assessment of overall corticospinal excitability and of plasticity of LICI and LAI circuits after ‘high dose’ PAS25/1 mV in HV compared with patients with dystonia
Seven patients (52.4 ± 6.0 years) and seven HV (40.3 ± 7.3 years; P = 0.1) participated. All except one patient had participated in experiment 1. RMT, size of a control test MEP, and measures of LICI and LAI were compared between baseline and 5–10 and 30–40 min after the end of a PAS intervention. Parameters of stimulation for PAS were the same as those used in experiments 1 and 2 except for the TMS intensity, which was higher. MEPs evoked during PAS were around 1 mV; this intervention will be referred as PAS25/1 mV.
PAS intervention method
PAS was achieved by pairing an electrical stimulus to the median nerve at the wrist (2.5 × PT, always below thumb twitch threshold) and a single TMS pulse targeting the FPB muscle. The interstimulus interval (ISI) between median nerve and TMS stimulation was set at 25 ms. This ISI has been shown to be optimal for inducing a sustained increase in motor cortex excitability (Stefan et al., 2000; Wolters et al., 2003). Two hundred and forty pairs of stimuli were delivered at 0.2 Hz over 20 min (Muller et al., 2007). We asked the subjects to count the stimulations in order to maintain attention. In experiment 1 and control experiment 1, TMS intensity was adjusted to evoke an MEP with a peak-to-peak amplitude about 0.5 mV in the relaxed FPB muscle (PAS25/0.5 mV). In control experiment 2, TMS intensity was adjusted to evoke a larger MEP with a peak-to-peak amplitude of 1–1.5 mV (PAS25/1 mV).
To assess the effect of the PAS interventions on overall corticospinal excitability we compared the size of a control MEP (mean size of ten MEPs) before and after PAS using the same stimulus intensity that produced a test MEP before PAS adjusted to 0.8–1 mV.
Set 2/Experiment 4: plasticity in inhibitory circuits of M1 during motor learning
Nine patients (50.5 ± 6.2 years) and 11 HV (44.8 ± 1.9 years, P = 0.15) participated. They had all participated in experiment 1 of the first set previously, and there was a minimum of 15 days between the studies. The size of a control test MEP, rMT and measures of LAI and LICI were compared at baseline (T0) and after each of three blocks of motor practice (L1, L2, L3).
Motor learning task
We used the motor task studied by Muellbacher et al. (2001, 2002). We chose this task as not susceptible to trigger dystonic movements in the patients in order to compare the physiological data between patients and HV at equal performance. It implies that the potential differences observed between the HV and the patients would not be linked to the phenotype yet may represent an endophenotype or a compensatory mechanism allowing the patients to maintain a normal performance despite their pathology.
Subjects were seated in front of a monitor with their supinated hand resting on an armrest. The practice task was a metronome-paced (0.5 Hz) ballistic pinch between the index finger and the thumb of the dominant hand. The subjects had to perform the pinch at or above a given acceleration (980 cm/s2). A green or a red light flashed after the end of each pinch according to whether the task was correctly (> 980 cm/s2) or incorrectly (< 980 cm/s 2) performed. There were three blocks of practice including 150 movements for the first two blocks and 300 for the third one. Acceleration of each pinch was recorded with an accelerometer (Endevco Corp., CA, USA) fixed to the dorsal aspect of the index finger. The first peak acceleration of each movement was analysed online and used for delivering the signal feedback and also stored for offline analysis (Labview Graphical programming language, National Instruments). Performance was assessed by measuring (i) the mean percentage of successful pinches (acceleration > 980 cm/s2) and (ii) and the mean peak acceleration (expressed in gravities, g) during the first, second and third block of practice.
To compare PAS or learning-induced changes of the 1-mV control MEP between HV and patients, MEP sizes at T1 (PAS experiment) or after L3 (learning experiment) were normalized to their baseline values. A factorial anova was then performed with ‘group’ (HV versus patients) and ‘muscle’ (FPB versus ADM) as between-subject variables.
SICI, LICI and LAI were expressed as percentages of the test MEP. Baseline levels of SICI, LAI, LICI and the mean percentage of successful pinches and mean peak acceleration were compared between HV and patients using factorial anova. Effects of PAS and learning on LAI, LICI and SICI, effects of learning on the mean percentage of successful pinches and the mean peak acceleration were first analysed in each group (HV and patients) separately (indeed, the difference in baseline level of the inhibitions may render interpretation difficult if both groups were entered in the same analysis) by using a repeated anova (ranova), with the values of inhibition at T0, T1 and T2 or the values of behavioural parameters at T0, L1, L2 and L3 forming the repeats (‘time’). Post-hoc analysis was done using a Fisher test to compare values two by two.
In a second step to compare the patient group and the control group, amounts of LAI and LICI, percentage of successful pinches and mean peak acceleration were normalized to their baseline value. An ranova was then computed with ‘group’ as between-subject variable and ‘time’ as the repeat. To study PAS-induced changes of SICI, an ranova was computed with the values of SICI at 0.5, 0.6 and 0.7 × RMT forming the repeats. For Post-hoc tests we used Fisher’s test. All results are presented as mean ± SEM.
As pathophysiology may differ between writer’s cramp and musician’s dystonia (Rosenkranz et al., 2005) a further statistical analysis was done after stratification of the patients into two groups: writer’s cramp (WC) and musician’s dystonia (MD). Patients with both WC and MD were classified to the MD group as their motor disturbance appeared first when playing music. This was done only for experiment 1, set 1 (WC: n = 8, MD: n = 5) as size of the samples was too small to get homogenous groups in the other experiments.
