OUP user menu

Motor cortex plasticity in Parkinson's disease and levodopa-induced dyskinesias

Francesca Morgante, Alberto J. Espay, Carolyn Gunraj, Anthony E. Lang, Robert Chen
DOI: http://dx.doi.org/10.1093/brain/awl031 1059-1069 First published online: 13 February 2006

Summary

Experimental models of Parkinson's disease have demonstrated abnormal synaptic plasticity in the corticostriatal system, possibly related to the development of levodopa-induced dyskinesias (LID). We tested the hypothesis that LID in Parkinson's disease is associated with aberrant plasticity in the human motor cortex (M1). We employed the paired associative stimulation (PAS) protocol, an experimental intervention involving transcranial magnetic stimulation (TMS) and median nerve stimulation capable of producing long-term potentiation (LTP) like changes in the sensorimotor system in humans. We studied the more affected side of 16 moderately affected patients with Parkinson's disease (9 dyskinetic, 7 non-dyskinetic) and the dominant side of 9 age-matched healthy controls. Motor-evoked potential (MEP) amplitudes and cortical silent period (CSP) duration were measured at baseline before PAS and for up to 60 min (T0, T30 and T60) after PAS in abductor pollicis brevis (APB) and abductor digiti minimi (ADM) muscles. PAS significantly increased MEP size in controls (+74.8% of baseline at T30) but not in patients off medication (T30: +0.07% of baseline in the non-dyskinetic, +27% in the dyskinetic group). Levodopa restored the potentiation of MEP amplitudes by PAS in the non-dyskinetic group (T30: +64.9% of baseline MEP) but not in the dyskinetic group (T30: −9.2% of baseline). PAS prolonged CSP duration in controls. There was a trend towards prolongation of CSP in the non-dyskinetic group off medications but not in the dyskinetic group. Levodopa did not restore CSP prolongation by PAS in the dyskinetic group. Our findings suggest that LTP-like plasticity is deficient in Parkinson's disease off medications and is restored by levodopa in non-dyskinetic but not in dyskinetic patients. Abnormal synaptic plasticity in the motor cortex may play a role in the development of LID.

  • levodopa-induced dyskinesia
  • Parkinson's disease
  • plasticity
  • motor cortex
  • transcranial magnetic stimulation
  • ADM = abductor digiti minimi
  • APB = abductor pollicis brevis
  • CSP = cortical silent period
  • DA = dopamine
  • LE = levodopa equivalents
  • LID = levodopa-induced dyskinesias
  • LTP = long-term potentiation
  • MEP = motor-evoked potentials
  • NMDA = N-methyl-d-aspartate
  • PAS = paired associative stimulation
  • RMT = resting motor threshold
  • TMS = transcranial magnetic stimulation
  • UPDRS = Unified Parkinson's Disease Rating Scale

Introduction

Chronic dopaminergic treatment of Parkinson's disease is complicated by the development of levodopa-induced dyskinesias (LID) in ∼40% of patients after 4–6 years of levodopa therapy (Ahlskog and Muenter, 2001). The commonest forms of LID are peak-dose dyskinesias that occur during the period of maximal benefit of parkinsonian symptoms when levodopa brain concentration is highest, are mainly choreic in nature and affect predominantly the upper limbs, usually starting on the side most affected by the disease (Marconi et al., 1994; Fahn, 2000).

Although several hypotheses have been formulated to explain these motor complications, the underlying synaptic mechanism of LID remains unclear (Obeso et al., 2000). A model of the basal ganglia motor circuit predicts that LID are associated with decreased firing in the globus pallidus pars interna (GPi) and in the subthalamic nucleus (STN), secondary to non-physiological stimulation of dopamine (DA) receptors in the denervated striatum (Wichmann and DeLong, 1996). However, observations such as the marked reduction of LID in parkinsonian patients after pallidotomy (Lang et al., 1999) are opposite to the prediction of the model. Moreover, the model does not take into account the role of a number of connections, most notably the corticostriatal projections to the basal ganglia (Obeso et al., 2000). Studies conducted in vitro (Calabresi et al., 2000) in animal models of parkinsonism (Engber et al., 1994) and in patients with Parkinson's disease (Verhagen et al., 1998) have suggested that the glutamatergic corticostriatal projection to medium spiny neurons in the striatum may play an important role in the priming and development of LID. Interactions between DA receptors and ionotropic glutamate receptors on the dendritic spines of the medium spiny neurons appear to be a major mechanism modulating synaptic transmission in the striatum as well as the firing pattern of corticostriatal neurons by activation of protein kinase and phosphorylation of ionotropic glutamate receptors (Kotter, 1994; Chase and Oh, 2000). Activation of ionotropic glutamate receptors, represented by α-amino- 3-hydroxyl-5-methylisoxazole-4-propionic acid (AMPA) and N-methyl-d-aspartate (NMDA) receptors, mediate long-term potentiation (LTP), a persistent enhancement in synaptic transmission, and long-term depression (LTD), where synaptic transmission is decreased (Malenka and Bear, 2004). Behavioral correlates of synaptic plasticity within the striatum include the production and storage of motor skills (Calabresi et al., 1996). In late stages of Parkinson's disease, chronic intermittent ‘pulsatile’ non-physiologic stimulation of DA receptors on striatal neurons (Chase et al., 2000) can induce post-synaptic modifications in NMDA and AMPA receptor firing, causing aberrant motor patterns leading to motor complications. Therefore, these studies suggest that a pathological form of striatal synaptic plasticity, related to abnormal NMDA receptors function, could cause the development of abnormal motor patterns leading to LID.

