effect of theta burst stimulation on sensorimotor cortex ...

EFFECT OF THETA BURST STIMULATION ON SENSORIMOTOR CORTEX IN HUMANS Kaoru Matsunaga,1 Murat Sağlam,2 Nobuki Murayama,2 Yuki Hayashida,2 Ryoji Nakanishi 1 1

2

Department of Neurology, Kumamoto Kinoh Hospital, Kumamoto, Japan Department of Human and Environmental Informatics, Graduate School of Science and Technology, Kumamoto University, Japan

Abstract - Objective: To study the after-effect of theta burst

transcranial magnetic stimulation (TBS) on sensorimotor cortex excitability as well as cortico-muscular synchronization in humans. Methods: We used a continuous TBS (cTBS) paradigm for 40 sec (600 pulses) (Huang et al., 2005). Somatosensory evoked potentials (SEPs) following electrical stimulation of right or left median nerve and motor evoked potentials (MEPs) in the right or left first dorsal interosseous (FDI) muscles were recorded before and after cTBS over the left motor (M1) or sensory (S1) cortex (2cm posterior to M1) in healthy subjects. In addition, coherence function was computed as a measure of cortico-muscular synchronization by recording electroencephalogram (EEG) and electromyogram (EMG) during the isometric contraction. Results: Amplitudes of P25/N33 (parietal SEP component) following right median nerve stimulation were significantly increased for at least 53 minutes after cTBS over left M1, whereas this component was suppressed for 13 min after cTBS over left S1. MEPs in right as well as left FDI muscles were suppressed in a similar time course after cTBS over left M1. The beta-band cortico-muscular coherence for the C3 scalp site, significantly decreased in 30-60 minutes after cTBS over left M1. Conclusion: cTBS can induce a long-lasting change in sensorimotor cortex excitability and cortico-muscular synchronization. Keywords – Theta burst stimulation; Transcranial magnetic stimulation; Motor evoked potentials; Somatosensory evoked potentials; Cortico-muscular coherence

I. INTRODUCTION In animal experiments, repetitive electrical stimulation of central nervous pathways with “theta burst” paradigms (3-5 pulses at 100 Hz repeated at 5 Hz) has been used to induce changes in the efficacy of synaptic transmission that outlast the period of stimulation. Recently Huang et al. [1] developed a “theta burst” paradigm to condition the human motor cortex using low intensity repetitive transcranial magnetic stimulation (rTMS). This theta burst stimulation (TBS) produces a long-lasting effect on motor cortex excitability and behaviour after an application period of only

-1-

978-1-4244-3316-2/09/$25.00 ©2009 IEEE

20-190 seconds in humans [1]. The pattern of delivery of TBS (continuous TBS versus intermittent TBS) is crucial in determining whether the excitability of the motor cortex, as monitored by the amplitude of transcranial motor evoked potentials (MEPs), is increased or decreased. Continuous TBS (cTBS) decreases the amplitude of MEPs, while they are increased by intermittent TBS. One question we address here is whether TBS has any lasting after-effect on somatosensory cortex. Another question is whether TBS can modulate a cortico-muscular synchronization during voluntary muscle contractions in rather near-natural conditions. The functional coupling between corticies and the corresponding muscles can be assessed by means of the coherence function in the frequency domain, or of the cumulant density function in the time domain [2]. The aim of the present paper was to test whether cTBS has after-effects on somatosensey evoked potentials (SEPs) as well as MEPs evoked from both hemispheres. In addition, we examined how cTBS can affect the cortico-muscular coherence during voluntary isometric contraction [3, 4]. II. SUBJECTS AND METHODS 2.1 Subjects Twelve healthy volunteers (10 men and 2 women; mean age (±SD), 42.0±6.3 years) were studied for the effect of cTBS on SEPs and MEPs (Experiment 1)[3]. The other 8 healthy volunteers (7 men, 1 woman; 25.6±2.2 years) were studied for the effect of cTBS on the beta-band corticomuscular coherence (Experiement 2)[4]. All subjects gave their informed consent to the studies, which were approved by the local ethical committee and conformed to the requirements of the Declaration of Helsinki. 2.2 Experimental Procedures [3, 4] In the first experiment (Experiment 1)[3], two different experiments (SEP or MEP experiments) were performed on separate days. In order to assess the time course of the cTBS effect over the sensorimotor cortex, SEPs or MEPs were recorded before and immediately (0), 10, 20, 30, 40, 50 and 60 min after the end of the cTBS trains in each experiment. cTBS was performed either over the scalp location of left M1 or S1 in each experiment: M1 was defined as the “motor