Pas-induced plasticity in inhibitory circuits
After PAS25/0.5 mV the rMT was not modified in either HV (rMTT0 = 46.6 ± 0.9% of maximum stimulator output vs. rMTT1 = 45.9 ± 0.9%) or in patients (rMTT0 = 46.2 ± 1.7% vs. rMTT1 = 45.1 ± 1.6%).
Effect of PAS25/0.5 mV on overall corticospinal excitability. In HV, a control FPB test MEP (MEPT0 = 1.0 ± 0.1 mV) was significantly enhanced after PAS25/0.5 mV (MEPT1 = 1.7 ± 0.2 mV) (ranova: ‘time’P < 0.004). The corresponding ADM MEP was enhanced only slightly (MEPT0 = 0.7 ± 0.1 mV vs. MEPT1 = 0.9 ± 0.2 mV at T1) (ranova: ‘time’P = 0.9). In contrast, PAS25/0.5 mV did not cause any change of control MEPs in patients either in FPB (MEPT0 = 1.0 ± 0.1 mV vs. MEPT1 = 1.0 ± 0.1 mV) (ranova: ‘time’P = 0.2) or ADM (MEPT0 = 0.4 ± 0.1 mV vs. MEPT1 = 0.4 ± 0.1) (ranova: ‘time’P = 0.9). This was confirmed by the statistical analysis done using normalized values of MEPs. It showed a significant effect of group for the PAS-induced change of control FPB MEP size (anova‘group’: P < 0.03, ‘muscle’: P n.s. (not significant), ‘muscle’ × ‘group’: P: n.s.).
Effects of PAS25/0.5 mV on LICI. In HV, (Fig. 2A) LICI decreased progressively in FPB after PAS25/0.5 mV (ranova: ‘time’P < 0.01, see results of Fisher test on Fig. 2A). No significant effect was observed in ADM (P = 0.4). Conversely, in patients, PAS25/0.5 mV did not significantly modify LICI in either FPB (ranova: ‘time’P = 0.2) or ADM (ranova: ‘time’P = 0.2) (Fig. 2A).
The amount of LICI at T0 (Fig. 2A, white bars) was not different in HV (FPB: LICIT0 = –60.1 ± 6.2% of test MEP, ADM: LICIT0 = –34.2 + 13.1%) and patients (FPB: LICIT0 = –46.7 ± 3.6%, ADM: LICIT0 = –6.6 ± 10.5%) (anova: FPB: ‘group’ = 0.09; ADM: ‘group’P = 0.11).
The between-groups comparison was done using the normalized values of LICI. In FPB muscle, ranova confirmed that PAS-induced modulation of LICI was lost in patients (‘group’P < 0.04, no effect of ‘time’P = 0.4, and no interaction). In ADM, ranova did not find any significant effect (‘group’P = 0.3, ‘time’P = 0.7, no interaction). No correlation was found between PAS-induced changes of the control test 1-mV MEP and those of LICI.
Effects of PAS25/0.5 mV on LAI. The same(Fig. 2B) 13 HV and patients, except one, who were tested for LICI, were also tested for LAI. In the HV group, LAI was decreased immediately after PAS25/0.5 mV and stayed decreased at T2 in both the FPB (ranova: ‘time’P < 0.02,) and the ADM (ranova: ‘time’P < 0.006, see results of Fisher tests on Fig. 2B). In the patient group, modulation induced by PAS25/0.5 mV was in the opposite direction as LAI increased after the intervention (FPB: ranova: ‘time’P < 0.001; ADM: ‘time’P = 0.1). The amount of LAI for FPB at T0 was significantly less in patients (FPB: LAIT0 = −14.5 ± 10.2%) than in HV (LAIT0 = −40.2 ± 6.0%, t-test: P < 0.04). A similar trend was observed for ADM but did not reach statistical significance (patients: LAIT0 = −10.9 ± 12.4%; HV: LAIT0 = −31.4 ± 6.6%) (P = 0.2) (Fig. 2B, white bars). To eliminate the confounding role of such a significant difference in baseline in the comparison of PAS-induced effects between the two groups, the comparison was done on sub-groups having similar mean baseline LAI. The three HV with the largest LAI and the three patients with the smallest LAI at T0 were discarded from analysis. In the remaining nine patients, LAIT0 = −30.6 ± 7.0% in FBP and −26.0 ± 8.7% in ADM while, in the ten remaining HV, LAIT0 = −32.6 ± 6.3% in FPB (P = 0.8) and −25.1 ± 6.3% in ADM (P = 0.9).
PAS-induced changes of LAI in FPB were then compared between these subgroups by a ranova that confirmed a significant ‘group’ effect (P < 0.01) (‘time’P = 0.4, no interaction). In ADM similar PAS-induced changes were observed (ranova, ‘group’P < 0.02, ‘time’P = 0.2, no interaction). There was no correlation between PAS-induced changes of the control 1-mV MEP and those of LAI.
Comparison of the writer’s cramp and musician’s dystonia patients. Neither the MD (n = 5) nor WC patients (n = 8) showed any PAS25/0.5 mV -induced modulation of the FPB MEPs (WC: MEPT0 = 1.0 ± 0.1, MEPT1 = 1.0 ± 0.1; MD: MEPT0 = 0.9 ± 0.1, MEPT1 = 1.0 ± 0.1) or ADM MEPs (WC: MEPT0 = 0.4 ± 0.1, MEPT1 = 0.4 ± 0.1; MD: MEPT0 = 0.4 ± 0.2, MEPT1 = 0.3 ± 0.1).
