Intermittent theta burst stimulation over primary motor cortex enhances movement-related beta synchronisation

Clinical Neurophysiology 122 (2011) 2260–2267

Contents lists available at ScienceDirect

Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph

Intermittent theta burst stimulation over primary motor cortex enhances movement-related beta synchronisation Ya-Fang Hsu a, Kwong-Kum Liao b, Po-Lei Lee a,c, Yun-An Tsai d, Chia-Lung Yeh c, Kuan-Lin Lai e, Ying-Zu Huang f,g, Yung-Yang Lin a,b, I-Hui Lee a,b,⇑ a

Institute of Brain Science, National Yang-Ming University, Taipei, Taiwan Department of Neurology, Taipei Veterans General Hospital, Taipei, Taiwan National Central University, Taoyuan County, Taiwan d Center for Neural Regeneration, Taipei Veterans General Hospital, Taiwan e Department of Neurology, Taipei Municipal Gandau Hospital, Taipei, Taiwan f Chang-Gung Memorial Hospital, Taoyuan County, Taiwan g Chang Gung University College of Medicine, Taoyuan County, Taiwan b c

a r t i c l e

i n f o

Article history: Accepted 9 March 2011 Available online 4 May 2011 Keywords: Theta burst stimulation (TBS) Transcranial magnetic stimulation (TMS) Event-related synchronisation (ERS) Event-related desynchronisation (ERD) Magnetoencephalography (MEG)

h i g h l i g h t s  Theta burst stimulation with doubled number of stimuli (TBS1200) modulates motor cortical excitability in parallel with the conventional protocols, with modestly prolonged efficacy.  Facilitatory intermittent theta burst stimulation (iTBS1200) alters motor task-dependent beta frequency oscillation.  Prolonged theta burst stimulation (TBS1200) may be suited to enhance the stability of its effect on motor cortical excitability.

a b s t r a c t Objective: The objective of this study is to investigate how transcranial magnetic intermittent theta burst stimulation (iTBS) with a prolonged protocol affects human cortical excitability and movement-related oscillations. Methods: Using motor-evoked potentials (MEPs) and movement-related magnetoencephalography (MEG), we assessed the changes of corticospinal excitability and cortical oscillations after iTBS with double the conventional stimulation time (1200 pulses, iTBS1200) over the primary motor cortex (M1) in 10 healthy subjects. Continuous TBS (cTBS1200) and sham stimulation served as controls. Results: iTBS1200 facilitated MEPs evoked from the conditioned M1, while inhibiting MEPs from the contralateral M1 for 30 min. By contrast, cTBS1200 inhibited MEPs from the conditioned M1. Importantly, empirical mode decomposition-based MEG analysis showed that the amplitude of post-movement beta synchronisation (16–26 Hz) was significantly increased by iTBS1200 at the conditioned M1, but was suppressed at the nonconditioned M1. Alpha (8–13 Hz) and low gamma-ranged (35–45 Hz) rhythms were not notably affected. Movement kinetics remained consistent throughout. Conclusions: TBS1200 modulated corticospinal excitability in parallel with the direction of conventional paradigms with modestly prolonged efficacy. Moreover, iTBS1200 increased post-movement beta synchronisation of the stimulated M1, and decreased that of the contralateral M1, probably through interhemispheric interaction. Significance: Our results provide insight into the underlying mechanism of TBS and reinforce the connection between movement-related beta synchronisation and corticospinal output. Ó 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction ⇑ Corresponding author. Address: Department of Neurology, Taipei Veterans General Hospital, No. 201, Sec. 2, Shih-Pai Road, Taipei 112, Taiwan. Tel.: +886 2 28712121x8109; fax: +886 2 28757579. E-mail address: [email protected] (I-H Lee).

Theta burst stimulation (TBS) is a form of repetitive transcranial magnetic stimulation (rTMS) consisting of bursts of three pulses at 50 Hz repeated at 5 Hz. Depending on different stimulation modalities, intermittent and continuous TBS (iTBS and cTBS) lead to

1388-2457/$36.00 Ó 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2011.03.027

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

facilitation or inhibition of cortical excitability, respectively (Agostino et al., 2008; Di Lazzaro et al., 2008a; Huang et al., 2005, 2009). The TBS form of TMS is adopted from a protocol with the same name that has long been used to induce synaptic long-term potentiation (LTP) in animal brain slices (Hess and Donoghue, 1996; Larson and Lynch, 1986; Vickery et al., 1997). The modulation by TBS is likely to involve the mechanisms of (LTP- or long-term depression (LTD)-like synaptic plasticity. This is supported by a pharmacological study showing that the effects of iTBS and cTBS were N-methyl-D-aspartic acid (NMDA) receptor-dependent (Huang et al., 2007), which is a critical component of neuroplasticity. Other studies using spinal epidural recordings have shown that facilitatory iTBS over the primary motor cortex (M1) increases later indirect (I) waves of the corticospinal volleys (Di Lazzaro et al., 2008a), whereas inhibitory cTBS suppresses the earliest I1-wave (Di Lazzaro et al., 2005), suggesting that iTBS and cTBS may affect different cortical targets, and, in turn, modulate corticospinal output. The I-waves are primarily derived from gamma-aminobutyric acid (GABA)-related interneuronal circuits within the human motor cortex (Ziemann et al., 1998). An increased concentration of GABA rather than glutamate/glutamine has been demonstrated by cTBS over human M1 using magnetic resonance (MR) spectroscopy (Stagg et al., 2009). However, the mechanisms underlying the effects of TBS on human M1 are largely unclear. To explore how TBS modulates intracortical motor circuitry, we adopted non-invasive and highly sensitive magnetoencephalography (MEG) to observe voluntary movement-related cortical oscillations, in particular, event-related desynchronisation (ERD) and event-related synchronisation (ERS) identified at the rolandic area (Pfurtscheller and Lopes da Silva, 1999). Pre-movement alpha ERD (8–13 Hz) has been implicated in somatosensory and motor planning (Pfurtscheller and Lopes da Silva, 1999); post-movement beta ERS (16–26 Hz) has been implicated in corticomuscular output (Baker et al., 1999; Schaefer et al., 2006) and deactivation/ inhibition of motor networks (Pfurtscheller et al., 1996); and perimovement gamma ERS (a broad range 35–150 Hz) has been associated with motor programming (Crone et al., 1998a) and output muscle activity (Pfurtscheller et al., 1993; Salenius et al., 1996). In primate M1, beta synchronous oscillation is characteristically observed by local field potential recording (Murthy and Fetz, 1992; Sanes and Donoghue, 1993). It is unknown whether beta oscillation will be affected by TBS given to the M1, and what the functional relevance will be, if so. In the present study, we doubled the stimulation duration of TBS to 1200 pulses per session and hypothesised that TBS1200 over