hot spot” for the right hand muscle determined by a singlepulse TMS and S1 was a point 2cm posterior from this site . This S1 location (M1+2cm) presumably lies over the crown of the postcentral gyrus [5]. In the second experiment (Experiment 2)[4], in order to assess the effects of cTBS on cortico-muscular coherence, EEG and EMG signals were simultaneously recorded before and after cTBS. cTBS was applied either over the scalp location of left M1 or S1, or as sham. First, to obtain baseline coherence values (as the magnitudes and the frequency bands) for each subject, two recording sessions were made before cTBS application. Three minutes after the pre0 session, we delivered cTBS, and then 2 min later, a post-hoc recording session was made. Afterwards, a recording session was repeated every 30 min (i.e. 30, 60, 90, and 120 min after cTBS). Motor task in the present experiments was similar to that described in previous studies [6]. Briefly, in each recording session, subjects were asked to maintain weak isometric contraction of their right first dorsal interosseous (FDI) muscle with ~15 % force level of the maximum voluntary contraction (MVC). A force sensor was placed between the thumb and the index finger for monitoring the force level during the EEG-EMG recording. The sensor output was shown also to the subject for her/him to adjust the force level of muscle contraction, but only prior to recording sessions in order to avoid a possible visualcognitive effect of exhibiting the force level [6]. 2.3 Somatosensory evoked potential (SEP) recording During the sessions, subjects lay down in a bed and were observed continuously. SEPs were recorded following electrical stimulation of the right or left median nerve at the wrist at 3 Hz with a pulse width of 0.2 ms. The intensity of stimulation was fixed at the motor threshold and was checked throughout the course of the experiment by monitoring the evoked EMG response in the Abductor pollicis brevis (APB) muscle. SEPs were recorded from scalp Ag-AgCl surface electrodes 2 cm posterior from C3 or C4 (parietal component) and 5 cm anterior from C3 or C4 (frontal component) referred to the contralateral earlobe according to the 10-20 electrode system for EEG placement. Recordings were made with a band-pass of 5 Hz to 1 kHz using a Synax 1200 (NEC, Japan). 500 responses were averaged in each session 2.4 Motor evoked potential (MEP) recording MEPs were recorded from Ag-AgCl surface electrodes over the right or left FDI or APB muscles. The signal was amplified and band-pass filtered (20-3000 Hz) by a Synax 1200 (NEC, Japan). MEPs were evoked at rest by a singlepulse TMS of the contralateral M1 with a High Power Magstim 200 machine and a figure of eight coil with mean loop diameters of 70mm (Magstim Co., Whitland, Dyfed, UK). The magnetic stimulus had a nearly monophasic pulse configuration with a rise time of approximately 100 µs, decaying back to zero over approximately 0.8 ms. The coil current during the rising phase of the magnetic field flowed toward the handle. We determined the optimum position for

