- Email: [email protected]
The cortical silent period: intrinsic variability and relation to the waveform of the transcranial magnetic stimulation pulse
Clinical Neurophysiology 115 (2004) 1076–1082 www.elsevier.com/locate/clinph
The cortical silent period: intrinsic variability and relation to the waveform of the transcranial magnetic stimulation pulse M. Orth*, J.C. Rothwell Sobell Department of Motor Neuroscience and Movement Disorders, The National Hospital for Neurology and Neurosurgery, Royal Free and University College Medical School, Institute of Neurology, Box 77, Queen Square, London WC1N 3BG, UK Accepted 20 December 2003
Abstract Objective: To assess the variability of the duration of the contralateral cortical silent period (CSP) between individuals and to assess the effect of different transcranial magnetic stimulation (TMS) pulse waveforms. Methods: In Expt. 1, CSP duration, and the motor-evoked potential (MEP) amplitude and area were measured in the first dorsal interosseous muscle (FDI) of 11 subjects on 3 separate occasions using a TMS intensity of 150% active motor threshold (AMT). In Expt. 2, the stimulation intensity was varied between 100% AMT and 150% AMT. In both sets of experiments, 3 types of TMS pulse were used: monophasic posterior-anterior (PA) induced current in the brain, monophasic anterior-posterior induced current (AP), and biphasic PA/AP stimulation. Results: Experiment 1: Between-subject variation in CSP duration was high. In addition, the duration after PA stimulation was significantly shorter than after AP or PA/AP stimulation. However, there was a good correlation between CSP duration and the area, or amplitude, of the MEP. This meant that calculating the ratio of duration/amplitude or duration/area reduced intersubject variability and eliminated differences between TMS pulses. Experiment 2: increasing stimulation intensity increased the mean value of all parameters, but with significantly lower values for PA than other forms of stimulation. The ratios of duration/amplitude or duration/area did not differ between current flow directions and were relatively constant for intensities 130– 150% AMT. Conclusions: Between-subject variation in the duration of the CSP is high. A given intensity of stimulation (expressed in %AMT) produces a shorter CSP for PA stimulation than for AP or PA/AP stimulation. Significance: If the ratio (CSP duration)/(MEP size) is calculated, then intersubject variability is reduced, and TMS pulse type differences are eliminated. q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Transcranial magnetic stimulation; Cortical silent period; Variability; Recruitment; Current flow direction
1. Introduction A transcranial magnetic stimulation (TMS) pulse given to the motor cortex during tonic contraction of a target muscle produces a motor-evoked potential (MEP) followed by a period of EMG silence before the activity resumes its prestimulus baseline level of activity. This period of silence is known as the cortical silent period (CSP). The duration of the CSP depends on the intensity of the TMS pulse, but is relatively unaffected by the level of ongoing muscle activity. For CSP durations longer than about 100 ms, * Corresponding author. Tel.: þ44-20-7837-3611x3947; fax: þ 44-207278-8772. E-mail address: [email protected] (M. Orth).
the resumption of EMG activity is thought to depend on recovery of motor cortical excitability from GABAergic inhibition following the TMS pulse (Chen et al., 1999; Fuhr et al., 1991; Inghilleri et al., 1993). The duration of the CSP is often measured in both clinical and basic science studies, and is used as a measure of excitability in cortical inhibitory circuits. However, the duration is highly variable from subject to subject. In this paper we examine the variation between subjects. In the first experiment we used a standard intensity of TMS (150% AMT) to evoke the silent period in the first dorsal interosseous muscle (FDI). We then tested whether we could reduce the variability in the measure of CSP duration by combining it with a measure of MEP amplitude, since
1388-2457/$30.00 q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2003.12.025
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
1077
previous studies have reported that the two factors are linked (Wilson et al., 1993). In the same experiment we also asked whether the CSP was differentially affected by different TMS pulse waveforms. For example, it is known that monophasic posterioranterior (PA) stimulation has a lower absolute threshold for MEP recruitment than anterior-posterior (AP) stimulation (Sakai et al., 1997), and that the former tends to recruit I1 wave input to corticospinal neurones whereas the latter favours I3 recruitment (Kammer et al., 2001). It has also been reported recently that recruitment of MEPs differs with monophasic AP stimulation versus PA stimulation (Sakai et al., 1997). Given that the CSP is usually measured at a given intensity above motor threshold, we tested whether it had the same duration at an intensity of 150% AMT with all 3 pulse waveforms. In the second experiment we examined how the CSP duration and MEP size changed over a range of stimulus intensities from 100– 150% AMT.
will be referred to as PA throughout the text. When the coil handle is rotated by 1808 from this position the coil handle points forward, and the current induced in the brain flows anterior-posterior. This will be referred to as AP throughout the text. Biphasic stimulus pulses were applied through the same coil using a Magstim Super Rapid stimulator. We positioned the coil with the handle pointing backwards. The initial current flow in the brain is determined by the up-slope component of the biphasic wave so that with this coil position the current induced in the brain first flows posterioranterior and then anterior-posterior. This will be referred to as PA/AP throughout the text. Fig. 1 illustrates the different characteristics of the stimulators and the current flow direction. The optimal spot for right FDI stimulation was marked with a felt pen. Note that the maximum absolute output of the Magstim Super Rapid stimulator is smaller than that of the High Power Magstim device, so that the values of absolute stimulus intensity should not be compared between the monophasic and biphasic devices.
2. Material and methods
2.4. Motor threshold and silent period measurements
2.1. Subjects
Resting motor threshold (RMT) was defined as the minimum intensity needed to evoke an MEP of . 50 mV in 5 out of 10 consecutive trials in the relaxed FDI. Active motor threshold (AMT) was defined as the minimum intensity needed to evoke an MEP of . 200 mV in 5 out of 10 trials in the tonically active FDI (, 20% of maximal contraction as assessed visually on an oscilloscope). Thresholds were approached from above threshold in steps of 1% stimulator output. Once no MEPs could be elicited the intensity was increased in steps of 1% stimulator output until a minimal MEP was observed. This intensity was taken as motor threshold. RMT and AMT were determined separately for each stimulator. Cortical silent periods were recorded from the tonically active right FDI.