2261

the M1 might induce longer-lasting effects on cortical excitability, as compared with the original TBS paradigms. Our results demonstrated lasting facilitatory effects of iTBS1200 on the corticospinal excitability, in association with a so far unreported increase of movement-related beta range synchronisation at the conditioned M1. Opposite changes were observed at the nonconditioned M1 through interhemispheric interaction. Our results shed light on understanding the mechanisms of intracortical modulation by iTBS. 2. Methods 2.1. Subjects and study design Ten healthy right-handed volunteers (male: female = 5:5, mean age: 29.0 ± 7.1 years) were included for the TMS study of iTBS, cTBS and sham stimulation, among whom five participated in the MEG study too; the others were not eligible due to metallic dental implants that could have interfered with the MEG recording. Ten age-matched subjects (male: female = 4:6, mean age: 28.4 ± 3.7 years) were selected for the MEG study. In total, 15 subjects were included, none of whom had any history of neurological illness or use of neuromodulatory medications. The study was approved by the Ethics Committee of Taipei Veterans General Hospital (VGHIRB 97-08-01A), and consent forms were signed in advance by all subjects. The schematic experimental design is shown in Fig. 1. For the TMS study, the volunteers were placed in the supine position. The abductor pollicis brevis (APB) muscle was chosen as the target muscle for localisation of the M1 hand region and for measurement of corticospinal excitability. The initial resting motor threshold (rMT), active motor threshold (aMT) and motor-evoked potentials (MEPs) were measured (details see later). Baseline MEPs were recorded every 10 min for 30 min, followed by TBS. Each subject underwent four different TBS conditions in random order on separate days (at least 5 days from each other), including iTBS, cTBS and sham stimulation over the left M1 for MEPs evoked from the conditioned M1 (iTBS-conditioned, cTBS-conditioned and sham stimulation), and an additional session of iTBS for MEPs evoked from the contralateral M1 (iTBS-nonconditioned). After TBS, MEPs were recorded every 5 min for at least 30 min. We extended the period of measurement for iTBS-conditioned M1 up to 60 min to confirm when the effects had returned to baseline. For the MEG study, the baseline movement-related cortical oscillations (pre-test) were measured for each subject. They then received iTBS and sham

Fig. 1. Schematic study design. In Experiment 1, TMS, resting motor threshold (rMT) and active motor threshold (aMT) of the right abductor pollicis brevis (APB) were first determined, and the baseline MEPs were averaged from 12 single-pulse TMS at an intensity of 120% rMT for three times every 10 min. Afterwards, four conditions of TBS each consisting of 1200 pulses at an intensity of 80% aMT were given on different days to the left M1. The end of TBS was defined as 0 and the time course of MEP changes was evaluated for 30–60 min. In Experiment 2, MEG, a pre-test of task-dependent MEG for unilateral self-paced index-finger lifting was recorded. Afterwards, the subjects were evaluated for aMT and given either iTBS or sham stimulation to the left M1 on different days. The end of TBS was defined as 0 and its effect on sensorimotor oscillation was immediately evaluated by a repeated task-dependent MEG measurement.