It is possible to induce LTP-like changes in the sensorimotor system at the level of the human motor cortex (M1), by means of a validated experimental intervention known as paired associative stimulation (PAS) developed by Stefan and coworkers (Stefan et al., 2000; Wolters et al., 2003, 2005). Low-frequency median nerve stimulation paired with transcranial magnetic stimulation (TMS) over the contralateral M1 area at interstimulus intervals of about 25 ms enhances M1 excitability and at interstimulus intervals of about 10 ms decreases M1 excitability (Wolters et al., 2003). Changes in motor cortex excitability induced by PAS occur rapidly; last for at least 60 min, are topographically specific, occurring selectively in median nerve innervated muscles, and are blocked by dextromethorphan (NMDA receptor antagonist) or nimodipine (L-type voltage-gated channel antagonist) (Stefan et al., 2002). The physiological and pharmacological profile of these modifications in M1 excitability suggests that a LTP/LTD-like mechanism may underlie cortical plasticity induced by PAS.

In the present study, we employed PAS to examine synaptic plasticity in patients with Parkinson's disease and tested the hypothesis that LID correlate with an aberrant form of plasticity in the human M1.

Methods

Subjects

We studied 16 patients (9 dyskinetic, 7 non-dyskinetic, 12 men, 4 women, aged 68.2 ± 6.9 years, range 50–80 years) with moderate Parkinson's disease without significant tremor and 9 age-matched healthy controls (6 men, 3 women, aged 63.1 ± 5.8 years, range 57–74 years). Patients were recruited from the Movement Disorders Clinic at the Toronto Western Hospital, and their clinical features are shown in Table 1. We excluded patients with obvious cognitive impairment, as assessed by their treating physicians, as well as those taking antidepressant medications. All subjects were right-handed according to the Edinburgh handedness inventory (Oldfield, 1971). The dyskinetic and non-dyskinetic patients were matched for disease severity and duration. We selected patients with non-disabling dyskinesias [score 1 for item 33 of Unified Parkinson's Disease Rating Scale (UPDRS) IV] to avoid excessive muscle activation during TMS.

View this table:
Table 1

Clinical features of Parkinson's disease patients

PatientAge (years)/ genderDisease durationUPDRSaUPDRS IIIbUPDRS IVcH & YdMedications (LE, mg/day)eMore affected side
174 M1040.528.502600 (L-dopa)Left
272 M633.514.502551 (L-dopa, PRA, AMA)Right
371 M1373.545.5031000 (L-dopa, AMA)Left
461 M6372402974 (L-dopa, PRA)Right
572 M6271502800 (L-dopa, PRA)Right
667 M8442702600 (L-dopa, ENT)Right
772 M13432803900 (L-dopa, PRA)Left
861 F859.537.5131250 (L-dopa)Left
963 M738.521.5121500 (L-dopa, PRA)Right
1072 F8282112600 (L-dopa)Left
1180 M2060.538.514951 (L-dopa, PER)Right
1250 F7513014738 (L-dopa, ROP, ENT)Right
1368 M1041.526.513800 (L-dopa)Right
1470 F14543123500 (L-dopa)Left
1566 M1044.521.512550 (L-dopa)Left
1673 M2055.529.5231050 (L-dopa, PER, AMA)Left
Non-dyskineticDyskinetic
Age69.8 ± 4.567 ± 8.5
UPDRS III-OFF26.1 ± 10.428.6 ± 6.6
UPDRS III-ON13.5 ± 6.912.3 ± 3.9
UPDRS42.6 ± 14.848.11 ± 10.81
UPDRS IV0.01.22 ± 0.4
H & Y OFF2.3 ± 0.52.9 ± 0.8
Duration Parkinson's disease8.9 ± 3.211.6 ± 5.2
LE (mg/day)775.1 ± 190.8882 ± 337.9
  • Patients 1–7: Non-dyskinetic Parkinson's disease; Patients 8–16: dyskinetic Parkinson's disease. Medications are abbreviated as follows: AMA = amantadine, ENT = entacapone, L-dopa = levodopa, ROP = ropinirole, PER = pergolide, PRA = pramipexole.

  • a Unified Parkinson's Disease Rating Scale (UPDRS): total score; bUPDRS-motor score ‘off medication’; cUPDRS-IV complication of therapy scale (item 32–33); dmodified Hoehn and Yahr scale; elevodopa equivalent (LE).