Eleven healthy right handed Caucasian volunteers were studied (mean age 29 years, range 20– 37 years, 6 women). Subjects were asked to refrain from caffeine-containing beverages on the day of the experiment. All subjects participated in Expt. 1; 10 subjects participated in the second experiment. Subjects gave informed written consent, and the local ethics committee approved the study protocol. 2.2. Electromyography recordings Surface electromyograms (EMG) were recorded from the right FDI using silver/silver-chloride disc surface electrodes (1 cm diameter) in a belly tendon montage. The EMG signal was amplified and analogue filtered (30 Hz to 3 kHz) with a Digitimer D150 amplifier (Digitimer Ltd., Welwyn Garden City, UK). Data (sampling rate 4 kHz) was digitized for offline analysis using Signal software (Cambridge Electronic Devices, Cambridge, UK). Peak-to-peak amplitude of MEPs, the area under the curve of the MEP and the silent period duration were measured with in-house software. 2.3. Transcranial magnetic stimulation Subjects were seated in a comfortable chair. Magnetic stimuli were given with a hand-held figure-of-8 coil (outer winding diameter 9 cm) connected to a High Power Magstim 200 stimulator (Magstim Co., Whitland, Dyfed, UK). This stimulator generates a magnetic pulse with monophasic waveform and in the brain induces a current with posterior-anterior flow when the coil handle is positioned at an angle of 458 pointing backwards. This
Fig. 1. Illustration of the coil positions, magnetic waveforms generated by the stimulators and the current flow directions induced in the brain. PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/ anterior-posterior.
1078
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
In Expt. 1, a fixed test stimulus intensity of 150% AMT was used and the session repeated twice (total of 3 sessions). In the second experiment, the test stimulus intensity was varied in each subject (100, 110, 120, 130, 140, 150% of the AMT measure with the respective stimulator). The order of the experiments was randomized. For each test stimulus intensity, 10 trials were collected in each subject in a single session with an interval of 4 s between trials. In each individual trial the duration of the silent period was measured from the beginning of the MEP evoked by the test stimulus to the resumption of (any level of) sustained EMG activity. In addition, the peak-to-peak amplitude and the area under the test stimulus MEP were determined. 2.5. Data analysis The following parameters were examined: CSP duration, MEP amplitude, MEP area. In addition, ratios were calculated from CSP duration and MEP amplitude, or MEP area, respectively. In the first experiment, we examined how variable CSP duration, MEP amplitude or MEP area were between sessions. For this purpose, a coefficient of variation was calculated from the mean of each session and the corresponding standard deviation (CV ¼ session SD/session mean £ 100). We went on to test how reproducible the results were between sessions. To this end, a repeated-measures analysis of variance (ANOVA) model was used with ‘session’ and ‘stimulator’ as the within-subject factors. As the difference between the sessions was not significant for any of the stimulation conditions, data from all 3 sessions were pooled for each subject. Next, we examined whether CSP duration, MEP amplitude or MEP area differed between subjects by calculating a CV for either CSP duration, or MEP amplitude or MEP area, as main factor. For each stimulator, a regression analysis was used to test how accurately CSP duration was predicted by either MEP amplitude or MEP area. In the second experiment, we sought to investigate the effect of different stimulus intensities on CSP duration, MEP amplitude or MEP area. For this purpose, we used a repeated-measures ANOVA model with ‘coil current’ and stimulus ‘intensity’ as within-subject factors. As the dataset for intensities smaller than 130% was incomplete, we restricted the analysis to the intensities 130, 140 and 150% AMT. A statistical difference in the ANOVAs was followed by a post hoc paired t test analysis with Bonferroni correction of the P value. Mauchly’s test examined sphericity in the repeated measures ANOVA models. Statistical significance levels were set to P ¼ 0:05, for non-spherical data the Greenhouse-Geisser correction was used. All statistical analysis was performed using SPSS 11 for Windows software package.
3. Results 3.1. Experiment 1: stimulation at 150% AMT With PA current flow, subjects had a mean RMT of 37% of stimulator output (SD 4.5, range 30 – 47) and a mean AMT of 26.9% (SD 4.7, range 17– 35). With AP current flow the mean RMT was 46.8% (SD 8, range 39 –63) and the mean AMT was 38.4% (SD 6.4, range 29 –51). With PA/AP current flow, subjects had a mean RMT of 49.3% (SD 6.5, range 39– 63) and a mean AMT of 38.9% (SD 6, range 29– 52). 3.1.1. Between-subject variability The results for all CSP and MEP parameters were very similar between the 3 sessions in each subject (repeated measures ANOVA, P . 0:05) with the between-session CV always smaller than 10%. Thus the data from the 3 sessions were pooled for statistical analysis (Table 1). There was considerable variability between subjects in the measures of ‘amplitude’ and ‘area’ of MEP and ‘duration’ of CSP for each of the current flow directions (Fig. 2) with the betweensubject CV ranging from 20.0 to 35.3% (Table 2). 3.1.2. Effect of stimulus waveform The duration of the cortical silent period and the area under the MEP differed significantly depending upon the direction of current flow (repeated-measures ANOVA, Table 2: P ¼ 0:027, F ¼ 4:44, df 2,18 and P ¼ 0:015, F ¼ 5:35, df 2,18, respectively). Post hoc analysis revealed that the CSP duration was shorter for PA than for either AP or PA/AP stimulation (paired t test, P ¼ 0:006 and P ¼ 0:05, respectively), and that the MEP area was smaller for PA than PA/AP stimulation (paired t test, P ¼ 0:001). There was no significant effect of waveform on MEP amplitudes, although as with area measures, the peak-topeak amplitude of the MEP did tend to be smaller for PA than for AP or PA/AP stimulation. The lack of effect may well have been due to saturation of the MEP amplitude in some subjects at 150% AMT, making area a better measure of response size. For each direction of current, the MEP amplitude and area significantly predicted the duration of the cortical silent Table 1 Coefficient of variation of CSP duration, MEP amplitude and MEP area Between-session variability
Duration MEP amplitude MEP area
PA
AP
PA/AP
3.8 7.1 6.1
4.7 7.3 8.4
4.6 6.1 7.1
The between-session CV was calculated from the median of the CV of 3 sessions in each of 11 subjects. PA, posterior-anterior; AP, anteriorposterior; PA/AP, posterior-anterior/anterior-posterior.