2262

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

stimulation over the left M1 in random order on different days, and an immediate MEG follow-up. 2.2. TBS, safety monitoring and corticospinal excitability The basic TBS pattern is a burst containing three pulses of 50 Hz given at 5 Hz frequency. In this study, iTBS was composed of trains of basic TBS for 2 s (i.e., 10 bursts) given every 10 s for a total of 1200 pulses (iTBS1200), and cTBS was composed of continuous trains of basic TBS for a total of 1200 pulses (cTBS1200). Both were given at an intensity of 80% aMT (Huang et al., 2005). aMT was defined as the minimum intensity required to evoke MEPs of more than 200 lV peak-to-peak amplitude in five out of 10 trials while the subject maintained a voluntary contraction of nearly 20% of the maximum monitored by visual feedback. The TBS was delivered using a Magstim Rapid2 simulator (Magstim Co., Whitland, UK) with a 70-mm figure-of-eight coil. Surface electromyography (EMG) was recorded from the APB muscle with Ag–AgCl electrodes and a band-pass filter from 3 Hz to 10 kHz (Synergy MEG/EP system). TBS or sham stimulation was applied to the hand area of the left M1 (conditioned hemisphere). The hand-representation area was identified by applying single-pulse TMS to elicit the maximal peak-to-peak amplitude of the MEPs in the APB muscle. The coil was held in a posterolateral direction, tangentially to the scalp to induce a posterior–anterior current direction in the brain, which has been shown to induce better iTBS effects (Talelli et al., 2007). The sham stimulation was given as per the iTBS, except that the coil was held perpendicularly to the scalp. Subjects were asked to relax with their eyes open during the intervention. To monitor the safety of the prolonged TBS protocol, we continuously recorded muscle action potentials of the APB by using surface EMG throughout the experiment, including when TBS was being given, and documented subjects’ self-reports of any adverse effects, such as local pain, headache, fatigue, dizziness or seizures after TBS (Rossi et al., 2009). Corticospinal excitability was evaluated using single-pulse TMS generated by a Magstim Rapid2 simulator. rMT was defined as the minimum intensity required to elicit MEPs of more than 50 lV peak-to-peak amplitude in five out of 10 trials at rest (Rothwell et al., 1999). At each time point, 12 successive MEPs evoked every 5 s for 1 min, by an intensity of 120% baseline rMT, were averaged. 2.3. MEG measurements We used a 306-channel whole-head neuromagnetometer (band-pass: 0.05–250 Hz; digitised at 1 kHz; Vectoview; Neuromag Ltd., Helsinki, Finland) to detect neural activity elicited by unilateral self-paced finger lifting in a simple reaction-time paradigm. Subjects were asked to briskly lift the index finger every 7 s with a range of motion of around 35–40° with their eyes open to suppress the occipital alpha rhythm. Each subject executed unilateral indexfinger lifting for two repeated sessions of 50 trials (in total 100 trials for each hand; 45 min for both hands) in a seated position, monitored by surface EMG (digitised at 1 kHz) recorded at the extensor digitorum communis. All subjects learned to consistently perform self-paced finger lifting before data acquisition. Electrooculograms were recorded from above the right orbit and below the left orbit to exclude trials with eye movements and blinks (over 100 lV). The head position with respect to the MEG sensor array was determined by measuring magnetic signals from four headposition-indicator coils placed on the scalp. Coil positions were identified by a three-dimensional (3-D) digitiser with respect to three predetermined landmarks: one nasion and two preauricular points. These data were used to superimpose MEG source signals onto individual MR imaging (MRI) images. Nearly 60 trials with explicit EMG activities were selected from a total of 100 trials, as

good epochs for subsequent signal processing. For consistent movement monitoring by surface EMG, the integral values of peak-to-peak amplitudes of compound muscle action potentials during –0.2 to 0.2 s relative to the movement onset (0 s) and the mean duration were averaged at pre-test, sham and iTBS conditions. Epochs comprised of MEG data from –4 to 3 s were segmented without applying any pre-filtering. We adopted the empirical mode decomposition approach (Lee et al., 2009) to extract single-trial noise-suppressed sensorimotor oscillatory activity. Two channels in the vicinity of the left and right sensorimotor areas were chosen as the left-hemisphere and righthemisphere channels of interest (LHCOIs and RHCOIs), respectively. Good epochs recorded from the LHCOIs and RHCOIs were decomposed into a series of intrinsic mode functions (IMFs) using empirical mode decomposition. A spatial map was created from each IMF by computing its correlation with raw data from all other channels. The sensorimotor-related IMFs, which had highly focussed spatial weights on left and right primary sensorimotor areas, were chosen to reconstruct noise-suppressed sensorimotor oscillatory activities (Lee et al., 2009). The noise-suppressed sensorimotor oscillatory activities were further band-pass filtered in alpha (8–13 Hs), beta (16–26 Hs) and low gamma (35–45 Hz) bands, determined by two Fourier spectra obtained from baseline and reactive periods (Pfurtscheller and Lopes da Silva, 1999). The baseline period was the same period for all frequency bands (–3 to –2 s), whereas the reactive periods were chosen differently for ERD (–0.5 to 0.5 s) and ERS (0.5– 1.5 s) computations. Task-specific frequencies in the alpha, beta and low gamma bands were determined by subtracting the Fourier spectrum of the baseline period from that of the reactive periods. Only significant frequencies (above 95% confidence, i.e., Z > 2.0, i < 0.05) around the alpha, beta and low gamma bands were selected as task-related oscillations for the subsequent ERD and ERS calculations (Lee et al., 2009; Pfurtscheller and Lopes da Silva, 1999). To estimate the sources of task-related oscillatory activities, we used the Neuromag software system (Vectoview; Neuromag Ltd., Helsinki, Finland) with spatiotemporal equivalent current dipoles (ECDs) modelling analysis (Hamalainen et al., 1993). A single dipole model was applied for each hemisphere to estimate the neural source at maximum activity in alpha, beta and low gamma task-related oscillatory activities by selecting the MEG sensors over the left and right sensorimotor areas. Individual anatomical high-resolution T1-weighted MR images (3.0-T Bruker MedSpec S300 system, Bruker, Karlsruhe, Germany; gradient-echo pulse sequence, TR/TE/TI = 88.1 ms/4.12 ms/650 ms, 128  128  128 matrix, field of view (FOV) = 250 mm) were acquired, and the MEG source signals were superimposed. Only the dipoles showing >80% goodness of fit were rendered onto individual MR images (Jensen and Vanni, 2002). The task-related oscillatory activity was quantified using the amplitude modulation (AM) method (Clochon et al., 1996):