Baseline LICI was not different between WC and MD patients (WC: LICIT0 = −46.3 ± 4.8, MD: LICIT0 = −49.9 ± 5.8, P = 0.4). In the two dystonic groups PAS25/0.5 mV-induced modulation of LICI was different from that observed in the HV group. Additionally, LICI modulation differed in WC patients compared with MD patients: FPB LICI was slightly decreased after PAS in WC (‘time’P < 0.05), while FPB LICI was increased in MD (‘time’P = 0.1) (ranova: MD vs. WC P < 0.04, ‘time’P = 0.4, ‘time’ × ‘group’ interaction P < 0.02). The same effect, even more marked, was observed for the ADM muscle (ranova: MD vs. WC P < 0.05, ‘time’P = 0.4, no interaction).
Baseline LAI was decreased in the two dystonic groups (WC: LAIT0 = −20.5 ± 14.3, MD: LAIT0 = −3.9 ± 13.3) compared with the HV group (LAIT0 = −44.0 ± 5.1) (anova: ‘group’P < 0.05, Fisher test: MC vs. HV P < 0.02, WC vs. HV P = 0.09, MC vs. WC P = 0.3). In both. dystonic groups PAS25/0.5 mV induced an increase of LAI (ranova: MD vs. WC P = 0.8, ‘time’P = 0.9, no interaction).
Experiment 2/control experiment 1
Effect of PAS25/0.5 mV on SICI. SICI was constantly evoked in patients for CS of 0.6 and 0.7 × rMT. PAS25/0.5 mV did not induce any significant change of SICI (CS = 0.5 × rMT: SICI = 101 ± 7%; CS = 0.6 × rMT: SICI = 70 ± 14%; CS = 0.7 × rMT: SICI = 40 ± 4) at any of the three intensities tested (ranova: ‘intensity’P < 0.0001, ‘time’P = 0.3, no interaction).
Experiment 3/control experiment 2
Contrasting with previous reports (Quartarone et al., 2003; Weise et al., 2006) our PAS intervention did not induce any change of the control 1-mV FPB test MEP in the patient group. As we used a smaller intensity for the TMS stimulation during the PAS than that used in the previous studies, we wondered whether it could explain the lack of effect in patients. In subgroups of seven patients and seven HV we tested the effect of a PAS intervention with similar methods as the PAS25/0.5 mV except a higher TMS intensity (PAS25/1 mV). FPB MEP size at T0 was not different in HV and patients (MEPT0 = 0.9 ± 0.1 and 1.1 ± 0.2 mV, respectively). It was enhanced at T1 and T2 in the HV group (MEPT1 = 1.6 ± 0.4 mV; MEPT2 = 1.9 ± 0.4 mV) (ranova: ‘time’P < 0.01; Fisher test T0 vs. T2: P < 0.04) and to the same extent or even more in the patient group (MEPT1 = 1.9 ± 0.2 mV; MEPT2 = 2.4 ± 0.3 mV) (ranova: ‘time’P < 0.0001, T0 vs. T1 P < 0.002, T0 vs. T2 P < 0.0001). PAS-induced changes in ADM were not significant yet we observed a trend for the ADM MEP to be facilitated after PAS in the patient group (MEPT0 = 0.4 ± 0.1, MEPT1 = 0.9 ± 0.3, MEPT2 = 0.9 ± 0.2 mV) (ranova: ‘time’P = 0.1) and not in the HV group (HV: MEPT0 = 0.5 ± 0.2, MEPT1 = 0.6 ± 0.2, MEPT2 = 0.6 ± 0.2 mV; ranova: ‘time’P = 0.3). When both groups were compared using normalized values of MEPs no differences were significant.
Effects of PAS25/1 mV on LICI. There was no significant PAS-induced changes of LICI either in the control group (LICIT0 = −38.2 ± 7.8%, LICIT1 = −54.0 ± 7.2%, LICIT2 = −40.0 ± 11.1%) (ranova: ‘time’P = 0.7) or in the patient group (LICIT0 = −40.7 ± 5.8%, LICIT1 = −35.7 ± 11.8%, LICIT2 = −47.7 ± 12.3%) (ranova: ‘time’P = 0.5). As baseline LICI was no different between the two groups (P = 0.4), the normalized values of LICI were compared using an ranova (‘time’P = 0.7, ‘group’P = 0.8, no interaction) that confirmed that both groups behave similarly.
Effects of PAS25/1 mV on LAI. There was no significant PAS-induced changes of LAI either in the control group (LAIT0 = −37.3 ± 10.3%, LAIT1 = −30.7 ± 6.3%, LAIT2 = −35.4 ± 9.0%; ranova: ‘time’P = 0.8) or in the patient group (LAIT0 = −12.1 ± 13.3%, LAIT1 = −23.2 ± 30.0%, LAIT2 = −16.6 ± 12.2%; ranova: ‘time’P = 0.9).
Learning-induced plasticity in inhibitory circuits
Performance. At baseline, (Fig. 3A and B) HV and patients had similar performance (number of correct movements: 47 ± 8% in HV and 54 ± 9% in patients, P = 0.5). Motor performance improved between the first and second block of practice and remained stable between the second and third block in both groups (ranova; HV: ‘time’P < 0.03; patients: ‘time’P < 0.05, see results of the Post-hoc tests in Fig. 3A). Statistical analysis on the normalized values confirmed that patients performed as fast and as well as HV (‘time’P: not significant, ‘group’P: not significant, no interaction). Mean peak acceleration (MPA) (Fig. 3B) was not different at baseline between the two groups (HV: MPA = 0.38 ± 0.05; patients: MPA = 0.36 ± 0.07, P = 0.7) and slightly increased after motor practice (ranova; HV: ‘time’P = 0.08; patients: ‘time’P < 0.05). After normalization there was no difference between the two groups.
Effect of learning on overall corticospinal excitability. After the three blocks of motor practice the rMT was not modified in either HV (rMTT0 = 45.0 ± 1.8% of maximum stimulator output vs. rMTL3 = 44.0 ± 1.6%) or patients (rMTT0 = 47.8 ± 2.5% vs. rMTL3 = 47.9 ± 2.4%).