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
1079
Fig. 2. Correlation of CSP duration with MEP area (A,C,E) or MEP amplitude (B,D,F), for PA current flow direction (A,B), AP current flow direction (C,D) and PA/AP current flow direction (E,F). Each data point represents the mean of 3 sessions in each subject (n ¼ 11). PA, posterior-anterior; AP, anteriorposterior; PA/AP, posterior-anterior/anterior-posterior.
Table 2 The effect of current flow direction in the brain at a fixed stimulation intensity of 150% AMT on CSP duration and MEP amplitude or MEP area, and on ratios calculated from CSP duration and the corresponding MEP amplitude or area, respectively PA
Duration (ms) MEP amplitude (mV) MEP area (mV ms) Duration/amplitude (ms/mV) Duration/area (ms/mV ms)
AP
PA/AP
Mean (SD)
CV
Mean (SD)
CV
Mean (SD)
CV
108.0 (38.1) 7.1 (1.7) 620.9 (215.3) 15.4 (2.2) 0.18 (0.02)
35.3 24.1 34.7 14.2 12.8
139.2 (30.8) 8.0 (2.0) 746.8 (223.4) 19.0 (5.2) 0.2 (0.05)
22.1 25.0 29.9 27.4 26.0
132.0 (37.3) 8.0 (1.6) 776.4 (205.5) 16.8 (2.7) 0.18 (0.03)
28.3 20.0 26.5 16.4 17.5
Results are from the mean of 3 repeated measurements in each of 11 subjects. CV, coefficient of variation; PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/anterior-posterior.
1080
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
period (Fig. 2, regression analysis). The r 2 values were greater for PA (Fig. 2A,B) compared with AP (Fig. 2C,D) or PA/AP (Fig. 2E,F) current. The strong correlation of CSP duration with MEP amplitude, and MEP area, led us to calculate ratios of duration/MEP amplitude and duration/MEP area (Table 2) for each direction of current. The ratios did not differ between different directions of current flow (repeatedmeasures ANOVA, P . 0:1 for both). In order to analyse whether calculation of the ratio duration/area reduced the between-subject variability, we went on to compare the CV of CSP duration with the CV of the corresponding CSP duration/MEP area, or MEP amplitude, ratio for each of the current flow directions. For PA and PA/AP current flow direction, the calculation of the duration/MEP area, or MEP amplitude ratio reduced intersubject variability whereas the intersubject CV remained high for AP current flow (Table 2). 3.2. Experiment 2: the effect of different stimulation intensities In the second experiment we varied the stimulation intensities to assess whether the differences that we had seen with stimulation at 150% AMT also occurred at other intensities. Since the data set was incomplete for lower stimulus intensities, we only analysed that from 130– 150% AMT in detail. Separate two-factor ANOVAs with intensity and coil current as main factors were performed on MEP and CSP measures. This revealed a significant main effect of stimulus intensity on the duration of the cortical silent period (Fig. 3A, P , 0:001, F ¼ 65:4, df 1.18, 10.63), MEP amplitude (data not shown, P ¼ 0:002, F ¼ 9:1, df 2,18) and MEP area (Fig. 3B, P , 0:001, F ¼ 15, df 2,18). All 3 parameters increased significantly with increasing stimulation intensities. There was also a significant main effect of coil current on the duration of the cortical silent period (Fig. 3A, P ¼ 0:002, F ¼ 9:2, df 2,18) and MEP area (Fig. 3B, P ¼ 0:019. F ¼ 5:0, df 2,18), with a marginal effect on MEP amplitude (P ¼ 0:073, F ¼ 3:03, df 2,18). There was no significant interaction, suggesting that responses to all 3 current directions were influenced by intensity in the same way. Post hoc pairwise comparisons on the main effect of coil current showed that the duration of the CSP was shorter for PA than AP (P ¼ 0:005, F ¼ 13:9, df 1,9) or PA/AP (P ¼ 0:011, F ¼ 10, df 1,9), and the MEP area was smaller with PA compared with AP (P ¼ 0:032, F ¼ 6:4, df 1,9) or PA/AP (P ¼ 0:009, F ¼ 10:8, df 1,9). MEP amplitude was also significantly smaller with PA compared with PA/AP (P ¼ 0:044, F ¼ 5:5, df 1,9) while the difference between PA and AP was marginal (P ¼ 0:051, F ¼ 5, df 1,9). Comparison of the ratios of either CSP duration/MEP area or CSP duration/MEP amplitude revealed no difference between stimulation intensities or current flow directions (Fig. 3C).
Fig. 3. The effect of different stimulation intensities on CSP parameters. At stimulation intensities smaller than 130% AMT, the data set is incomplete (segment on the left). At stimulation intensities above 120% AMT, (segment on the right) CSP duration (A) and MEP area (B) are significantly smaller with PA current flow (straight line) compared with AP (regularly dotted line) or PA/AP (irregularly dotted line) while the ratio calculated from CSP duration and MEP area is not different (C). Values are means of a single session in 10 subjects, error bars are SDs. PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/anterior-posterior.
4. Discussion Most commonly, the cortical silent period is recorded with the subject voluntarily contracting the target muscle between 20 and 30% of the maximal force, and with a test stimulus intensity of around 130% RMT or 150% AMT (Kessler et al., 2002). In the present study we used such a conventional experimental set up in order to avoid fluctuations of voluntary contraction and saturation of the MEP.