Patients were studied on two separate days in the practically defined off state, after overnight withdrawal of dopaminergic medications (‘off med’ condition), and ∼1 h after administration of the levodopa equivalent (LE) (as immediate release levodopa-carbidopa) of their first morning dose (‘on med’ condition), once clinical benefits were fully documented by neurological examination. The patients' medications were converted into LEs calculated with the following formula: LE = total dose of immediate release levodopa (with peripheral decarboxylase inhibitor) + (0.75 × dose of controlled-release levodopa) + (10 × dose of bromocriptine) + (100 × dose of pergolide) + (67 × dose of pramipexole) + (16.7 × dose of ropinirole) (Fine et al., 2000). Amantadine was withheld for the 3 days preceding the experimental session (three patients: two non-dyskinetic, one dyskinetic). If patients had wearing off, an additional dose of levodopa was administered during the experiment before their expected clinical decline (five patients: one non-dyskinetic, four dyskinetic).

Parkinsonism was assessed in both ‘off med’ and ‘on med’ conditions with the motor section (item 18–31) of the UPDRS, the modified Hoehn and Yahr scale and the Schwab and England scale. Dyskinesias were also assessed with the Lang-Fahn Activities of Daily Living Dyskinesias Scale and UPDRS part IV dyskinesias ratings (item 32–34).

All patients gave written informed consent. The protocol was approved by the University Health Network Research Ethics Board in accordance with the Declaration of Helsinki on the use of human subjects in experiments.

Experimental set-up

Transcranial magnetic stimulation

TMS was performed with a 7-cm figure of eight coil connected to a Magstim 200 stimulator (The Magstim Company, Whitland, UK). The handle of the coil pointed backwards and laterally at about 45° from the midline, thus generating a posterior–anterior current in the brain.

We established the optimal position for activating the contralateral abductor pollicis brevis (APB) muscle by moving the coil in 0.5 cm steps around the presumed hand motor area. The optimal coil position that elicited the largest motor-evoked potentials (MEPs) with the steepest slope was marked on the scalp. The same procedure was then repeated for the abductor digiti minimi (ADM) muscle. This set-up allowed us to compare the effect of PAS in a median nerve innervated muscle (APB, homotopic PAS) with an ulnar nerve innervated muscle (ADM, heterotopic PAS). MEPs were recorded from the more affected arm of the patients and the dominant side of the controls. TMS was adjusted to produce MEPs of ∼1 mV in the relaxed contralateral APB muscle.

EMG recording

EMG was recorded from APB and ADM muscles contralateral to the stimulated motor cortex, using disposable disc electrodes with a belly-tendon montage. EMG was amplified (Intronix Technologies Corporation Model 2024F, Bolton, Ontario, Canada), filtered (band pass 2 Hz–2.5 kHz), digitized at 5 kHz (Micro 1401, Cambridge Electronics Design, Cambridge, UK) and stored in a personal computer for off-line analysis. Subjects were asked to relax throughout the experimental session with EMG monitoring on a computer screen and via loudspeakers.

Median nerve stimulation

The median nerve contralateral to TMS was stimulated at the wrist with standard bar electrodes with the cathode positioned proximally. The pulses were constant current square wave pulses with a pulse width of 200 μs. Stimulus intensity was set at 300% of the perceptual threshold.

Paired associative stimulation

The PAS protocol involved pairing median nerve stimulation at the wrist with TMS over the contralateral motor cortex 21.5 ms later at a rate of 0.01 Hz for 30 min (180 pairs). This interstimulus interval was chosen since PAS maximally enhances cortical excitability at the interstimulus interval that is equal to N20 latency of the somatosensory evoked potential (Ziemann et al., 2004).

Experimental design and data analysis

For each experimental session (healthy controls, patients ‘off med’, patients ‘on med’), we tested the effect of PAS on resting motor threshold (RMT), MEP amplitude and cortical silent period (CSP) duration in the APB and ADM muscles. These parameters of corticospinal excitability were assessed before PAS (baseline) and for up to 60 min (T0, T30 and T60) after PAS (Fig. 1).

Fig. 1

Overview of the experimental paradigm. Patients with Parkinson's disease were studied in separate sessions ‘off’ and ‘on’ medications. Measure of cortical excitability (RMT, MEP amplitude and CSP) were assessed from the APB and ADM muscles at baseline and for up to 1 h after the end of the PAS protocol. PAS involved median nerve stimulation preceding TMS over the motor spot for APB by 21.5 ms, at 0.01 Hz, for 30 min.

The RMT was defined as the minimum stimulator intensity eliciting MEPs of ≥50 μV in 5 out of 10 consecutive trials in the relaxed target muscles (Rossini et al., 1994). MEPs were recorded from APB and ADM in separate trials with 20 stimuli at 0.1 Hz over the contralateral M1. The stimulus intensity was adjusted to produce MEPs of ∼1 mV in the relaxed target muscles at baseline and was kept constant at the different times of assessment during the experiment (baseline, T0, T30, T60). MEP amplitudes were measured peak to peak and then averaged.