In HV, motor practice led to an enhancement of the FPB MEP (MEPT0 = 0.8 ± 0.1 mV, MEPL3 = 1.6 ± 0.2 mV; ranova: ‘time’P < 0.004). The corresponding ADM MEP was also enhanced (MEPT0 = 0.4 ± 0.1 mV, MEPL3 = 0.7 ± 0.1 mV; ranova: ‘time’P = 0 .2) but not significantly. In patients, FPB and ADM MEPs were marginally increased (FPB: MEPT0 = 1.0 ± 0.1 mV, MEPL3 = 1.1 ± 0.3 mV; ranova: ‘time’P = 0.6; ADM: MEPT0 = 0.4 ± 0.1 mV, MEPL3 = 0.5 ± 0.3 mV; ranova: ‘time’P = 0.4). This was confirmed by the statistical analysis on normalized values (MEPL3/MEPT0), which showed a significant effect of group for the learning-induced change of FPB MEP size (anova: FPB ‘group’: P < 0.05; ADM ‘group’P = 0.3).
Effects of learning on LICI. In HV, (Fig. 4A) learning the motor task led to a decrease of LICI in the FBP engaged in the task (Fig. 4A) (ranova: ‘time’P < 0.01, see results of Post-hoc tests on Fig. 4A). The same was observed in the patients (ranova: ‘time’P < 0.03, see results of Post-hoc tests on Fig. 4A). As shown in Fig. 4A, the decrease continued after the task was already learned. Indeed, subjects had peaked in learning after the first block of practice (Fig. 3A and B) while LICI decrease was maximal after the third block of practice. In ADM, there was also a trend for LICI to decrease during learning in HV yet it did not reach statistical significance (ranova: HV: ‘time’P = 0.1; patients: ‘time’P = 0.8).
After normalization of LICI values (LICIL1/LICIT0, LICIL2/LICIT0, LICIL3/LICIT0), ranova confirmed that LICI decreased to the same extent in HV and patients (FPB: ranova: ‘time’P = 0.3, ‘group’P = 0.7, no interaction). We did not find any correlation between the behavioural parameters and the learning-induced changes of LICI.
Effects of learning on LAI. In HV, (Fig. 4B) LAI decreased after the first block of practice, and continued to decrease slightly after the second and third blocks. It occurred in both muscles but was statistically significant only in FPB (FPB: ranova: ‘time’P < 0.004; ADM: ‘time’P = 0.09, see results of Post-hoc tests on Fig. 4B). Dystonic patients behaved differently, as LAI in FPB or in ADM were not modified by motor learning (ranova: FPB: ‘time’P = 0.6, ADM: ‘time’P = 0.4) (Fig. 4B).
After normalization, the ranova comparing variations of LAI (LAIL1/LAIT0, LAIL2/LAIT0, LAIL3/LAIT0) between HV and patients confirmed that LAI decreased more in HV than in patients (ranova: FPB: ‘time’P = 0.4, ‘group’P < 0.03, no interaction).
In HV, the larger the decrease of LAI, the better was the performance in the following or preceding motor practice. Indeed, there was a positive correlation between the increase of successful pinches between the first (L1) and the second (L2) motor practice and the decrease of LAI after the first [(LAIL1/LAIT0) × 100, R2 = 0.7, P < 0.002; Fig. 5A] or the second block of practice [(LAIL2/LAIT0) × 100, R2 = 0.6, P < 0.006; Fig. 5B]. The correlation was still present yet weaker between the performance improvement after the third block of practice (L3 vs. L1) and the decrease of LAI between the first and third block of practice [(LAIL3/LAIT0) × 100, R2 = 0.4, P < 0.04; Fig. 5C]. Such correlations were specific to the trained muscle (here the FPB) as they did not exist in ADM (Fig. 5G–I). Such correlations did not exist in patients for FPB (Fig. 5D–F) or ADM (Fig. 5J–L). The increase of the mean peak acceleration did not correlate with the changes in LAI.
There are two important findings of this study. (i) In healthy subjects the learning of a simple motor task involving the FPB muscle was accompanied, in the motor representation of the FPB, by a long-lasting decrease of excitability of inhibitory circuits supporting LAI and LICI. Learning-induced changes of LAI correlated with performance improvement. These LAI and LICI changes were in the same direction as those observed after a PAS intervention targeting the FPB muscle. (ii) In dystonic patients, despite motor performance similar to that of controls, there was no learning-induced modulation of LAI while LICI was normally decreased. After PAS, plasticity was impaired in both LAI and LICI circuits as there was a paradoxical increase of LAI and no decrease of LICI.
Mechanism of PAS or learning-induced changes of LAI and LICI
In a previous paper (Russmann et al., 2009) we showed that PAS-induced changes of LAI and LICI are probably due to the development of a bi-directional associative long-term depression (LTD)-like plasticity in inhibitory pathways. Based on the facts (i) that PAS shares physiological properties of LTP/LTD (Stefan et al., 2000), (ii) that LTP is a mechanism of motor memory (Hess & Donoghue, 1996; Rioult-Pedotti et al., 1998, 2000) and (iii) that the same subset of neurons are involved in after-effects of PAS and voluntary motor activity/learning (Kujirai et al., 2006; Stefan et al., 2006; Rosenkranz et al., 2007; Kennedy & Carson, 2008), we speculated that learning and PAS-induced changes of LAI and LICI rely on similar mechanisms.