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
The data from Expt. 1 confirmed a previous report showing that motor thresholds with PA current flow are lower than with AP current flow in the brain (Sakai et al., 1997). The data also revealed 4 important features of the CSP: (1) repeated measures of CSP duration (and MEP size) are relatively stable between sessions (coefficient of variation less than 10%) for any one individual; (2) between-subject variation is much higher for all measures; (3) for a given intensity of stimulus (i.e. 150% AMT), CSP duration, MEP area and MEP amplitude are all smaller with PA than other forms of stimulation; (4) for PA and PA/AP current flow direction the ratios CSP duration/MEP amplitude or CSP duration/MEP area are much less variable between subjects than the raw measures alone as there is a strong correlation between CSP duration and the size of the preceding MEP. Data from Expt. 2 indicate that these features are true for the range of stimulus intensities from 130 to 150% AMT. Feature (1) may mean that measures of CSP duration are adequate in repeated measures studies where the effect of a certain treatment may be evaluated on a given group of subjects. Feature (2) indicates that measures of CSP duration are much less sensitive in revealing differences between different groups of subjects and feature (3) implies that we should not compare the duration of the CSP or size of MEP made with different stimulus waveforms, even if the stimulus intensity has been adjusted to the same intensity relative to threshold in each case. Feature (4) suggests that a way to improve the power of between-subject comparisons is to calculate the ratio of CSP duration to MEP size (either amplitude or area). Experiment 2 went on to test how well this reasoning would apply at different intensities of stimulation between 100 and 150% AMT. Higher intensities were not studied because of the risk of saturating MEP size. The data showed that PA stimulation gave lower values for all parameters for all intensities, especially from 130 to 150% AMT. However, the ratio of CSP duration to MEP size was fairly constant across the intensities 130 –150% AMT which are conventionally used for CSP measurements. 4.1. Correlation between CSP duration and MEP size It has been recognized in the past that CSP duration and MEP amplitude both relate to the intensity of the stimulation (Wilson et al., 1993). In our study we observed that this correlation holds between subjects even at a fixed intensity of stimulation, and is independent of the pulse waveform of the TMS stimulus. Thus, subjects who have a large MEP to a given stimulus also tend to have a longer CSP. The relation is much stronger for PA than for PA/AP, and is relatively weak for AP current flow. The good correlation between MEP area and CSP duration led us to calculate ratios of CSP duration and the corresponding MEP area, and this reduced both the between-subject and the between-stimulator variability for
1081
PA and PA/AP current flow direction. This suggests that the factors that cause variation in the MEP are the same as those that cause variation in the duration of the CSP. One simple possibility is that the corticospinal outflow that produces the MEP is also responsible for the CSP. Corticospinal neurones give off recurrent axon collaterals that have a range of effects on cortical circuits. In particular, in the cat, recurrent collaterals of fast corticospinal axons have a predominantly inhibitory effect on the firing of neurones that have slower conducting axons (Phillips and Porter, 1977). Neurones with large axons appear to be activated by TMS pulses, and if tonic voluntary contractions are supported by the activity of slow axoned neurones, this could explain the occurrence of the silent period. However, further studies are needed to test this hypothesis in detail. 4.2. The difference between PA and AP or PA/AP stimulation We observed the strongest relation of MEP size and CSP duration for PA stimulation. The relation was good for PA/AP but weak for AP current flow direction. In addition, with stimulation intensities from 130 to 150% AMT all parameters were significantly smaller with PA current flow compared to AP or PA/AP current flow even though the ratios of CSP duration/MEP amplitude, or CSP duration/ MEP area, remained very similar. This difference, in the recruitment curves between different stimulus waveforms, was unexpected. One explanation might be that different current flow directions preferentially activate different subsets of cortical neurones and that the properties of these neuronal populations are different. Indeed, it is known that PA stimulation tends to recruit I1 input to corticospinal neurones whereas AP stimulation tends to recruit I3 inputs (Kammer et al., 2001). If the recruitment curves of these elements differed then it might account for the effects observed on the MEP measures. The effect on the CSP duration would be secondary to this effect if it were caused, as argued above, by the same corticospinal output as determined the MEP. Whatever the explanation, it is clear that not only should CSP durations not be directly compared using different stimulus waveforms, but neither should MEP parameters. Nevertheless, the ratio between these values may be of interest. For example, in the simple model above in which PTN discharge produces inhibition of neighbouring PTNs, an increase in MEPs without a corresponding lengthening of the CSP would indicate that there had been a reduction in excitability of the recurrent inhibitory pathway. Alternatively, in a model where separate mechanisms produce MEPs and CSP, the same change in ratio would be compatible with an independent effect on excitatory and inhibitory circuits in the cortex. In conclusion, we would like to suggest that the calculation of ratios of CSP duration and MEP area can
1082
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
help to reduce the between-subject and between-stimulator variability of silent period measures. Acknowledgements We wish to thank our subjects for their participation in this study. The work was funded by the Medical Research Council. References Chen R, Lozano AM, Ashby P. Mechanism of the silent period following transcranial magnetic stimulation. Evidence from epidural recordings. Exp Brain Res 1999;128:539 –42. Fuhr P, Agostino R, Hallett M. Spinal motor neuron excitability during the silent period after cortical stimulation. Electroenceph clin Neurophysiol 1991;81:257–62.
Inghilleri M, Berardelli A, Cruccu G, Manfredi M. Silent period evoked by transcranial stimulation of the human cortex and cervicomedullary junction. J Physiol 1993;466:521 –34. Kammer T, Beck S, Thielscher A, Laubis-Herrmann U, Topka H. Motor thresholds in humans: a transcranial magnetic stimulation study comparing different pulse waveforms, current directions and stimulator types. Clin Neurophysiol 2001;112:250–8. Kessler KR, Schnitzler A, Classen J, Benecke R. Reduced inhibition within primary motor cortex in patients with poststroke focal motor seizures. Neurology 2002;59:1028– 33. Phillips CG, Porter R. Corticospinal neurones. Their role in movement. Monographs of the Physiological Society, No. 34. London: Academic Press; 1977. Sakai K, Ugawa Y, Terao Y, Hanajima R, Furubayashi T, Kanazawa I. Preferential activation of different I waves by transcranial magnetic stimulation with a figure-of-eight-shaped coil. Exp Brain Res 1997;113: 24 –32. Wilson SA, Lockwood RJ, Thickbroom GW, Mastaglia FL. The muscle silent period following transcranial magnetic cortical stimulation. J Neurol Sci 1993;114:216–22.