To assess CSP duration, 10 trials were recorded during isometric contraction (∼15% of maximum voluntary contraction kept constant with visual feedback through an oscilloscope) of the APB and ADM muscles in separate trials with TMS at an intensity of 130% of RMT. For each trial 200 ms of pre-stimulus EMG was recorded. We used an automated method to measure CSP duration, as previously described by Daskalakis et al. (Daskalakis et al., 2003). The CSP was measured for each trial from the MEP onset to the initial return of EMG activity. The trials were high-pass filtered to remove movement artefacts. Then squaring of the trace was employed to magnify and rectify the EMG activity and to further isolate the MEP and the EMG flat period. The CSP offset was determined from the processed waveform and was the first point to exceed 25% of the mean pre-stimulus EMG that consistently enclosed the EMG flat period after the end of the MEP. This method is objective and avoids the potential biases and discrepancies inherent to the more conventional visually guided CSP measuring techniques.

The background EMG area in the 40 ms preceding the TMS pulse was measured in all the subjects for each of the 20 trials used for measurement of MEP amplitude to ensure comparability of the baseline activity between controls, non-dyskinetic and dyskinetic patients and at different times of the experiment. Background EMG area was also evaluated in the 200 ms preceding TMS to ensure a constant degree of contraction during the CSP recording. A trial was excluded if the pre-stimulus EMG area exceeded mean + 2 SD (standard deviation) of the pre-stimulus EMG area measured at baseline. The experimental block was excluded from the analysis if >50% of trials were rejected.

Statistical analysis

UPDRS scores, Hoehn and Yahr, disease duration and LEs were compared between dyskinetic and non-dyskinetic patients by the Mann–Whitney U-test. Wilcoxon signed rank test was used to compare UPDRS scores in the ‘off’ and ‘on’ condition within the patient groups. For the other variables, Kolmogorov–Smirnov normality tests were applied. Non-parametric tests were used if the test was significant (P < 0.05); otherwise parametric tests were used. The Mann–Whitney U-test was used to compare background EMG area among the groups tested. Group differences in TMS variables at baseline were compared by unpaired t-test.

To test the effect of PAS protocol on motor cortex excitability in Parkinson's disease, we obtained four blocks of measurements (baseline, T0, T30 and T60) for each measure of cortical excitability (RMT, MEP, CSP). Separate repeated measures analyses of variance (ANOVA) were performed for RMT, MEP amplitudes and CSP. We computed repeated measures ANOVA with time (four levels: baseline, T0, T30, T60) as within-subjects factors and group (healthy controls, dyskinetic ‘off med’, non-dyskinetic ‘off med’, dyskinetic ‘on med’, non-dyskinetic ‘on med’) as between-subjects factors. If the factors group or time showed significant main effects or significant interaction, the effect of time in each group was explored with separate one-way repeated measures ANOVA.

The effect of levodopa treatment on PAS was analysed with a repeated measures ANOVA with time as within-subjects factors, and treatment (‘off’ versus ‘on’) and group (dyskinetic patients versus non-dyskinetic patients) as between-subjects factors.

Conditional on a significant f-value, post hoc multiple paired t-tests were performed to explore differences between groups. Because of multiple comparisons, significance was set at P < 0.01. For the other statistical tests, significance was set at P < 0.05. Unless otherwise stated, data are given as mean ± standard error of the mean (SEM).

Results

Clinical features

The two groups of patients did not differ in age, total UPDRS scores, UPDRS motor score (on and off med), modified Hoehn and Yahr scale (on and off med), duration of Parkinson's disease and dosage of dopaminergic medications expressed as LE (Table 1). In the dyskinetic patients, the Lang-Fahn Activities of Daily Living Dyskinesias Scale scores were 2.3 ± 1.2.

Resting motor threshold

RMT at baseline was similar among the five groups in both APB (controls 38.5 ± 3.4%, non-dyskinetic ‘off med’ 33.8 ± 2.5%, dyskinetic ‘off med’ 39.4 ± 2.5%, non-dyskinetic ‘on med’ 35 ± 2.2%, dyskinetic ‘on med’ 39.4 ± 2.9%) and ADM muscles (controls 39.1 ± 3.3%, non-dyskinetic ‘off med’ 33.1 ± 2.6%, dyskinetic ‘off med’ 40.1 ± 2.7%, non-dyskinetic ‘on med’ 36.5 ± 2.6%, dyskinetic ‘on med’ 40 ± 2.5%). ANOVA showed no significant effect of the factors time after PAS or group on RMT in both APB and ADM muscles.

MEP amplitudes

Examples from representative subjects are shown in Fig. 2, and the group data are shown in Fig. 3 and Table 2. For the APB muscle, PAS induced a significant increase in MEP amplitudes in control subjects (+74.8% of baseline at T30) but not in either the non-dyskinetic (T30: +0.07% of baseline) or dyskinetic patients off medication (T30: +27% of baseline). Levodopa restored potentiation of MEP amplitudes by PAS in the non-dyskinetic group, leading to an increase in MEP amplitude in the APB muscle (T30: +64.9% of baseline MEP). In contrast, dyskinetic patients on medications exhibited no increase in MEP after PAS (−9.2% of baseline at T30).