Plastic changes in circuits of LAI and LICI during motor learning in healthy subjects
This is the first time, to our knowledge, that changes of LAI and LICI have been studied after motor learning in humans. A decrease of LAI correlated with performance improvement: the larger the decrease of LAI after each block of practice, the better the performance improvement. This result indicates that the role of somatosensory processing in shaping the motor output is prominent at the early stages of motor learning and diminishes in proportion as the task is learned. Circuits supporting LAI are not precisely known, but may involve cortical (secondary somatosensory, premotor and posterior parietal cortices) or sub-cortical (basal ganglia, cerebellum) sensory integration areas (Sailer et al., 2003). In this regard, imaging studies have confirmed the involvement of associative cortical (prefrontal and parietal) (Tracy et al., 2001) and subcortical (basal ganglia) (Lehericy et al., 2005) regions in the early stage, the first 5–10 min, of performance.
There was no correlation between the degree of decrease of LICI and the improvement of performance. This is not unexpected given the complexity of the relationship between plasticity of GABA synapses and development of LTP (Stelzer & Shi, 1994; Mendoza et al., 2006).
Inhibitory synaptic plasticity possibly develops at postsynaptic GABAA and GABAB synapses as well as presynaptic GABAB synapses on GABAA terminals. GABAA transmission is modified during motor learning, as shown by (i) the decrease of SICI after motor practice (Nordstrom & Butler, 2002; Liepert et al., 2004; Perez et al., 2004; Rosenkranz, 2006; Rosenkranz et al., 2007) and (ii) by the impairment of practice-dependent plasticity by the GABAA agonist lorazepam (Teo et al., 2009). This study suggests also that GABAB transmission is durably decreased by motor learning and after artificial induction of plasticity by PAS. Plasticity of inhibitory synapses during motor learning or after PAS, as tested by LICI, probably involves postsynaptic GABAB receptors. Indeed, the concomitant decrease of SICI (see above) and of LICI (our results) argues against a presynaptically mediated effect. Also, baclofen, an agonist of post-synaptic GABAB receptors, decreased PAS-induced plasticity in human motor cortex (McDonnell et al., 2007).
Plastic changes in circuits of LAI during motor learning in dystonic patients
Although performance of patients was not different from that of HV, long-term modulation of the MEPs and of LAI was lacking. This may seem contradictory as performance in HV correlated with changes in LAI, suggesting a causal link between motor learning and changes in LAI. A first explanation is that the lack of PAS-induced changes of LAI is not due to a defective plasticity development. (i) An obvious explanation would be that, according to the homeostatic rule of plasticity (Muller et al., 2007), the decreased baseline LAI opposes the development of an LTD-like plasticity and favours that of LTP-like plasticity. Such a view fits with the opposite directions of PAS-induced changes of LAI in patients (LAI is increased) and HV (LAI is decreased). However, this view is not supported by the lack of PAS-induced modulation of LAI in the subgroup of patients with normal baseline LAI. Such an explanation also does not hold for LICI, as LICI levels at baseline are similar in patients and HVs. It also implies that regulatory mechanisms of synaptic plasticity operate adequately in dystonic patients, which is not the case (Quartarone et al., 2005; Kang et al., 2011). Other possibilities are (ii) a ceiling effect, and (iii) masking of the inhibition by a superimposed facilitation. This would fit with the observation that inhibition was sometimes replaced by a facilitation in patients with hand dystonia (Abbruzzese et al., 2001). (iv) There is also an inhibitory interaction between LICI and LAI (Sailer et al., 2003), with LAI inhibiting LICI, indicating that the two pathways have at least one common interneuron. If the PAS-induced effect on LICI and LAI developed at the level of this interneuron, the lack of PAS-induced change of LICI and LAI reflects the same phenomenon. This is unlikely as, after learning LAI is not modulated while LICI is normally decreased. If the lack of change of LAI and MEP size after PAS and learning does reflect a defective plasticity, how do the patients can maintain a normal performance? Other pathways in different cortical or subcortical areas might take over the function of the deficient structures and by-pass the corticospinal output. Such a view is supported by imaging studies showing that activity of the sensorimotor cortex is not enhanced while that of cerebellum and basal ganglia is abnormally elevated in dystonic patients performing repeated finger tapping (Blood et al., 2004) or in non-manifesting carriers of the DYT1 gene mutations performing an adaptation task at levels matching those of controls (Carbon et al., 2008). When the motor task is more complex or when the disease is more severe, the abilities of compensatory areas may be not enough and the performance would worsen.
Increased threshold for induction of LTP in dystonic patients
A striking result was the lack of effect of PAS in the dystonic patients: neither the target FPB, nor the neighbour ADM MEPs were modified after PAS. Previous results in the literature using different PAS protocols, reported that PAS-induced effects in the muscle receiving spatially congruent peripheral somatosensory input, are either as large as in healthy subjects (Kang et al., 2011) or even enhanced (Quartarone et al., 2003; Weise et al., 2006) while a spread of the effects to the neighbour muscles not receiving the sensory inputs was absent in healthy subjects and constantly found in patients (Quartarone et al., 2003; Weise et al., 2006, 2011). It stresses that effects of one given plasticity induction protocol cannot be transferred to another one or even to the same yet applied with different intensities or frequency or duration of stimulation. It has been shown that the interaction between the age of the subjects and the ‘dose’ of the plasticity-induction protocol (i.e. number of stimulations, intensity of the TMS) influences the degree of PAS-induced MEP changes: the combination older age and high ‘dose’ protocol being the more prejudicial for plasticity (Muller-Dahlhaus et al., 2008). Our PAS protocol delivered more pairs of stimulation at a higher frequency than in most of the previous studies yet we used lower intensity of stimulation (0.5 mV MEPs instead of 1 mV MEPs). The latter could be the key factor, as when we increased the intensity to evoke a 1 mV MEP, a robust facilitation became evident in patients target muscle with a trend for the facilitation to spread to the non-target ADM. It suggests that the threshold for evoking plasticity was higher in the patients. As soon as this threshold was surpassed, the plasticity increased more rapidly than normal. This result fits with a previous study (Edwards et al., 2006) showing that the response to a theta-burst stimulation protocol was less in asymptomatic carriers of the DYT1 gene mutation compared to HV while it was more in the symptomatic patients. It may also explain why, in genetically predisposed subjects, the repetition of a motor task is necessary before dystonic symptoms occur. Also the learning task did not induce an increase of the control 1 mV MEP as it did in HV strengthening the view that in physiological conditions LTP-like plasticity may have an elevated threshold of induction.