The cortical silent period: intrinsic variability and relation to the waveform of the transcranial magnetic stimulation pulse M. Orth*, J.C. Rothwell Sobell Department of Motor Neuroscience and Movement Disorders, The National Hospital for Neurology and Neurosurgery, Royal Free and University College Medical School, Institute of Neurology, Box 77, Queen Square, London WC1N 3BG, UK Accepted 20 December 2003
Abstract Objective: To assess the variability of the duration of the contralateral cortical silent period (CSP) between individuals and to assess the effect of different transcranial magnetic stimulation (TMS) pulse waveforms. Methods: In Expt. 1, CSP duration, and the motor-evoked potential (MEP) amplitude and area were measured in the first dorsal interosseous muscle (FDI) of 11 subjects on 3 separate occasions using a TMS intensity of 150% active motor threshold (AMT). In Expt. 2, the stimulation intensity was varied between 100% AMT and 150% AMT. In both sets of experiments, 3 types of TMS pulse were used: monophasic posterior-anterior (PA) induced current in the brain, monophasic anterior-posterior induced current (AP), and biphasic PA/AP stimulation. Results: Experiment 1: Between-subject variation in CSP duration was high. In addition, the duration after PA stimulation was significantly shorter than after AP or PA/AP stimulation. However, there was a good correlation between CSP duration and the area, or amplitude, of the MEP. This meant that calculating the ratio of duration/amplitude or duration/area reduced intersubject variability and eliminated differences between TMS pulses. Experiment 2: increasing stimulation intensity increased the mean value of all parameters, but with significantly lower values for PA than other forms of stimulation. The ratios of duration/amplitude or duration/area did not differ between current flow directions and were relatively constant for intensities 130– 150% AMT. Conclusions: Between-subject variation in the duration of the CSP is high. A given intensity of stimulation (expressed in %AMT) produces a shorter CSP for PA stimulation than for AP or PA/AP stimulation. Significance: If the ratio (CSP duration)/(MEP size) is calculated, then intersubject variability is reduced, and TMS pulse type differences are eliminated. q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Transcranial magnetic stimulation; Cortical silent period; Variability; Recruitment; Current flow direction
1. Introduction A transcranial magnetic stimulation (TMS) pulse given to the motor cortex during tonic contraction of a target muscle produces a motor-evoked potential (MEP) followed by a period of EMG silence before the activity resumes its prestimulus baseline level of activity. This period of silence is known as the cortical silent period (CSP). The duration of the CSP depends on the intensity of the TMS pulse, but is relatively unaffected by the level of ongoing muscle activity. For CSP durations longer than about 100 ms, * Corresponding author. Tel.: þ44-20-7837-3611x3947; fax: þ 44-207278-8772. E-mail address: [email protected] (M. Orth).
the resumption of EMG activity is thought to depend on recovery of motor cortical excitability from GABAergic inhibition following the TMS pulse (Chen et al., 1999; Fuhr et al., 1991; Inghilleri et al., 1993). The duration of the CSP is often measured in both clinical and basic science studies, and is used as a measure of excitability in cortical inhibitory circuits. However, the duration is highly variable from subject to subject. In this paper we examine the variation between subjects. In the first experiment we used a standard intensity of TMS (150% AMT) to evoke the silent period in the first dorsal interosseous muscle (FDI). We then tested whether we could reduce the variability in the measure of CSP duration by combining it with a measure of MEP amplitude, since
1388-2457/$30.00 q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2003.12.025
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
1077
previous studies have reported that the two factors are linked (Wilson et al., 1993). In the same experiment we also asked whether the CSP was differentially affected by different TMS pulse waveforms. For example, it is known that monophasic posterioranterior (PA) stimulation has a lower absolute threshold for MEP recruitment than anterior-posterior (AP) stimulation (Sakai et al., 1997), and that the former tends to recruit I1 wave input to corticospinal neurones whereas the latter favours I3 recruitment (Kammer et al., 2001). It has also been reported recently that recruitment of MEPs differs with monophasic AP stimulation versus PA stimulation (Sakai et al., 1997). Given that the CSP is usually measured at a given intensity above motor threshold, we tested whether it had the same duration at an intensity of 150% AMT with all 3 pulse waveforms. In the second experiment we examined how the CSP duration and MEP size changed over a range of stimulus intensities from 100– 150% AMT.
will be referred to as PA throughout the text. When the coil handle is rotated by 1808 from this position the coil handle points forward, and the current induced in the brain flows anterior-posterior. This will be referred to as AP throughout the text. Biphasic stimulus pulses were applied through the same coil using a Magstim Super Rapid stimulator. We positioned the coil with the handle pointing backwards. The initial current flow in the brain is determined by the up-slope component of the biphasic wave so that with this coil position the current induced in the brain first flows posterioranterior and then anterior-posterior. This will be referred to as PA/AP throughout the text. Fig. 1 illustrates the different characteristics of the stimulators and the current flow direction. The optimal spot for right FDI stimulation was marked with a felt pen. Note that the maximum absolute output of the Magstim Super Rapid stimulator is smaller than that of the High Power Magstim device, so that the values of absolute stimulus intensity should not be compared between the monophasic and biphasic devices.
2. Material and methods
2.4. Motor threshold and silent period measurements
2.1. Subjects
Resting motor threshold (RMT) was defined as the minimum intensity needed to evoke an MEP of . 50 mV in 5 out of 10 consecutive trials in the relaxed FDI. Active motor threshold (AMT) was defined as the minimum intensity needed to evoke an MEP of . 200 mV in 5 out of 10 trials in the tonically active FDI (, 20% of maximal contraction as assessed visually on an oscilloscope). Thresholds were approached from above threshold in steps of 1% stimulator output. Once no MEPs could be elicited the intensity was increased in steps of 1% stimulator output until a minimal MEP was observed. This intensity was taken as motor threshold. RMT and AMT were determined separately for each stimulator. Cortical silent periods were recorded from the tonically active right FDI.