-2-

activation of the FDI or APB muscles for the "motor hot spot". Stimulus intensities for TMS were determined so that the amplitude of the MEP was about 1 mV peak to peak for the baseline before cTBS. In Experiment 1, we recorded the MEP from the right as well as the left FDI muscles. 2.5 EEG and EMG recording for the coherence measurement EEG signals were recorded from 19 scalp electrodes mounted on a cap (Electro-cap International, Inc., Eaton, OH). Earlobe Ag–AgCl surface electrodes served as the reference. Surface EMG was recorded from the right FDI muscle. EEG and EMG signals were recorded by a bioamplifier (MME-3124; Nihon Kohden, Tokyo, Japan) with passbands of 0.5-200 Hz and 5-300 Hz, respectively, and with 1-kHz sampling frequency. The EMG signals were rectified to be used for analyses [6, 7]. As mentioned above, simultaneous recordings of EEG and EMG signals were made in Experiment 2. 2.6 Theta burst stimulation We applied a continuous TBS paradigm (cTBS) for 40 seconds (600 pulses) to the left sensorimotor cortex. The main element of cTBS is a burst of 3 stimuli at 50Hz which is repeated at 5Hz as described by Huang et al. [1]. TBS was performed using a figure of eight coil with mean loop diameter of 70 mm, connected to a Magstim Super Rapid stimulator (Magstim Co., Whitland, Dyfed, UK). The magnetic stimulus had a biphasic waveform with a pulse width of approximately 300 µs. During the first phase of the stimulus, the current in the centre of the coil flowed toward the handle. Each individual’s active motor threshold (AMT) over the M1 for the right FDI or APB muscle was determined prior to cTBS using the Magstim Super Rapid stimulator and the coil orientation with the handle pointing backwards and laterally at a 45° angle away from the midline. AMT was defined as the lowest stimulus intensity at which 5 of 10 consecutive stimuli elicited reliable MEP (about 200 µV in amplitude) during slight (about 20% maximum) tonic contraction of the target muscle. A total number of 600 stimuli at 80% AMT for the FDI or APB muscles were applied during a single cTBS session. Sham stimulation was used for Experiment 2. A figure-ofeight coil connected to an uncharged magnetic stimulator was placed over the left M1. In addition, another coil connected to the charged stimulator was held 10cm above the scalp, and was allowed to pass the same electric current as the real cTBS. 2.7 Data analysis 2.7.1 SEP and MEP analysis Peak-to-peak amplitudes of SEP components and MEPs were measured. Mean amplitudes of each component of two SEPs or 40 MEPs recorded before cTBS were used as baseline values. The amplitude of SEPs or the mean amplitude of 20 MEPs recorded on each side at each time point after cTBS was compared to the baseline values. Because the absolute latencies of SEP components varied among subjects, grand average waveforms were constructed

by adjusting the time scale with respect to the peak latencies of the P14 far-field component of each waveform. 2.7.2 Coherence analysis Coherence function was used to quantify the synchronization between signals of EEG and EMG. Coherence is the squared magnitude of the cross-power spectrum of a signal pair normalized by the product of their auto-power spectra, and is described by the expression: 2

S XY ( f ) κ (f)= S XX ( f ) SYY ( f ) 2 XY

.

Here, κ 2XY ( f ) is the coherency; S XY ( f ) represents the cross-spectral density function between signals x and y; S XX ( f ) and SYY ( f ) stand for the auto-spectral density function (spectral power) of the signals x and y, respectively. Since coherence is a normalized measure of the 2

cross-correlation between the signal pair, κ XY ( f ) = 1 represents a perfect linear dependence and 2 κ XY ( f ) = 0 indicates

a lack of linear dependence within the signal pair. To calculate EEG-EMG coherence, the EEG and EMG signals recorded in Experiment 1 were segmented into non-overlapping epochs of 1024-msec duration. For the EEG, the current source density (CSD) method was utilized in order to achieve spatially sharpened signals [6, 7]. The fast Fourier transform with the epoch size of 1024, resulting in frequency resolution of 0.98 Hz, was used to convert the signals in time domain into the frequency-domain signals. Cross-spectra between the EMG signal and each of the 19 EEG signals, as well as auto-spectra of those signals, were calculated for the EEG-EMG coherence to be obtained by the above-mentioned equation. Coherence magnitudes for

leads are all clearly visible. Afterc TBS over left M1, the amplitudes of P25/N33 components increased, whereas P14/N18, N18/P22, P14/N20 and N20/P25 components were relatively unaffected (Fig. 1). In contrast, TBS over left S1 reduced P25/N33 (Fig. 2). Peak latencies of all components were unchanged. The clearest effects were on the parietal P25/N33 component (see mean data in Fig. 3) which showed significant two and three-way interactions for time course x position (p=0.002) and time x side x position (p=0.011). Follow-up ANOVAs showed that these interactions were due to the fact that (a) cTBS had an effect on SEPs evoked from right but not from left median nerve, and (b) for right median nerve stimulation, cTBS over M1 increased the P25/N33 whereas it was decreased by cTBS over S1. Thus, two-way repeated measures ANOVAs on the data from the right and left median nerve separately revealed a significant time x position interaction after right median nerve stimulation (F7, 70=5.7, p0