Fig. 2

Representative MEPs from the APB muscle of a control subject and from a dyskinetic and non-dyskinetic patient off and on medication at baseline and 30 min after PAS (T30). Each MEP represents average of 20 trials. PAS led to increased MEP amplitude in the control subject but not in the dyskinetic or the non-dyskinetic patients off medications. In the on medication state, the effect of PAS on MEP amplitude was restored in the non-dyskinetic patient but not in the dyskinetic patient.

View this table:
Table 2

PAS-induced time-dependent changes in MEPs (mV)

BaselineT0T30T60
Controls
APB0.9 ± 0.051.3 ± 0.181.7 ± 0.09*1.9 ± 0.26*
ADM0.8 ± 0.091.08 ± 0.091.03 ± 0.061.19 ± 0.16
Non-dyskinetic patients
APB
    Off1.1 ± 0.061.1 ± 0.071.18 ± 0.11.15 ± 0.11
    On1.1 ± 0.091.7 ± 0.21*1.86 ± 0.26*1.4 ± 0.23
ADM
    Off1 ± 0.091.2 ± 0.051.3 ± 0.11.2 ± 0.1
    On1.1 ± 0.171.5 ± 0.211.45 ± 0.171.54 ± 0.2
Dyskinetic patients
APB
    Off1.2 ± 0.131.3 ± 0.131.5 ± 0.191.4 ± 0.15
    On1.4 ± 0.191.1 ± 0.171.3 ± 0.251.5 ± 0.2
ADM
    Off1.1 ± 0.11.4 ± 0.241.4 ± 0.271.4 ± 0.21
    On1.1 ± 0.121.6 ± 0.251.4 ± 0.171.4 ± 0.18
  • * P < 0.01 by paired t-test compared with baseline. Values are mean ± SEM. Off = off medication; On = on medication.

These observations were confirmed by repeated measures ANOVA showing a main effect of time (F = 13.41, P < 0.0001) and a time × group interaction (F = 2.5, P = 0.0005). The effect of group was not significant. To further explore these effects, we computed separate ANOVAs for each group with time as the repeated measure. There was a significant effect of time on MEP amplitude in the control (F = 8.7, P = 0.0004) and non-dyskinetic on med (F = 6.0, P = 0.005) groups, whereas the effect was not significant in the other groups (non-dyskinetic off med, dyskinetic off or on med). In the control subjects the increase in MEP size was significant at T30 and T60 (P < 0.01, paired t-test), whereas in the non-dyskinetic Parkinson's disease on med was significant at T0 and T30 (P < 0.01, paired t-test) (Table 2 and Fig. 3).

Fig. 3

Effect of PAS on MEP amplitudes. The data is plotted as a ratio to the baseline MEP amplitude. Ratios higher than 1 indicate facilitation and ratios below 1 indicate inhibition of MEP amplitude. In the APB muscle, PAS increased MEP amplitudes in controls but not in dyskinetic and non-dyskinetic patients off medication. In the on medication state, potentiation of MEP by PAS in the APB muscle was restored in the non-dyskinetic group but not in the dyskinetic group. PAS did not significantly increase the MEP amplitude in the ADM muscle in controls and both patients' groups either in the ‘off’ or ‘on’ medication states. *P < 0.01, **paired t-test compared with baseline. The data for the control group is duplicated in the OFF and ON graphs for clarity of comparison. Error bars represent standard error.

For the ADM muscle, all groups showed a slight increase in MEP amplitude after PAS (Fig. 3 and Table 2). Repeated measures ANOVA showed a significant effect of time (F = 8.3, P < 0.0001) but no significant effect of group or time × group interaction. Separate repeated measures ANOVA for each group did not show a significant effect of time in any of the groups (control, non-dyskinetic off or on med, dyskinetic off or on med).

CSP duration

Figure 4 shows examples from representative subjects, and the group data are shown in Table 3 and Fig. 5. For the APB muscle, ANOVA showed a main effect of time (F = 3.2, P = 0.025) on CSP duration but no significant effect of group or time × group interaction. Separate ANOVA for each group showed significant effect of time on CSP duration in the control group (F = 4.1, P = 0.02) but not in the other groups (non-dyskinetic off or on med, dyskinetic off or on med) (Table 3).

Fig. 4

Effect of PAS and medication on CSP duration. The recordings are from the APB muscle in a control subject and a dyskinetic and a non-dyskinetic patient off and on medication at baseline and at T30. Each trace is the average of 10 trials. PAS prolonged the CSP in the control subject and in the non-dyskinetic patient but had no effect in the dyskinetic patient. Medication (on versus off) change prolonged the CSP but did not change the effect of PAS on CSP.