AðtÞ ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi xðtÞ2 þ HðxðtÞÞ2 ;

where A(t) is the AM waveform of task-related oscillatory activity x(t) and H() denotes the Hilbert transformation operator. The vector sum of AM waveform (VAMW) at each sensor unit was calculated by computing the square root of AM waveforms in each qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi gradiometer pair; that is, Vði; tÞ ¼ mx ði; tÞ2 þ my ði; tÞ2 , where Vði; tÞ is the VAMW at the ith sensor unit. The VAMWs of task-related oscillatory activities filtered within task-specific alpha, beta and low gamma bands were computed and denoted as VAMWalpha, VAMWbeta and VAMWgamma, respectively. To evaluate the performance of the extracted task-related oscillatory activities, the alpha ERD in VAMWalpha, beta ERS in VAMWbeta and low gamma ERS in

2263

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

VAMWgamma were detected, in which alpha ERD was defined as the amplitude difference between the mean amplitude in the baseline period (Ampbaseline_alpha) and the alpha amplitude valley (minimal amplitude) within the movement onset period (Ampvalley_alpha) in VAMWalpha, (i.e., alpha ERD = Ampbaseline_alpha  Ampvalley_alpha); the beta ERS was defined as the amplitude difference between the mean amplitude in the baseline period (Ampbaseline_beta) in VAMWbeta and the beta peak amplitude within the post-movement period (Amprebound_beta) (i.e., beta ERS = Amprebound_beta  Ampbaseline_beta); and the low gamma ERS was defined as the amplitude difference between the mean amplitude in the baseline period (Ampbaseline_gamma) in VAMWgamma and the gamma peak amplitude within the post-movement period (Amprebound_gamma) (i.e., gamma ERS = Amprebound_gamma  Ampbaseline_gamma). 2.4. Statistical analysis We used the Statistical Package for Social Sciences (SPSS) software package (Windows version 16.0.0, Chicago, IL, USA) for all data analyses. A two-way repeated measures analysis of variance (ANOVA) test was performed to evaluate the changes of MEP amplitude (as absolute values) among the four TBS conditions, that is, iTBS-conditioned, cTBS-conditioned, iTBS-nonconditioned and sham stimulation (the between-subject factor was ‘TBS CONDITION’; and the within subject factor ‘TIME’ at pre-test and post-0, 5, 10, 15, 20, 25 and 30 min), followed by a post hoc Bonferroni’s analysis. For the time course of amplitude change of MEPs (as absolute values) following individual TBS conditions, a one-way repeated measures ANOVA test was performed (the within subject factor was ‘TIME’ at pre-test and post-0, 5, 10, 15, 20, 25 and 30 min; prolonged observation only for the iTBS-conditioned hemisphere: additional post-40, 50 and 60 min), followed by post hoc paired t-tests with Bonferroni’s correction between individual time points and the baseline. To compare changes in alpha ERD, beta ERS and low gamma ERS (as normalised values to the baseline) among pre-test, iTBS and sham stimulation, a one-way ANOVA test was performed followed by a post hoc Bonferroni’s analysis. 3. Results

Fig. 2. Temporal changes in MEPs after theta burst stimulation. iTBS of 1200 pulses significantly facilitated the MEP size evoked from the conditioned M1 (iTBSconditioned, j, p < 0.001), whereas inhibited that from the nonconditioned M1 (iTBS-nonconditioned, h, p < 0.001). Conversely, cTBS of 1200 pulses inhibited the MEP size from the conditioned M1 (cTBS-conditioned, , p < 0.01). No significant change of MEP size after sham stimulation (sham stimulation, , p = 0.795). The above facilitatory and inhibitory effects lasted for 30 min. The error bars represent standard error of means.

ANOVA test followed by post hoc Bonferroni’s analysis. The MEPs from iTBS-conditioned hemispheres were significantly facilitated for 30 min and gradually recovered to the baseline by 50 min ((F(10, 80) = 5.725; p < 0.001); p < 0.05 at post-iTBS 0, 5, 10, 15, 20, 25 and 30 min and p > 0.05 at 40, 50 and 60 min, respectively, as compared with the baseline). In parallel to the facilitatory effect at the iTBS-conditioned hemispheres, there was an inhibitory effect at the contralateral hemispheres lasting for 30 min. Similarly, the inhibitory effect at cTBS-conditioned hemispheres lasted for 30 min ((F(7, 63) = 22.289; p < 0.001; F(7, 63) = 24.207; p < 0.001); p < 0.05 at post-cTBS 0, 5, 10, 15, 20, 25 and 30 min, respectively, as compared with the baseline). There was no change in MEPs following sham stimulation (F(7, 63) = 0.548; p = 0.795). For safety monitoring, there was no muscle action potential recorded from surface EMG during any TBS. According to the subjects’ self-reports, there were no significant adverse effects after receiving TBS, except for mild local pain in three and transient