Eleven healthy right handed Caucasian volunteers were studied (mean age 29 years, range 20– 37 years, 6 women). Subjects were asked to refrain from caffeine-containing beverages on the day of the experiment. All subjects participated in Expt. 1; 10 subjects participated in the second experiment. Subjects gave informed written consent, and the local ethics committee approved the study protocol. 2.2. Electromyography recordings Surface electromyograms (EMG) were recorded from the right FDI using silver/silver-chloride disc surface electrodes (1 cm diameter) in a belly tendon montage. The EMG signal was amplified and analogue filtered (30 Hz to 3 kHz) with a Digitimer D150 amplifier (Digitimer Ltd., Welwyn Garden City, UK). Data (sampling rate 4 kHz) was digitized for offline analysis using Signal software (Cambridge Electronic Devices, Cambridge, UK). Peak-to-peak amplitude of MEPs, the area under the curve of the MEP and the silent period duration were measured with in-house software. 2.3. Transcranial magnetic stimulation Subjects were seated in a comfortable chair. Magnetic stimuli were given with a hand-held figure-of-8 coil (outer winding diameter 9 cm) connected to a High Power Magstim 200 stimulator (Magstim Co., Whitland, Dyfed, UK). This stimulator generates a magnetic pulse with monophasic waveform and in the brain induces a current with posterior-anterior flow when the coil handle is positioned at an angle of 458 pointing backwards. This
Fig. 1. Illustration of the coil positions, magnetic waveforms generated by the stimulators and the current flow directions induced in the brain. PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/ anterior-posterior.
1078
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
In Expt. 1, a fixed test stimulus intensity of 150% AMT was used and the session repeated twice (total of 3 sessions). In the second experiment, the test stimulus intensity was varied in each subject (100, 110, 120, 130, 140, 150% of the AMT measure with the respective stimulator). The order of the experiments was randomized. For each test stimulus intensity, 10 trials were collected in each subject in a single session with an interval of 4 s between trials. In each individual trial the duration of the silent period was measured from the beginning of the MEP evoked by the test stimulus to the resumption of (any level of) sustained EMG activity. In addition, the peak-to-peak amplitude and the area under the test stimulus MEP were determined. 2.5. Data analysis The following parameters were examined: CSP duration, MEP amplitude, MEP area. In addition, ratios were calculated from CSP duration and MEP amplitude, or MEP area, respectively. In the first experiment, we examined how variable CSP duration, MEP amplitude or MEP area were between sessions. For this purpose, a coefficient of variation was calculated from the mean of each session and the corresponding standard deviation (CV ¼ session SD/session mean £ 100). We went on to test how reproducible the results were between sessions. To this end, a repeated-measures analysis of variance (ANOVA) model was used with ‘session’ and ‘stimulator’ as the within-subject factors. As the difference between the sessions was not significant for any of the stimulation conditions, data from all 3 sessions were pooled for each subject. Next, we examined whether CSP duration, MEP amplitude or MEP area differed between subjects by calculating a CV for either CSP duration, or MEP amplitude or MEP area, as main factor. For each stimulator, a regression analysis was used to test how accurately CSP duration was predicted by either MEP amplitude or MEP area. In the second experiment, we sought to investigate the effect of different stimulus intensities on CSP duration, MEP amplitude or MEP area. For this purpose, we used a repeated-measures ANOVA model with ‘coil current’ and stimulus ‘intensity’ as within-subject factors. As the dataset for intensities smaller than 130% was incomplete, we restricted the analysis to the intensities 130, 140 and 150% AMT. A statistical difference in the ANOVAs was followed by a post hoc paired t test analysis with Bonferroni correction of the P value. Mauchly’s test examined sphericity in the repeated measures ANOVA models. Statistical significance levels were set to P ¼ 0:05, for non-spherical data the Greenhouse-Geisser correction was used. All statistical analysis was performed using SPSS 11 for Windows software package.
3. Results 3.1. Experiment 1: stimulation at 150% AMT With PA current flow, subjects had a mean RMT of 37% of stimulator output (SD 4.5, range 30 – 47) and a mean AMT of 26.9% (SD 4.7, range 17– 35). With AP current flow the mean RMT was 46.8% (SD 8, range 39 –63) and the mean AMT was 38.4% (SD 6.4, range 29 –51). With PA/AP current flow, subjects had a mean RMT of 49.3% (SD 6.5, range 39– 63) and a mean AMT of 38.9% (SD 6, range 29– 52). 3.1.1. Between-subject variability The results for all CSP and MEP parameters were very similar between the 3 sessions in each subject (repeated measures ANOVA, P . 0:05) with the between-session CV always smaller than 10%. Thus the data from the 3 sessions were pooled for statistical analysis (Table 1). There was considerable variability between subjects in the measures of ‘amplitude’ and ‘area’ of MEP and ‘duration’ of CSP for each of the current flow directions (Fig. 2) with the betweensubject CV ranging from 20.0 to 35.3% (Table 2). 3.1.2. Effect of stimulus waveform The duration of the cortical silent period and the area under the MEP differed significantly depending upon the direction of current flow (repeated-measures ANOVA, Table 2: P ¼ 0:027, F ¼ 4:44, df 2,18 and P ¼ 0:015, F ¼ 5:35, df 2,18, respectively). Post hoc analysis revealed that the CSP duration was shorter for PA than for either AP or PA/AP stimulation (paired t test, P ¼ 0:006 and P ¼ 0:05, respectively), and that the MEP area was smaller for PA than PA/AP stimulation (paired t test, P ¼ 0:001). There was no significant effect of waveform on MEP amplitudes, although as with area measures, the peak-topeak amplitude of the MEP did tend to be smaller for PA than for AP or PA/AP stimulation. The lack of effect may well have been due to saturation of the MEP amplitude in some subjects at 150% AMT, making area a better measure of response size. For each direction of current, the MEP amplitude and area significantly predicted the duration of the cortical silent Table 1 Coefficient of variation of CSP duration, MEP amplitude and MEP area Between-session variability
Duration MEP amplitude MEP area
PA
AP
PA/AP
3.8 7.1 6.1
4.7 7.3 8.4
4.6 6.1 7.1
The between-session CV was calculated from the median of the CV of 3 sessions in each of 11 subjects. PA, posterior-anterior; AP, anteriorposterior; PA/AP, posterior-anterior/anterior-posterior.