View this table:
Table 3

PAS-induced time-dependent changes in CSP (ms)

BaselineT0T30T60
Controls
APB157.3 ± 13.3177.5 ± 13.9*177 ± 14.3*176.3 ± 13.8*
ADM155.9 ± 9.7166.1 ± 10.7*169.5 ± 9.4*165.4 ± 8.2
Non-dyskinetic patients
APB
    Off140.6 ± 12143.9 ± 12.7153.4 ± 12.7148.6 ±11.8
    On161.4 ± 13.9167 ± 12.7172.2 ± 11.3167.9 ± 13.8
ADM
    Off138.9 ± 8.3153.8 ± 10.3*156 ± 10.5*152.9 ± 11.9
    On172 ± 11.4173.7 ± 10.5173.4 ± 12.2171.7 ± 10.9
Dyskinetic patients
APB
    Off148.2 ± 13.7143.3 ± 12.4142.4 ± 14.2143.4 ± 11.9
    On169 ± 5.9175.4 ± 9.1172.7 ± 9.6176.9 ± 12.6
ADM
    Off146.6 ± 14.2138.1 ± 11.4141.3 ± 13.7143.5 ± 12.3
    On166.1 ± 6.5164.2 ±10.5159.3 ± 11.2170.2 ± 11.3
  • * P < 0.01 by paired t-test compared with baseline values. Values are mean ± SEM. Off = off medication; On = on medication.

For the ADM muscle, there was a significant effect of time (F = 3.1, P = 0.029) and time × group interaction (F = 2.2, P = 0.021) on CSP duration but there was no significant effect of group (Table 3 and Fig. 5). Separate ANOVA for each group showed significant effect of time on CSP duration for the control group (F = 3.5, P = 0.032) and the non-dyskinetic off med group (F = 7.4, P = 0.002). The effect of time in the other groups (non-dyskinetic on med, dyskinetic off or on med) was not significant.

Fig. 5

Effect of PAS on CSP duration. PAS prolonged CSP duration in both the APB and ADM muscles in the control group. In non-dyskinetic patients, there was a tendency for increase in CSP duration following PAS. CSP did not change following PAS in the dyskinetic group, either on or off dopaminergic medications. Data are shown as a ratio to baseline CSP duration. Ratios higher than 1 indicate prolongation of CSP following PAS. *P < 0.01, paired t-test compared with baseline values. The data for the control group is duplicated in the OFF and ON graphs for clarity of comparison. Error bars represent standard error.

Effects of levodopa on MEP amplitude and CSP duration

A separate ANOVA was performed in patients with Parkinson's disease to specifically examine the effect of levodopa treatment on motor cortex plasticity.

MEP amplitudes

For the APB muscle, there was a significant effect of time (F = 4.2, P = 0.008) on MEP amplitude but no main effect of group (non-dyskinetic and dyskinetic) and treatment (off and on med). There were significant time × group (F = 3.0, P = 0.035) and time × treatment × group (F = 5.4, P = 0.002) interactions. The significant interactions can be explained by increase in MEP amplitude following PAS in the non-dyskinetic on med group, but not in the other groups (non-dyskinetic off med, dyskinetic off or on med) (Table 2 and Fig. 3).

For the ADM muscle, there was a significant effect of time (F = 7.0, P = 0.0003) on MEP amplitude but no main effect of group and treatment. There were no significant interactions. This confirmed the finding that there was a small increase in MEP amplitude following PAS in all groups but there was no significant difference between non-dyskinetic and dyskinetic Parkinson's disease patients and no significant effect of dopaminergic medications.

CSP duration

In the APB muscle, there was a significant effect of treatment (on med, off med, F = 4.6, P = 0.041) on CSP duration (Table 3 and Fig. 4). The effects of group and time were not significant and there were no significant interactions. In the ADM muscle, there was a marginal effect of treatment (F = 4.1, P = 0.054) but no main effect of group or time. There was a significant time × group interaction (F = 3.3, P = 0.025). The effect of treatment (on or off) confirmed that dopaminergic medications increased the CSP in both non-dyskinetic and dyskinetic groups both before and after PAS. In the ADM muscle, the significant time × group interaction is probably due to the non-dyskinetic group showing a tendency to increase CSP duration especially at T0, whereas the dyskinetic group tended to have decreased CSP following PAS especially at T0.

Background EMG area

EMG background activity preceding TMS was not different at baseline between controls, non-dyskinetic and dyskinetic patients both on and off medications in APB and ADM muscles (Table 4). The assessment of the MEP in APB off medication at T0 of Patient 14 and of the MEP in APB on medication at T30 of Patient 15 were excluded from the analysis because more than 50% of the trials were rejected for increased background EMG activity.

View this table:
Table 4

Pre-stimulus background EMG area (μV*ms)

APB-OFFAPB-ONADM-OFFADM-ON
Controls0.19 ± 0.0040.17 ± 0.008
Non- dyskinetic0.19 ± 0.020.18 ± 0.0020.19 ± 0.010.17 ± 0.04
Dyskinetic0.19 ± 0.010.22 ± 0.020.21 ± 0.020.27 ± 0.08
  • Values are mean ± SEM.