were assumed to be statistically significant only if they were above the 95% confidence level, which was calculated as described in Rosenberg et al. [8]. 2.7.3 Statistical analysis The effects of cTBS on SEP, MEP or peak coherence magnitudes were evaluated employing repeated measures of ANOVA (analyses of variance). When necessary, the Greenhouse-Geisser correction was used to correct the nonsphericity. The post-hoc test with the Bonferroni correction for multiple comparisons was used to compare the SEP, MEP or coherence values between sessions. In all analyses, the statistical significance was assumed for p values smaller than 0.05.

RESULTS 3.1. After-effects on SEPs [3] Grand average waveforms of SEPs following right median nerve stimulation obtained from 11 subjects are shown in Figures 1 and 2. The P14, N18, P22 and N30 in the frontal leads and P14, N20, P25, N33 and P40 peaks in the parietal

-3-

Fig. 1 Effect of cTBS over left M1 on right median SEPs. 3.2 After-effects on MEPs [3] Fig. 4 summarizes the after-effects of cTBS on the MEPs. cTBS over left M1 decreased the amplitude of MEPs in both the right and left FDI muscles, whereas there was no significant effect after cTBS of left S1. Thus, a three-way repeated measures ANOVA with time course, side of recording and position of cTBS revealed a significant main effect of position (F1, 9=9.1, p=0.015) and significant interactions of position x time (F 7, 63=3.4, p=0.004) and side

x time (F7,63=2.3, p=0.036). A separate two-way repeated measures ANOVA on the MEP data before and after cTBS over M1 with time and side of recording as main factors showed a significant main effect of time (F7, 63=3.8, p=0.001), but failed to show a significant interaction of side and time (p=0.7). This was due to the fact that MEPs recorded from right and left FDI muscles decreased with a similar time course after cTBS over left M1. Post hoc tests indicated that the MEPs in the right FDI muscle were significantly suppressed for up to 42 min after cTBS over left M1 and the MEPs in the left FDI muscle were significantly suppressed for more than 50 min starting about 12 min after cTBS (Fig. 4). In contrast, two-way repeated measures ANOVA on the MEP data before and after S1 cTBS with main factors of time and side of recording failed to show any significant effects, suggesting there is no effect of cTBS at this site on MEPs in either hand.

TBS over left M1

A m 1.0 pli tu de0.5 (m V)

**

**

V)3 µA m2 pli tu 1 de

**

**

102030405013 Time 23after33 43 53 TBS (min)

6063

baselin e

36

1323334316 Time 26after 36 46 TBS (min)

5356

V) 3 µ A m2 pli tu 1 de

**

(min)

5053

6063

baseline 3-6

13-16

**

**

baseline 2-4

12-14 22-24 32-34 42-44 52-54 62-64

Time after TBS (min)

Lt FDI

0.0

10-12 20-22 30-32 40-42 50-52 60-62

baseline 2-4

12-14 22-24 32-34 42-44 52-54 62-64

Time after TBS (min)

Fig. 4 Effect of cTBS over left M1 or S1 on MEPs. 3.3 After-effects on cortico-muscular coherence [4] For all subjects, significant magnitude of coherence were always observed for the C3 site, and the peak of coherence spectra were within the beta frequency band (13-30Hz), before and after the cTBS. On the C3 site, peak coherence values were normalized with the control value which was obtained by averaging the peak coherence values in the pre30 and pre0 sessions. The control values from the eight subjects (mean ± S.E.) were 0.067±0.011, 0.068±0.015, 0.054±0.012 for cTBS-on-M1, cTBS-on-S1 and sham, respectively. The normalized peak coherence before and after cTBS are shown in Fig. 5. Two-way repeated measures of ANOVA revealed a significant interaction between stimulation site and time (F12,84=2.23, p