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
1079
Fig. 2. Correlation of CSP duration with MEP area (A,C,E) or MEP amplitude (B,D,F), for PA current flow direction (A,B), AP current flow direction (C,D) and PA/AP current flow direction (E,F). Each data point represents the mean of 3 sessions in each subject (n ¼ 11). PA, posterior-anterior; AP, anteriorposterior; PA/AP, posterior-anterior/anterior-posterior.
Table 2 The effect of current flow direction in the brain at a fixed stimulation intensity of 150% AMT on CSP duration and MEP amplitude or MEP area, and on ratios calculated from CSP duration and the corresponding MEP amplitude or area, respectively PA
Duration (ms) MEP amplitude (mV) MEP area (mV ms) Duration/amplitude (ms/mV) Duration/area (ms/mV ms)
AP
PA/AP
Mean (SD)
CV
Mean (SD)
CV
Mean (SD)
CV
108.0 (38.1) 7.1 (1.7) 620.9 (215.3) 15.4 (2.2) 0.18 (0.02)
35.3 24.1 34.7 14.2 12.8
139.2 (30.8) 8.0 (2.0) 746.8 (223.4) 19.0 (5.2) 0.2 (0.05)
22.1 25.0 29.9 27.4 26.0
132.0 (37.3) 8.0 (1.6) 776.4 (205.5) 16.8 (2.7) 0.18 (0.03)
28.3 20.0 26.5 16.4 17.5
Results are from the mean of 3 repeated measurements in each of 11 subjects. CV, coefficient of variation; PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/anterior-posterior.
1080
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
period (Fig. 2, regression analysis). The r 2 values were greater for PA (Fig. 2A,B) compared with AP (Fig. 2C,D) or PA/AP (Fig. 2E,F) current. The strong correlation of CSP duration with MEP amplitude, and MEP area, led us to calculate ratios of duration/MEP amplitude and duration/MEP area (Table 2) for each direction of current. The ratios did not differ between different directions of current flow (repeatedmeasures ANOVA, P . 0:1 for both). In order to analyse whether calculation of the ratio duration/area reduced the between-subject variability, we went on to compare the CV of CSP duration with the CV of the corresponding CSP duration/MEP area, or MEP amplitude, ratio for each of the current flow directions. For PA and PA/AP current flow direction, the calculation of the duration/MEP area, or MEP amplitude ratio reduced intersubject variability whereas the intersubject CV remained high for AP current flow (Table 2). 3.2. Experiment 2: the effect of different stimulation intensities In the second experiment we varied the stimulation intensities to assess whether the differences that we had seen with stimulation at 150% AMT also occurred at other intensities. Since the data set was incomplete for lower stimulus intensities, we only analysed that from 130– 150% AMT in detail. Separate two-factor ANOVAs with intensity and coil current as main factors were performed on MEP and CSP measures. This revealed a significant main effect of stimulus intensity on the duration of the cortical silent period (Fig. 3A, P , 0:001, F ¼ 65:4, df 1.18, 10.63), MEP amplitude (data not shown, P ¼ 0:002, F ¼ 9:1, df 2,18) and MEP area (Fig. 3B, P , 0:001, F ¼ 15, df 2,18). All 3 parameters increased significantly with increasing stimulation intensities. There was also a significant main effect of coil current on the duration of the cortical silent period (Fig. 3A, P ¼ 0:002, F ¼ 9:2, df 2,18) and MEP area (Fig. 3B, P ¼ 0:019. F ¼ 5:0, df 2,18), with a marginal effect on MEP amplitude (P ¼ 0:073, F ¼ 3:03, df 2,18). There was no significant interaction, suggesting that responses to all 3 current directions were influenced by intensity in the same way. Post hoc pairwise comparisons on the main effect of coil current showed that the duration of the CSP was shorter for PA than AP (P ¼ 0:005, F ¼ 13:9, df 1,9) or PA/AP (P ¼ 0:011, F ¼ 10, df 1,9), and the MEP area was smaller with PA compared with AP (P ¼ 0:032, F ¼ 6:4, df 1,9) or PA/AP (P ¼ 0:009, F ¼ 10:8, df 1,9). MEP amplitude was also significantly smaller with PA compared with PA/AP (P ¼ 0:044, F ¼ 5:5, df 1,9) while the difference between PA and AP was marginal (P ¼ 0:051, F ¼ 5, df 1,9). Comparison of the ratios of either CSP duration/MEP area or CSP duration/MEP amplitude revealed no difference between stimulation intensities or current flow directions (Fig. 3C).
Fig. 3. The effect of different stimulation intensities on CSP parameters. At stimulation intensities smaller than 130% AMT, the data set is incomplete (segment on the left). At stimulation intensities above 120% AMT, (segment on the right) CSP duration (A) and MEP area (B) are significantly smaller with PA current flow (straight line) compared with AP (regularly dotted line) or PA/AP (irregularly dotted line) while the ratio calculated from CSP duration and MEP area is not different (C). Values are means of a single session in 10 subjects, error bars are SDs. PA, posterior-anterior; AP, anterior-posterior; PA/AP, posterior-anterior/anterior-posterior.
4. Discussion Most commonly, the cortical silent period is recorded with the subject voluntarily contracting the target muscle between 20 and 30% of the maximal force, and with a test stimulus intensity of around 130% RMT or 150% AMT (Kessler et al., 2002). In the present study we used such a conventional experimental set up in order to avoid fluctuations of voluntary contraction and saturation of the MEP.