Discussion

Our results confirm that PAS at an interstimulus interval of 21.5 ms is capable of enhancing motor cortex excitability with increased MEP amplitude and prolonged CSP duration in normal subjects for at least 1 h (Stefan et al., 2000), confirming that PAS-induced after-effects are long lasting and that they may progressively increase with time. Since the changes induced by PAS share the features of the Hebbian LTP, such as associativity, input specificity and cooperativity (Bear and Malenka, 1994), the dynamic changes we have observed in motor cortex after PAS represent a form of motor cortex plasticity (Stefan et al., 2000). We also found topographical specificity of the PAS effect with greater changes in the homotopic APB muscle than the heterotopic ADM muscle, but this was not absolute, similar to previous reports (Quartarone et al., 2003). There was a prolongation of CSP and an increase in MEP size of about 33% at 30 min after PAS in the ADM muscle, although this is less than the changes observed in the APB muscle. Our study assessed cortical excitability and cortical inhibition for up to an hour after the end of PAS.

Parkinson's disease and motor cortex plasticity

A novel finding of our study is that both dyskinetic and non-dyskinetic patients had deficient LTP-like effects in the human motor cortex. Dopaminergic deficiency may prevent the motor cortex from changing the strength of synaptic connection when it is primed by a repetitive, low-frequency stimulation. The changes may be similar to the effects of dopaminergic denervation on striatal plasticity that have been characterized in experimental models of parkinsonism in vitro. DA was found to be essential for the induction of striatal LTP and LTD since corticostriatal slices obtained from 6-hydroxydopamine-treated rats did not exhibit LTP (Centonze et al., 1999) and LTD (Calabresi et al., 1992) when they were subjected to high-frequency cortical stimulation. The DA dependency of striatal plasticity may be explained by the tight connection on the striatal medium spiny neurons between dopaminergic transmission and ionotropic glutamatergic transmission (mediated by AMPA and NMDA receptors) originating from motor cortex (Parent and Hazrati, 1995). Several studies have shown that striatal DA regulates glutamate release, modulating the opening, distribution and anchoring to plasma membranes of NMDA and AMPA receptors (Raymond et al., 1993, 1994).

Several mechanisms may account for reduced synaptic plasticity in the motor cortex in patients with Parkinson's disease off medications. An altered pattern of neuronal discharge in the basal ganglia may lead to abnormal plasticity in the motor cortex. Indeed, recent models of the basal ganglia (Bergman and Deuschl, 2002; Brown, 2003) suggest that abnormal, low-frequency (at 8–30 Hz), oscillations within the basal ganglia nuclei might affect motor cortex plasticity itself, preventing facilitation of cortical activity necessary to select and perform an appropriate movement (Brown and Marsden, 1998).

There may also be a direct facilitatory action of dopamininergic terminals on NMDA transmission in the upper cortical layers of the primary motor cortex, similarly to the situation in the striatum. This would be consistent with the idea that PAS induces plastic changes in the upper cortical layers (Wolters et al., 2005) and with the findings that induction of LTP in the pre-frontal cortex in vitro requires DA (Otani et al., 2003) and that in Parkinson's disease there is a reduction of dopaminergic innervations in upper layers of motor and prefrontal cortices (Gaspar et al., 1991).

Another explanation could be related to abnormalities within the pathway that mediates the PAS paradigm. Since repetitive stimulation of somatosensory afferents is crucial to induce dynamic changes in M1, a deficient sensorimotor integration may explain a lack of motor cortex plasticity in Parkinson's disease. However, this is less likely because sensory-motor integration, measured with the technique of short latency afferent inhibition in the human M1, was found to be normal in parkinsonian patients off medications (Sailer et al., 2003).

Dopaminergic medications restore motor cortex plasticity in non-dyskinetic Parkinson's disease patients

In non-dyskinetic patients, treatment with levodopa improved motor symptoms and restored the potentiation of MEP amplitudes induced by PAS in the APB muscle without any loss of topographical specificity, since the ADM excitability was not increased. The finding in the APB muscle is similar to that recently reported by Ueki et al. (Ueki et al., 2006). This is consistent with the hypothesis that DA is necessary for LTP-like changes in the motor cortex. It is not known whether the restoration of motor cortex plasticity by levodopa was related to a direct effect on M1, on the corticocortical connections or on the corticobasal ganglia-cortical motor loop. Since LTP in M1 may support the formation and the storage of motor programmes (Rioult-Pedotti et al., 1998), it is conceivable that restoration of motor cortex plasticity in non-dyskinetic patients is related to a beneficial action of DA on motor memory. A recent paper demonstrated that a single dose of levodopa restored motor memory in elderly subjects to the levels of younger subjects (Floel et al., 2005).

The duration of PAS after-effects on MEP amplitudes was shorter in the non-dyskinetic patients, lasting for up to half an hour, whereas it lasted up to 1 h in the control subjects. This may be related to ‘wearing off’ of the effects of levodopa in some patients towards the end of the experiment.

Deficient motor cortex plasticity in patients with LID

Motor cortex plasticity measured with both MEP amplitude and CSP duration after PAS is reduced in patients with LID in the off medication condition and is not restored by levodopa administration. Since the non-dyskinetic and dyskinetic groups of patients were matched for disease severity and duration, these factors are unlikely to account for the differences between the two groups. Furthermore, our patients did not have cognitive impairment that is usually related to presence of cortical pathology (Braak et al., 2005). Therefore, our results suggest that deficient synaptic plasticity in the motor cortex is associated with the development of LID.