Abnormal plasticity in inhibitory circuits in dystonia
In HV, PAS-induced plasticity of inhibitory circuits was related to the ‘dose’ of the plasticity-induction protocol, as already shown for other plasticity-induction protocols (McAllister et al., 2009). With a high-‘dose’ protocol (TMS set to evoke a 1-mV MEP) PAS induced a facilitation of the MEPs and no change in LICI or LAI. Results obtained in patients were thus similar to that of HV except for a trend of a spread of the effects to neighbouring muscles.
In HV, with a ‘low-dose’ plasticity-induction protocol (TMS set to evoke a 0.5-mV MEP), MEP facilitation was present and LTD-like plasticity of inhibitory circuits supporting LICI and LAI was evident. In contrast, in dystonic patients, MEPs were not modified and LICI levels remained unchanged (WC) or even increased (MD). Before concluding that intracortical inhibitory pathways have an abnormal susceptibility to develop plasticity in dystonia, we need to consider that PAS-induced changes of LAI and LICI might be driven by the plasticity induced in excitatory intracortical pathways, which is abnormal in dystonia (Quartarone et al., 2003; Weise et al., 2006). This raises the issue of whether the abnormal plasticity in inhibitory pathways of dystonic patients is secondary to the absence of LTP-like plasticity of excitatory synapses or reflects a genuine impairment of plastic mechanisms in inhibitory synapses with a possible uncoupling of excitatory synapses and inhibitory synapses. In slice experiments induction of LTP in excitatory synapses is coupled with complex plastic changes in inhibitory synapses (Perez et al., 1999; Patenaude et al., 2003; Woodin et al., 2003; Huang et al., 2005; Mendoza et al., 2006; Saraga et al., 2008). In our experimental conditions, MEPs are probably generated by a descending volley comprising several successive I-waves. I-waves are trans-synaptic in origin, and result from the excitation of glutamatergic intracortical pathways with separate pathways generating the early (I1) and later waves (I2, I3, …) (Ziemann & Rothwell, 2000) (see Fig. 6). LICI and LAI act by decreasing the later components of the corticospinal volley (I3 waves) and have no effect on early components (D wave and I1 waves) (Chen et al., 1999; Di Lazzaro et al., 2002). By analogy with ‘high-dose’ PAS, for which it was demonstrated directly by epidural recordings (Di Lazzaro et al., 2009), ‘low-dose’ PAS probably exerts its effects by influencing the cortico-cortical connections of the motor cortex that generate later I-waves. The relative ‘weight’ of I3 components in the test volley is then higher after PAS than before and the relative effect of the conditioning volley involved in LICI (and LAI) decreases. Accordingly, if the relative ‘weight’ of I3 components in the test volley is higher after ‘low-dose’ than after ‘high-dose’ PAS, this explains why LAI and LICI are differentially affected after ‘low-dose’ and ‘high-dose’ PAS. In dystonic patients, I3 waves are not modified by inhibitory inputs as in HV (Di Lazzaro et al., 2009), a possible explanation for the lack of effects of PAS.
The fact that the same modulation of MEPs as seen after PAS25/0.5mV and learning can be accompanied by different modulations of LICI/LAI levels (MEPs increased after PAS25/0.5mV and PAS25/1 mV in HV while LICI is respectively decreased or unchanged; MEPs unchanged after PAS25/0.5 mV and learning in patients while LICI is respectively unchanged or decreased) argues against a dependency of the plastic effects in inhibitory pathways to the LTP-like plasticity development.
In keeping with the explanations given above, both deficient PAS-induced effects on the MEPs and on the LICI level are probably related to an impaired ability of synapses in the late I wave pathways (at B on Fig. 6) to develop plasticity.
While healthy subjects had decreased LICI after PAS, writer’s cramp patients had less or no modulation of LICI and musician’s dystonia patients showed increased LICI. This indicates that the baseline cortical organization, which is different in musicians and non-musicians probably because of many years of practice in the former group, may influence the expression of plasticity in inhibitory circuits involving GABAB receptors. Accordingly, intracortical inhibition as tested by SICI was reduced in healthy musicians compared with non-musicians (Nordstrom & Butler, 2002). The pattern of sensorimotor interaction was also shown to be different in healthy non-musicians compared with healthy musicians (Rosenkranz et al., 2005). Here, lasting changes in sensorimotor interactions, as tested by LAI, were found to have similar abnormal direction and extent in both dystonic groups compared with healthy non-musician subjects.
Comparison of PAS and learning-induced effects
We are aware of the fact that the two plasticity-induction protocols used in this study are not completely comparable. Nevertheless, they do probably share physiological mechanisms (LTP/LTD-like plasticity), and PAS-induced changes might help to get an idea of what happens when a highly selective contraction has to be encoded and stored. In this view abnormalities observed in patients are prominent. The susceptibility to plasticity of the different sets of inhibitory synapses involved in each protocol is different. This raises the issue of the validity of interpreting abnormalities of plasticity observed after artificial induction of plasticity in relation to physiological learning.