3.1. Effects of TBS1200 on the corticospinal excitability Four conditions of TBS (iTBS-conditioned, cTBS-conditioned, iTBS-nonconditioned and sham), each consisting of a total of 1200 pulses at an intensity of 80% aMT, were given to the left M1 of the same subjects on different days. There was a significant effect of ‘TBS CONDITION’ on MEP amplitude following stimulation ((F(3, 27) = 34.995; p < 0.001) with marked post hoc difference between each TBS condition and sham stimulation (p < 0.05, respectively, Fig. 2). The MEPs from the iTBS-conditioned hemispheres were facilitated, whereas those from the cTBS-conditioned hemispheres were suppressed. Notably, the MEPs from the iTBS-nonconditioned M1 were suppressed to a similar degree and there was no obvious difference in MEP size between those from the cTBS-conditioned and iTBS-nonconditioned hemispheres (‘TBS CONDITION’ (F(1, 9) = 0.066, p = 0.803); ‘TBS CONDITION’  ‘TIME’ interaction (F(7, 63) = 1.630, p = 0.143)). There was also a significant effect of the ‘TIME’ factor (F(7, 63) = 8.053; p < 0.001) and significant ‘TBS CONDITION’  ‘TIME’ interaction (F(21, 189) = 9.571; p < 0.001) on MEP amplitude following stimulation. The baseline rMT, aMT and MEPs had no notable differences among any TBS or sham conditions (Table 1a). The duration of after-effects after each individual ‘TBS CONDITION’ was further examined by a one-way repeated measures

Table 1 (a) Baseline TMS parameters. (b) Finger lifting monitored by surface EMG in MEG experiments. TBS condition

rMT (%)

aMT (%)

MEPs (mV)

(a) iTBS-conditioned Sham stimulation cTBS-conditioned iTBS-nonconditioned

42.9 ± 4.6 42.2 ± 4.1 42.2 ± 3.3 40.9 ± 2.9

35.8 ± 3.2 37.3 ± 3.9 37.1 ± 2.2 –

0.97 ± 0.07 1.01 ± 0.12 1.02 ± 0.07 1.00 ± 0.09

Condition, side of finger lifting

Power (V  s)

Duration (s/trial)

(b) Pre-test, right Pre-test, left iTBS, right iTBS, left Sham, right Sham, left

1.117 ± 0.006 1.118 ± 0.005 1.121 ± 0.008 1.119 ± 0.009 1.121 ± 0.007 1.123 ± 0.024

6.9 ± 1.1 7.2 ± 0.4 7.1 ± 0.9 6.9 ± 0.7 6.8 ± 0.6 6.9 ± 0.8

All values are mean ± S.D. rMT and aMT: resting and active motor threshold given as percentage of maximum stimulator output (%). MEPs: motor-evoked potentials at an intensity of 120% rMT. Power: the integral values of peak-to-peak amplitude of compound muscle action potentials during 0.2 to 0.2 s relative to the movement onset (0 s).

2264

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

Fig. 3. Representative spatial cortical oscillations from magnetoencephalography during unilateral finger lifting. A healthy subject performed (A) right and (B) left indexfinger lifting at pre-test and after sham stimulation or iTBS over the left primary motor cortex. Note the movement-related oscillation is distributed over the contralateral sensorimotor cortex mainly in alpha (8–13 Hz), beta (16–26 Hz) and low gamma (35–45 Hz) rhythms. Colour bars show the degree of correlation coefficient value with either left or right sensorimotor channel of interest. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Representative source localisation of the sensorimotor oscillations. Sources of the maximum cortical activities during right index-finger lifting were superimposed onto the anatomical T1-weigted MRI in a healthy subject. The dipoles of alpha event-related desynchronisation (ERD) are located at left post-central gyrus (primary somatosensory cortex), beta event-related synchronisation (ERS) at left pre-central gyrus (primary motor cortex), and low gamma ERS at left peri-central sulcus (d: alpha ERD; N: beta ERS; j: gamma ERS). Blue arrows indicate the central sulcus. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

2265

Fig. 5. Temporal cortical oscillations at primary sensorimotor cortex during unilateral finger lifting. (A) Averaged temporal waveforms of alpha, beta and low gamma bands from the primary sensorimotor cortex of 10 healthy subjects. After iTBS over the left M1, the amplitude of beta event-related synchronisation (ERS) was increased markedly at the conditioned M1 (p < 0.01), while decreased at the nonconditioned M1 (p < 0.01) compared to pre-test or sham conditions (B), suggesting interhemispheric inhibition comparable to the changes of MEPs. No notable differences in alpha or lower gamma rhythms. The amplitude is normalised to the baseline period (3 to 2 s, 0 as onset of finger lifting). The error bars represent standard error of means.

headache in one out of a total of 15 subjects, suggesting that the new TBS1200 protocols are safe in healthy volunteers. 3.2. Sources of sensorimotor oscillations from MEG Movement-related oscillations during unilateral index-finger lifting were consistently and predominantly distributed in the contralateral sensorimotor cortex. Task-specific ERD and ERS responses were in alpha (8–13 Hz), beta (16–26 Hz) and low gamma (35–45 Hz) frequencies. High gamma (over 60 Hz) rhythm was barely detected (lower than 10 fT cm–1) in our MEG study (data not shown). After a single session of iTBS1200 or sham stimulation over the left M1, movement-related magnetic fields were not substantially altered from the pre-test condition in terms of spatial distribution in either hemisphere (Fig. 3). The sources of the maximum movement-related oscillations from the 10 healthy subjects were superimposed onto individual anatomical T1-weighted MR images (Fig. 4). The pre-test alpha ERD at the post-central gyrus (x, y, z = –44.2 ± 3.5, –20.4 ± 4.6, 95.5 ± 5.6 mm; goodness of fit = 93.1%), the beta ERS at the pre-central gyrus (x, y, z = –39.6 ± 3.3, –5.4 ± 6.6, 90.5 ± 3.6 mm; goodness of fit = 89.1%) and the low gamma ERS at the peri-central sulcus (x, y, z = 41.5 ± 6.3, –15.4 ± 7.6, 98.5 ± 5.3 mm; goodness-of-fit = 82.8%), were all compatible with previously reported locations (Jurkiewicz et al., 2006). The head coordinate system (x, y, z) for dipole location was anchored by the head-position indicator. The positive x-axis points to the right hemisphere through the bilateral preauricular points; the positive y-axis points anteriorly traversing the nasion and perpendicular to the x-axis; and the positive z-axis points upwards and perpendicularly to the x–y plane. 3.3. Intermittent TBS-enhanced beta synchronisation Temporal waveforms (–3 to 3 s, 0 s as movement onset) of the movement-related rhythms (alpha, beta and low gamma bands) were averaged from the 10 healthy volunteers in three conditions: pre-test, after iTBS and sham stimulation (Fig. 5). The mean peak latency of alpha ERD was at –0.13 ± 0.21 s, low gamma ERS at 0.36 ± 0.28 s and beta ERS at 1.61 ± 0.13 s. After iTBS1200 over the left M1, the ipsilateral beta ERS amplitude was significantly enhanced as compared with those in pre-test and after sham stimulation (increased 20% to pre-test condition, p < 0.01, respectively),