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
The data from Expt. 1 confirmed a previous report showing that motor thresholds with PA current flow are lower than with AP current flow in the brain (Sakai et al., 1997). The data also revealed 4 important features of the CSP: (1) repeated measures of CSP duration (and MEP size) are relatively stable between sessions (coefficient of variation less than 10%) for any one individual; (2) between-subject variation is much higher for all measures; (3) for a given intensity of stimulus (i.e. 150% AMT), CSP duration, MEP area and MEP amplitude are all smaller with PA than other forms of stimulation; (4) for PA and PA/AP current flow direction the ratios CSP duration/MEP amplitude or CSP duration/MEP area are much less variable between subjects than the raw measures alone as there is a strong correlation between CSP duration and the size of the preceding MEP. Data from Expt. 2 indicate that these features are true for the range of stimulus intensities from 130 to 150% AMT. Feature (1) may mean that measures of CSP duration are adequate in repeated measures studies where the effect of a certain treatment may be evaluated on a given group of subjects. Feature (2) indicates that measures of CSP duration are much less sensitive in revealing differences between different groups of subjects and feature (3) implies that we should not compare the duration of the CSP or size of MEP made with different stimulus waveforms, even if the stimulus intensity has been adjusted to the same intensity relative to threshold in each case. Feature (4) suggests that a way to improve the power of between-subject comparisons is to calculate the ratio of CSP duration to MEP size (either amplitude or area). Experiment 2 went on to test how well this reasoning would apply at different intensities of stimulation between 100 and 150% AMT. Higher intensities were not studied because of the risk of saturating MEP size. The data showed that PA stimulation gave lower values for all parameters for all intensities, especially from 130 to 150% AMT. However, the ratio of CSP duration to MEP size was fairly constant across the intensities 130 –150% AMT which are conventionally used for CSP measurements. 4.1. Correlation between CSP duration and MEP size It has been recognized in the past that CSP duration and MEP amplitude both relate to the intensity of the stimulation (Wilson et al., 1993). In our study we observed that this correlation holds between subjects even at a fixed intensity of stimulation, and is independent of the pulse waveform of the TMS stimulus. Thus, subjects who have a large MEP to a given stimulus also tend to have a longer CSP. The relation is much stronger for PA than for PA/AP, and is relatively weak for AP current flow. The good correlation between MEP area and CSP duration led us to calculate ratios of CSP duration and the corresponding MEP area, and this reduced both the between-subject and the between-stimulator variability for
1081
PA and PA/AP current flow direction. This suggests that the factors that cause variation in the MEP are the same as those that cause variation in the duration of the CSP. One simple possibility is that the corticospinal outflow that produces the MEP is also responsible for the CSP. Corticospinal neurones give off recurrent axon collaterals that have a range of effects on cortical circuits. In particular, in the cat, recurrent collaterals of fast corticospinal axons have a predominantly inhibitory effect on the firing of neurones that have slower conducting axons (Phillips and Porter, 1977). Neurones with large axons appear to be activated by TMS pulses, and if tonic voluntary contractions are supported by the activity of slow axoned neurones, this could explain the occurrence of the silent period. However, further studies are needed to test this hypothesis in detail. 4.2. The difference between PA and AP or PA/AP stimulation We observed the strongest relation of MEP size and CSP duration for PA stimulation. The relation was good for PA/AP but weak for AP current flow direction. In addition, with stimulation intensities from 130 to 150% AMT all parameters were significantly smaller with PA current flow compared to AP or PA/AP current flow even though the ratios of CSP duration/MEP amplitude, or CSP duration/ MEP area, remained very similar. This difference, in the recruitment curves between different stimulus waveforms, was unexpected. One explanation might be that different current flow directions preferentially activate different subsets of cortical neurones and that the properties of these neuronal populations are different. Indeed, it is known that PA stimulation tends to recruit I1 input to corticospinal neurones whereas AP stimulation tends to recruit I3 inputs (Kammer et al., 2001). If the recruitment curves of these elements differed then it might account for the effects observed on the MEP measures. The effect on the CSP duration would be secondary to this effect if it were caused, as argued above, by the same corticospinal output as determined the MEP. Whatever the explanation, it is clear that not only should CSP durations not be directly compared using different stimulus waveforms, but neither should MEP parameters. Nevertheless, the ratio between these values may be of interest. For example, in the simple model above in which PTN discharge produces inhibition of neighbouring PTNs, an increase in MEPs without a corresponding lengthening of the CSP would indicate that there had been a reduction in excitability of the recurrent inhibitory pathway. Alternatively, in a model where separate mechanisms produce MEPs and CSP, the same change in ratio would be compatible with an independent effect on excitatory and inhibitory circuits in the cortex. In conclusion, we would like to suggest that the calculation of ratios of CSP duration and MEP area can
1082
M. Orth, J.C. Rothwell / Clinical Neurophysiology 115 (2004) 1076–1082
help to reduce the between-subject and between-stimulator variability of silent period measures. Acknowledgements We wish to thank our subjects for their participation in this study. The work was funded by the Medical Research Council. References Chen R, Lozano AM, Ashby P. Mechanism of the silent period following transcranial magnetic stimulation. Evidence from epidural recordings. Exp Brain Res 1999;128:539 –42. Fuhr P, Agostino R, Hallett M. Spinal motor neuron excitability during the silent period after cortical stimulation. Electroenceph clin Neurophysiol 1991;81:257–62.
Inghilleri M, Berardelli A, Cruccu G, Manfredi M. Silent period evoked by transcranial stimulation of the human cortex and cervicomedullary junction. J Physiol 1993;466:521 –34. Kammer T, Beck S, Thielscher A, Laubis-Herrmann U, Topka H. Motor thresholds in humans: a transcranial magnetic stimulation study comparing different pulse waveforms, current directions and stimulator types. Clin Neurophysiol 2001;112:250–8. Kessler KR, Schnitzler A, Classen J, Benecke R. Reduced inhibition within primary motor cortex in patients with poststroke focal motor seizures. Neurology 2002;59:1028– 33. Phillips CG, Porter R. Corticospinal neurones. Their role in movement. Monographs of the Physiological Society, No. 34. London: Academic Press; 1977. Sakai K, Ugawa Y, Terao Y, Hanajima R, Furubayashi T, Kanazawa I. Preferential activation of different I waves by transcranial magnetic stimulation with a figure-of-eight-shaped coil. Exp Brain Res 1997;113: 24 –32. Wilson SA, Lockwood RJ, Thickbroom GW, Mastaglia FL. The muscle silent period following transcranial magnetic cortical stimulation. J Neurol Sci 1993;114:216–22.