Decreases of LICI and of LAI of the trained muscle probably participate in the increase of excitability and the shaping of the representation of the trained muscle at the early stage of learning. In HV, a decrease of LAI at the early stage of the learning may contribute to a permissive effect for the afferent volley to shape the motor output. In dystonic patients, the increased threshold of induction of LTP-like plasticity with the deficient somatosensory-motor coupling may leave the excitability of the representation of the target or trained muscle unchanged. Contrast between the excitability of the motor representation of the trained muscle (FPB) and that of the neighbouring muscle (ADM) is enhanced in HV during motor learning while this contrast did not occur in patients: a similar small ADM MEP increase as in HV and no change of FPB MEP size (see Fig. 4A and B, lower part). This might oppose the building up of adapted motor programmes where a precise adjustment between neighbouring cortical motor representations is necessary.
We thank Devera Schoenberg for editing of the English text. We thank Nguyet Dang for technical assistance. This research was funded by the Intramural Research Program of the National Institutes of Health, National Institute of Neurological Disorders and Stroke. H.R. was funded by the Swiss National Funds PBSKB-104264, the Swiss Parkinson Society and Intramural NIH. J.-C.L. was funded by the Fondation pour la Recherche Médicale (FRM) and Intramural NIH. S.M. was funded by Intramural NIH and INSERM.
abductor digiti minimi
conditioning stimulus to test stimulus at 90 ms
conditioning stimulus to test stimulus at 2.5 ms
flexor pollicis brevis
long afferent inhibition
long intracortical inhibition
mean peak acceleration
paired associative stimulation
resting motor threshold
short intracortical inhibition
transcranial magnetic stimulation
- 2001) Abnormalities of sensorimotor integration in focal dystonia: a transcranial magnetic stimulation study. Brain, 124, 537–545. , , , & (
- 2008) Short intracortical and surround inhibition are selectively reduced during movement initiation in focal hand dystonia. J. Neurosci., 28, 10363–10369. , , , , & (
- 2004) Basal ganglia activity remains elevated after movement in focal hand dystonia. Ann Neurol., 55, 744–748. , , , , , , & (
- 1992) Optimal focal transcranial magnetic activation of the human motor cortex: effects of coil orientation, shape of the induced current pulse, and stimulus intensity. J. Clin. Neurophysiol., 9, 132–136. , , , , & (
- 2008) Increased cerebellar activation during sequence learning in DYT1 carriers: an equiperformance study. Brain, 131, 146–154. , , , , & (
- 1997) Impaired inhibition in writer’s cramp during voluntary muscle activation. Neurology, 49, 1054–1059. , , & (
- 1999) Mechanism of the silent period following transcranial magnetic stimulation. Evidence from epidural recordings. Exp. Brain Res., 128, 539–542. , & (
- 2009) Afferent-induced facilitation of primary motor cortex excitability in the region controlling hand muscles in humans. Eur. J. Neurosci., 30, 439–448. , , , , , , & (
- 2002) Short-term reduction of intracortical inhibition in the human motor cortex induced by repetitive transcranial magnetic stimulation. Exp. Brain Res., 147, 108–113. , , , , , , , & (
- 2009) Reduced cerebral cortex inhibition in dystonia: direct evidence in humans. Clin. Neurophysiol., 120, 834–839. , , , , , , , & (
- 2003) Different patterns of electrophysiological deficits in manifesting and non-manifesting carriers of the DYT1 gene mutation. Brain, 126, 2074–2080. , , , & (
- 2006) Abnormalities in motor cortical plasticity differentiate manifesting and nonmanifesting DYT1 carriers. Mov. Disord., 21, 2181–2186. , , , & (
- 2002) Two phases of intracortical inhibition revealed by transcranial magnetic threshold tracking. Exp. Brain Res., 143, 240–248. , , , & (
- 2008) Inhibitory circuits and the nature of their interactions in the human motor cortex a pharmacological TMS study. J. Physiol., 586, 495–514. , , & (
- 2003) Impaired sequence learning in carriers of the DYT1 dystonia mutation. Ann Neurol., 54, 102–109. , , , , , , & (
- 2007) Short-term cortical plasticity in patients with dystonia: a study with repetitive transcranial magnetic stimulation. Mov. Disord., 22, 1436–1443. , , , , & (
- 2006) Pathophysiology of dystonia. J Neural Transm., 77 (Suppl), 485–488. (
- 1996) Long-term depression of horizontal connections in rat motor cortex. Eur. J. Neurosci., 8, 658–665. & (
- 2005) Common molecular pathways mediate long-term potentiation of synaptic excitation and slow synaptic inhibition. Cell, 123, 105–118. , , , , , , & (
- 2011) Deficient homeostatic regulation of practice-dependent plasticity in writer’s cramp. Cereb. Cortex, 21, 1203–1212. , , , & (
- 2008) The effect of simultaneous contractions of ipsilateral muscles on changes in corticospinal excitability induced by paired associative stimulation (PAS). Neurosci. Lett., 445, 7–11. & (
- 2006) Associative plasticity in human motor cortex during voluntary muscle contraction. J. Neurophysiol., 96, 1337–1346. , , & (
- 2005) Distinct basal ganglia territories are engaged in early and advanced motor sequence learning. Proc Natl Acad Sci USA, 102, 12566–12571. , , , , , & (
- 1998) Task-dependent changes of intracortical inhibition. Exp. Brain Res., 118, 421–426. , , & (
- 2004) Exercise-induced changes of motor excitability with and without sensory block. Brain Res., 1003, 68–76. , , , & (