but contralateral beta ERS amplitude was suppressed compared with those in pre-test and after sham stimulation (decreased 20% to pre-test condition, p < 0.01, respectively). The peak latency of beta ERS was not affected, nor were the latency and amplitude of alpha ERD or low gamma ERS. The surface EMG monitoring during finger lifting showed that the mean integral values of peakto-peak amplitude of compound muscle action potentials and duration of movements were not notably different among the three conditions (Table 1b), probably because we asked all subjects to perform stable movement throughout. Therefore, the changes of beta ERS likely resulted from iTBS rather than by performance variation. 4. Discussion We showed that iTBS1200 produced significant and lasting facilitation of corticospinal excitability at the conditioned hemisphere for 30 min or so, whereas it suppressed that at the nonconditioned hemisphere via interhemispheric inhibition. Conversely, cTBS1200 significantly suppressed corticospinal excitability at the conditioned hemisphere for 30 min. Sham iTBS1200 did not affect MEPs. Importantly, the facilitatory effects of iTBS1200 were associated with enhancement of post-movement beta ERS at the conditioned hemisphere, whereas they were associated with suppression of post-movement beta ERS at the contralateral hemisphere. By contrast, alpha ERD and low gamma ERS were not changed. Our results are contrary to those of a recent study suggesting that the facilitatory effects of iTBS and inhibitory effects of cTBS were converted into inhibitory and facilitatory effects, respectively, when the stimulation duration was doubled from 600 to 1200 pulses (Gamboa et al., 2010). We did not observe such reverse after-effects of iTBS1200 and cTBS1200. Instead, iTBS1200 produced a modestly prolonged facilitatory after-effect in concordance with ordinary iTBS600 (Huang et al., 2005). We can only speculate upon reasons for the contradictory results. A major difference we found between the protocols of Gamboa’s and our study was that they acquired isometric contraction of the target muscle for aMT measurement shortly (5 min) before TBS, whereas we waited for 30 min to give TBS. In addition, we measured aMT for 2 min, a shorter duration than the 5 min they used. It has been demonstrated that contraction immediately before (Gamboa et al.,

2266

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267

2010; Iezzi et al., 2008) or after (Huang et al., 2008) TBS can reverse its usual effect, probably through rapid polarity-reversing metaplasticity. Different durations of the break before TBS and of the target muscle contraction time may explain the different results. In addition, we found that iTBS1200 induced transcallosal inhibition at the nonconditioned hemisphere, which is compatible with a previous study using iTBS with 600 pulses (Suppa et al., 2008). This also supports the fact that iTBS1200 facilitates the conditioned hemisphere as in ordinary iTBS. Moreover, our MEG results of iTBS1200 showed opposite findings to those of suppressive 1-Hz rTMS (see later), which indirectly suggests a facilitatory effect of iTBS1200. Lastly, previous reports have demonstrated dose-dependent effects on corticospinal excitability by facilitatory 5-Hz rTMS (1800 vs. 150 pulses at the same stimulation intensity) (Peinemann et al., 2004) and inhibitory cTBS (600 vs. 300 pulses at the same intensity) (Huang et al., 2005). However, the dose-dependent effect of different TBS durations was not the purpose of this study. More comprehensive investigations are required to elucidate the full range of effects of TBS1200 on cortical excitability. Intriguingly, we found an increased amplitude of beta ERS at the iTBS-conditioned M1 only after iTBS but not sham stimulation, whereas a decreased amplitude of beta ESR at the iTBS-nonconditioned M1 was observed compared with the pre-test condition. Concerned with the possible impact of performance variation on cortical oscillation, we asked the subjects to practice consistent finger-lifting tasks throughout the MEG experiment in terms of movement velocity and force monitored with surface EMG. It is therefore less likely that the change of beta ERS was attributed to a change in motor performance. The finger-lifting task has been reported to not be affected by iTBS given over the M1 in terms of movement amplitude, peak velocity and peak acceleration (Agostino et al., 2008). Our results suggest that iTBS alters motor task-dependent oscillation at M1 where the maximum beta ERS is located, and further induces interhemispheric interaction (Reis et al., 2008). In concordance with our findings, inhibitory 1-Hz rTMS over M1 has been shown to significantly reduce beta ERS in association with decreased corticospinal excitability in healthy subjects (Tamura et al., 2005). Furthermore, our findings are in line with other spinal epidural recording studies, which showed that 1Hz rTMS suppressed the later I-waves (Di Lazzaro et al., 2008b), whereas iTBS facilitated them (Di Lazzaro et al., 2008a). Whether enhancement of beta ERS may reflect activated or augmented motor circuitry through transsynaptic activation by iTBS, and whether beta ERS may be linked with later I-waves or corticospinal output require further studies to elucidate. On the other hand, iTBS did not seem to influence alpha ERD or low gamma ERS. This is probably because iTBS was targeted at M1 (source of maximum beta ERS), rather than at the primary sensory cortex (source of maximum alpha ERD) or peri-central sulcus (source of maximum low gamma ERS). iTBS over M1 has been reported to have no effect on sensory evoked potentials (Katayama and Rothwell, 2007). In addition, the simple finger-lifting task may not involve much sensorimotor planning or motor programming in which alpha ERD or gamma ERS is implicated (Crone et al., 1998b; Pfurtscheller and Lopes da Silva, 1999). However, we are not sure whether iTBS affected high gamma oscillation (over 60 Hz) at M1 because the signals were hardly detected in our MEG analysis (lower than 10 fT cm–1). Furthermore, it might be possible that our surface EMG monitoring did not disclose subtle changes of kinematics in the three conditions. In future studies, it might be interesting to explore the isolated effect of iTBS on spontaneous cortical oscillation by analysing resting state MEG (de Pasquale et al., 2010; Liu et al., 2010). In summary, we demonstrated robust and lasting after-effects of TBS with 1200 pulses in parallel with the direction of those seen with the ordinary paradigms with 600 pulses. The facilitatory iTBS