Previous studies have demonstrated that a pathological form of synaptic plasticity in the striatum related to supersensitivity of NMDA receptors could cause the development of atypical motor patterns leading to LID (Chase et al., 2000, 2004). Anatomical and physiological interactions between DA and NMDA receptors on dendritic spines of striatal spiny neurons probably explain why dopaminergic denervation can cause an abnormal corticostriatal synaptic plasticity in Parkinson's disease. Chronic non-physiologic stimulation of DA receptors on striatal neurons (Chase et al., 2000) can induce modifications in NMDA receptors firing and thus the development of aberrant motor patterns leading to motor complications. This has provided the rationale for the use of NMDA receptor antagonists such as amantadine for treating LID in Parkinson's disease (Verhagen et al., 1998). Moreover, a study conducted in corticostriatal slices from rats experiencing LID has shown that dyskinesia is associated with the inability to downregulate LTP in the striatum (Picconi et al., 2003).

It is not possible to rule out the possibility that the changes in dyskinetic patients are secondary to the abnormal involuntary movements. However, the finding that patients with focal hand dystonia, a hyperkinetic movement disorder characterized by excessive muscle contraction (Marsden and Sheehy, 1990), exhibit a very different pattern of excessive increase in cortical excitability and a complete loss of topographical specificity in response to PAS makes this explanation unlikely (Quartarone et al., 2003). Furthermore, dyskinetic patients did not show any difference in the background EMG area compared with the non-dyskinetic group.

The LTP-like plasticity in the motor cortex is probably not related to changes in parkinsonian motor symptoms, since dopaminergic medications induced a similar degree of motor improvement in the non-dyskinetic and dyskinetic groups as measured by motor UPDRS.

Our results suggest that the primary motor cortex may play a role in the development of LID. Deficient LTP-like plasticity may impair the establishment of desired motor patterns, leading to the formation of undesired motor patterns seen in LID. The lack of LTP-like plasticity in the M1 may be secondary to altered patterns of firing within the basal ganglia loop associated with LID, reflecting abnormal synchronized activities between motor cortex and basal ganglia nuclei. Local field potential studies in patients with Parkinson's disease have found that LID is associated with decreased oscillatory activity below 30 Hz in the pallidum (Silberstein et al., 2005) and decreased synchronization of 8–30 Hz activity between the STN and the pallidum (Foffani et al., 2005).

Alternatively, altered corticostriatal input may be a factor in inducing aberrant plasticity in the striatum. This raises an intriguing possibility that deficient motor cortex plasticity, in spite of levodopa treatment, may constitute an endophenotypic trait predisposing to and causing LID. There are powerful connections from the frontal motor areas to the STN, originating from the axon collaterals of pyramidal tract neurons and forming the ‘hyperdirect pathway’. These projections enhance activity in the STN and may promote inhibition of competing motor programmes at the onset of a movement (Nambu et al., 2002). It is conceivable that a hypo-active or abnormal pattern in cortico-subthalamic pathway could be responsible for LID, with a loss of inhibition of unintended motor programmes and production of purposeless, involuntary movements. Further studies assessing the pattern of motor cortex plasticity in ‘de novo’ patients with Parkinson's disease longitudinally may shed light on this issue.

Intracortical inhibition and motor cortex plasticity in Parkinson's disease

The early part of the CSP is partly related to changes in spinal excitability and the later part of the CSP is due to changes in cortical excitability. Therefore, the CSP is widely used as a measure of cortical inhibition. MEP and CSP are mediated by different circuits in M1. MEP amplitude reflects the excitability of the corticospinal system, whereas CSP is influenced by inhibitory cortical interneurons and is probably mediated by the activation of post-synaptic GABA-B receptors (Werhahn et al., 1999). The prolongation of CSP in control subjects is similar to previous studies, and is probably due to facilitation of inhibitory interneurons (Stefan et al., 2000). Similar to previous reports, we confirmed that CSP duration is shortened in Parkinson's disease and is prolonged by levodopa treatment (Priori et al., 1994).

In addition, there was an interesting difference between non-dyskinetic and dyskinetic patients. In the non-dyskinetic patients off medication, PAS induced a modest but significant increase in CSP duration in the heterotopic ADM muscle. In the homotopic APB muscle, there was also a modest but non-significant prolongation of CSP duration, comparable with the effect observed in control subjects. In contrast, dyskinetic patients showed no prolongation of CSP by PAS both off and on medication in ABP and ADM. The failure of PAS to prolong CSP in dyskinetic patients both on and off medications further suggests that there is a deficient cortical inhibitory system in the motor cortex in this group of patients and this may contribute to LID.

Conclusions

Parkinson's disease is associated with deficient plasticity in the human motor cortex induced by the PAS protocol. Abnormal synaptic plasticity in the motor cortex may play a role in the development of LID.

References

View Abstract