- 2009) Selective modulation of intracortical inhibition by low-intensity Theta Burst Stimulation. Clin. Neurophysiol., 120, 820–826. , & (
- 2007) Suppression of LTP-like plasticity in human motor cortex by the GABAB receptor agonist baclofen. Exp. Brain Res., 180, 181–186. , & (
- 2006) Differential induction of long term synaptic plasticity in inhibitory synapses of the hippocampus. Synapse, 60, 533–542. , , , , & (
- 2001) Role of the human motor cortex in rapid motor learning. Exp. Brain Res., 136, 431–438. , , , & (
- 2002) Early consolidation in human primary motor cortex. Nature, 415, 640–644. , , , , , , , & (
- 2007) Homeostatic plasticity in human motor cortex demonstrated by two consecutive sessions of paired associative stimulation. Eur. J. Neurosci., 25, 3461–3468. , , & (
- 2008) Interindividual variability and age-dependency of motor cortical plasticity induced by paired associative stimulation. Exp. Brain Res., 187, 467–475. , , & (
- 1997) Intracortical facilitation and inhibition after transcranial magnetic stimulation in conscious humans. J. Physiol., 498, 817–823. , , & (
- 2002) Reduced intracortical inhibition and facilitation of corticospinal neurons in musicians. Exp. Brain Res., 144, 336–342. & (
- 2003) The variability of intracortical inhibition and facilitation. Clin. Neurophysiol., 114, 2362–2369. , & (
- 2003) GABAB receptor- and metabotropic glutamate receptor-dependent cooperative long-term potentiation of rat hippocampal GABAA synaptic transmission. J. Physiol., 553, 155–167. , , , & (
- 1999) Differential induction of long-lasting potentiation of inhibitory postsynaptic potentials by theta patterned stimulation vs. 100-Hz tetanization in hippocampal pyramidal cells in vitro. Neuroscience, 90, 747–757. , , , & (
- 2004) Motor skill training induces changes in the excitability of the leg cortical area in healthy humans. Exp. Brain Res., 159, 197–205. , , & (
- 2003) Abnormal associative plasticity of the human motor cortex in writer’s cramp. Brain, 126, 2586–2596. , , , , , , , , & (
- 2005) Homeostatic-like plasticity of the primary motor hand area is impaired in focal hand dystonia. Brain, 128, 1943–1950. , , , , , , , , & (
- 2009) Abnormal sensorimotor plasticity in organic but not in psychogenic dystonia. Brain, 132, 2871–2877. , , , , , , , & (
- 1995) Changes in excitability of motor cortical circuitry in patients with Parkinson’s disease. Ann Neurol., 37, 181–188. , & (
- 1998) Strengthening of horizontal cortical connections following skill learning. Nat. Neurosci., 1, 230–234. , , & (
- 2000) Learning-induced LTP in neocortex. Science, 290, 533–536. , & (
- 2005) Pathophysiological differences between musician’s dystonia and writer’s cramp. Brain, 128, 918–931. , , , , & (
- 2006) Differences between the effects of three plasticity inducing protocols on the organization of the human motor cortex. Eur. J. Neurosci., 23, 822–829. & (
- 2007) Differential modulation of motor cortical plasticity and excitability in early and late phases of human motor learning. J. Neurosci., 27, 12058–12066. , & (
- 1994) Non-invasive electrical and magnetic stimulation of the brain, spinal cord and roots: basic principles and procedures for routine clinical application. Report of an IFCN committee. Electroencephalogr. Clin. Neurophysiol., 91, 79–92. , , , , , , , , , , , , , , , (
- 1999) Magnetic stimulation: motor evoked potentials. The International Federation of Clinical Neurophysiology. Electroencephalogr. Clin. Neurophysiol., (Suppl), 52, 97–103. , , , , & (
- 2009) Associative plasticity in intracortical inhibitory circuits in human motor cortex. Clin. Neurophysiol., 120, 1204–1212. , , , & (
- 2003) Short and long latency afferent inhibition in Parkinson’s disease. Brain, 126, 1883–1894. , , , , & (
- 2008) Inhibitory synaptic plasticity regulates pyramidal neuron spiking in the rodent hippocampus. Neuroscience, 155, 64–75. , , , & (
- 2000) Induction of plasticity in the human motor cortex by paired associative stimulation. Brain, 123, 572–584. , , , & (
- 2002) Mechanisms of enhancement of human motor cortex excitability induced by interventional paired associative stimulation. J. Physiol., 543, 699–708. , , , & (
- 2006) Temporary occlusion of associative motor cortical plasticity by prior dynamic motor training. Cereb. Cortex, 16, 376–385. , , , , , & (
- 1994) Impairment of GABAA receptor function by N-methyl-D-aspartate-mediated calcium influx in isolated CA1 pyramidal cells. Neuroscience, 62, 813–828. & (
- 2004) Elevated threshold for intracortical inhibition in focal hand dystonia. Mov. Disord., 19, 1312–1317. & (
- 2009) Differing effects of intracortical circuits on plasticity. Exp. Brain Res., 193, 555–563. , , , & (
- 2001) A comparison of ‘Early’ and ‘Late’ stage brain activation during brief practice of a simple motor task. Brain Res. Cogn. Brain Res., 10, 303–316. , , , , & (
- 1992) Human motor evoked responses to paired transcranial magnetic stimuli. Electroencephalogr. Clin. Neurophysiol., 85, 355–364. , , & (
- 2006) The two sides of associative plasticity in writer’s cramp. Brain, 129, 2709–2721. , , , , , & (
- 2011) Loss of topographic specificity of LTD-like plasticity is a trait marker in focal dystonia. Neurobiol. Dis., 42, 171–176. , , , & (
- 2003) A temporally asymmetric Hebbian rule governing plasticity in the human motor cortex. J. Neurophysiol., 89, 2339–2345. , , , , , , & (
- 2003) Coincident pre- and postsynaptic activity modifies GABAergic synapses by postsynaptic changes in Cl- transporter activity. Neuron, 39, 807–820. , & (
- 2000) I-waves in motor cortex. J. Clin. Neurophysiol., 17, 397–405. & (