significantly increased post-movement beta synchronisation at the iTBS-conditioned M1, and decreased that at the nonconditioned M1, probably via interhemispheric inhibition. Alpha ERD, low gamma ERS, latencies or source locations of the maximum ERD/ ERS were not affected by iTBS. Our results provide an insight into the underlying mechanism of TBS and suggest a correlation between beta ERS and corticospinal excitability. Acknowledgements This work was supported by National Science Council (NSC-972314-B-480-002-MY3 and 97-3111-B-010-004), Taipei Veterans General Hospital (V98B2-008, V98E1-009, V99C1-069 and V99E1-010) and National Health Research Institutes (NHRI-EX999913EC) in Taiwan. We have no conflict of interest to disclose. References Agostino R, Iezzi E, Dinapoli L, Suppa A, Conte A, Berardelli A. Effects of intermittent theta-burst stimulation on practice-related changes in fast finger movements in healthy subjects. Eur J Neurosci 2008;28:822–8. Baker SN, Kilner JM, Pinches EM, Lemon RN. The role of synchrony and oscillations in the motor output. Exp Brain Res 1999;128:109–17. Clochon P, Fontbonne J, Lebrun N, Etevenon P. A new method for quantifying EEG event-related desynchronization: amplitude envelope analysis. Electroencephalogr Clin Neurophysiol 1996;98:126–9. Crone NE, Miglioretti DL, Gordon B, Lesser RP. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Eventrelated synchronization in the gamma band. Brain 1998a;121:2301–15. Crone NE, Miglioretti DL, Gordon B, Lesser RP. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Eventrelated synchronization in the gamma band. Brain 1998b;121(Pt. 12):2301–15. de Pasquale F, Della Penna S, Snyder AZ, Lewis C, Mantini D, Marzetti L, et al. Temporal dynamics of spontaneous MEG activity in brain networks. Proc Natl Acad Sci USA 2010;107:6040–5. Di Lazzaro V, Pilato F, Dileone M, Profice P, Oliviero A, Mazzone P, et al. The physiological basis of the effects of intermittent theta burst stimulation of the human motor cortex. J Physiol 2008a;586:3871–9. Di Lazzaro V, Pilato F, Dileone M, Profice P, Oliviero A, Mazzone P, et al. Low frequency rTMS suppresses specific excitatory circuits in the human motor cortex. J Physiol 2008b;586:4481–7. Di Lazzaro V, Pilato F, Saturno E, Oliviero A, Dileone M, Mazzone P, et al. Theta-burst repetitive transcranial magnetic stimulation suppresses specific excitatory circuits in the human motor cortex. J Physiol 2005;565:945–50. Gamboa OL, Antal A, Moliadze V, Paulus W. Simply longer is not better: reversal of theta burst after-effect with prolonged stimulation. Exp Brain Res 2010;204:181–7. Hamalainen M, Hari R, Ilmoniemi RJ, Knuutila J, Lounasmaa OV. Magnetoencephalography—theory, instrumentation, and applications to noninvasive studies of the working human brain. Rev Mod Phys 1993;65:413–97. Hess G, Donoghue JP. Long-term potentiation and long-term depression of horizontal connections in rat motor cortex. Acta Neurobiol Exp (Wars) 1996;56:397–405. Huang YZ, Chen RS, Rothwell JC, Wen HY. The after-effect of human theta burst stimulation is NMDA receptor dependent. Clin Neurophysiol 2007;118:1028–32. Huang YZ, Edwards MJ, Rounis E, Bhatia KP, Rothwell JC. Theta burst stimulation of the human motor cortex. Neuron 2005;45:201–6. Huang YZ, Rothwell JC, Edwards MJ, Chen RS. Effect of physiological activity on an NMDA-dependent form of cortical plasticity in human. Cereb Cortex 2008;18:563–70. Huang YZ, Sommer M, Thickbroom GW, Hamada M, Pascual-Leonne A, Paulus W, et al. Consensus: new methodologies for brain stimulation. Brain Stimul 2009;2:2–13. Iezzi E, Conte A, Suppa A, Agostino R, Dinapoli L, Scontrini A, et al. Phasic voluntary movements reverse the after-effects of subsequent theta-burst stimulation in humans. J Neurophysiol 2008;100:2070–6. Jensen O, Vanni S. A new method to identify multiple sources of oscillatory activity from magnetoencephalographic data. Neuroimage 2002;15:568–74. Jurkiewicz MT, Gaetz WC, Bostan AC, Cheyne D. Post-movement beta rebound is generated in motor cortex: evidence from neuromagnetic recordings. Neuroimage 2006;32:1281–9. Katayama T, Rothwell JC. Modulation of somatosensory evoked potentials using transcranial magnetic intermittent theta burst stimulation. Clin Neurophysiol 2007;118:2506–11. Larson J, Lynch G. Induction of synaptic potentiation in hippocampus by patterned stimulation involves two events. Science 1986;232:985–8. Lee PL, Shang LZ, Wu YT, Shu CH, Hsieh JC, Lin YY, et al. Single-trial analysis of cortical oscillatory activities during voluntary movements using Empirical

Y.-F. Hsu et al. / Clinical Neurophysiology 122 (2011) 2260–2267 Mode Decomposition (EMD)-based spatiotemporal approach. Ann Biomed Eng 2009;37:1683–700. Liu Z, Fukunaga M, de Zwart JA, Duyn JH. Large-scale spontaneous fluctuations and correlations in brain electrical activity observed with magnetoencephalography. Neuroimage 2010;51:102–11. Murthy VN, Fetz EE. Coherent 25- to 35-Hz oscillations in the sensorimotor cortex of awake behaving monkeys. Proc Natl Acad Sci USA 1992;89:5670–4. Peinemann A, Reimer B, Loer C, Quartarone A, Munchau A, Conrad B, et al. Longlasting increase in corticospinal excitability after 1800 pulses of subthreshold 5 Hz repetitive TMS to the primary motor cortex. Clin Neurophysiol 2004;115:1519–26. Pfurtscheller G, Lopes da Silva FH. Event-related EEG/MEG synchronization and desynchronization: basic principles. Clin Neurophysiol 1999;110:1842–57. Pfurtscheller G, Neuper C, Kalcher J. 40-Hz oscillations during motor behavior in man. Neurosci Lett 1993;164:179–82. Pfurtscheller G, Stancak Jr A, Neuper C. Post-movement beta synchronization. A correlate of an idling motor area? Electroencephalogr Clin Neurophysiol 1996;98:281–93. Reis J, Swayne OB, Vandermeeren Y, Camus M, Dimyan MA, Harris-Love M, et al. Contribution of transcranial magnetic stimulation to the understanding of cortical mechanisms involved in motor control. J Physiol 2008;586:325–51. Rossi S, Hallett M, Rossini PM, Pascual-Leone A. Safety, ethical considerations, and application guidelines for the use of transcranial magnetic stimulation in clinical practice and research. Clin Neurophysiol 2009;120:2008–39. Rothwell JC, Hallett M, Berardelli A, Eisen A, Rossini P, Paulus W. Magnetic stimulation: motor evoked potentials. The International Federation of Clinical

2267

Neurophysiology. Electroencephalogr Clin Neurophysiol Suppl 1999;52:97–103. Salenius S, Salmelin R, Neuper C, Pfurtscheller G, Hari R. Human cortical 40 Hz rhythm is closely related to EMG rhythmicity. Neurosci Lett 1996;213:75–8. Sanes JN, Donoghue JP. Oscillations in local field potentials of the primate motor cortex during voluntary movement. Proc Natl Acad Sci USA 1993;90:4470–4. Schaefer AT, Angelo K, Spors H, Margrie TW. Neuronal oscillations enhance stimulus discrimination by ensuring action potential precision. PLoS Biol 2006;4:e163. Stagg CJ, Wylezinska M, Matthews PM, Johansen-Berg H, Jezzard P, Rothwell JC, et al. Neurochemical effects of theta burst stimulation as assessed by magnetic resonance spectroscopy. J Neurophysiol 2009;101:2872–7. Suppa A, Ortu E, Zafar N, Deriu F, Paulus W, Berardelli A, et al. Theta burst stimulation induces after-effects on contralateral primary motor cortex excitability in humans. J Physiol 2008;586:4489–500. Talelli P, Cheeran BJ, Teo JT, Rothwell JC. Pattern-specific role of the current orientation used to deliver theta burst stimulation. Clin Neurophysiol 2007;118:1815–23. Tamura Y, Hoshiyama M, Nakata H, Hiroe N, Inui K, Kaneoke Y, et al. Functional relationship between human rolandic oscillations and motor cortical excitability: an MEG study. Eur J Neurosci 2005;21:2555–62. Vickery RM, Morris SH, Bindman LJ. Metabotropic glutamate receptors are involved in long-term potentiation in isolated slices of rat medial frontal cortex. J Neurophysiol 1997;78:3039–46. Ziemann U, Tergau F, Wischer S, Hildebrandt J, Paulus W. Pharmacological control of facilitatory I-wave interaction in the human motor cortex. A paired transcranial magnetic stimulation study. Electroencephalogr Clin Neurophysiol 1998;109:321–30.