G Model
ARTICLE IN PRESS
NSL 31395 1–5
Neuroscience Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Research article
1
Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction
2
3
4
Q1
5
Katja Airaksinen a,b,∗ , Tuuli Lehti a , Jussi Nurminen a , Jarkko Luoma a , Liisa Helle c , Samu Taulu c,d,e , Eero Pekkonen f , Jyrki P. Mäkelä a a
BioMag Laboratory, HUS Medical Imaging Center, University of Helsinki and Helsinki University Hospital, Finland Clinical Neurosciences, Neurology, University of Helsinki and Helsinki Univer-sity Hospital, Finland c Elekta Oy, Helsinki, Finland d Institute for Learning and Brain Sciences, University of Washington, Seattle, USA e Department of Physics, University of Washington, Seattle, USA f Department of Neurology, University of Helsinki and Helsinki University Hospital, Finland
6
b
7 8 9 10 11 12
h i g h l i g h t s
13 14
• Corticokinematic coherence (CKC) can be measured during isometric hand activity. • Cortico-muscular coherence (CMC) and CKC of the postural tremor resemble each other. • The sources of CMC and CKC of postural tremor colocalize in the cortex.
15 16 17 18
a r t i c l e
19 32
i n f o
a b s t r a c t
20
Article history: Received 17 February 2015 Received in revised form 7 May 2015 Accepted 17 June 2015 Available online xxx
21 22 23 24 25 26
31
Keywords: Cortico-muscular coherence Corticokinematic coherence Magnetoencepha-lography Postural tremor
33
1. Introduction
27 28 29 30
34Q3 35 36 37 38 39 40 41
Q2
Corticokinematic coherence (CKC), i.e., coherence calculated between MEG and an accelerometer signal recording movement kinematics can be used for functional mapping of the sensorimotor cortex. Cortical sources of CKC, induced by both voluntary and passive movements, localize at the proximity of sensorimotor cortex. We tested the CKC during a static contraction to compare it with simultaneously measured cortico-muscular coherence (CMC) estimated between MEG and surface EMG to study the role of postural tremor in CMC in ten healthy volunteers. CKC was detectable also during this static task. CKC and CMC spectra had similar power distributions, and sources of CMC and CKC colocalized at the cortex in close proximity of the central sulcus. During the static hold task, the accelerometer signal originates from the postural tremor. The similarities between CMC and CKC indicate that postural tremor is related to CMC in healthy subjects. © 2015 Published by Elsevier Ireland Ltd.
Physiological tremor is considered a general property of the neuromuscular system. It occurs during normal conditions, and is increased e.g., by stress [1]. Physiological tremor originates from complex interactions between neural events and limb-related mechanical properties [2]. However, it is unclear whether the neural or the mechanical properties are most prominent factors in the tremor. Synchronous oscillations in the central nervous system may play an important role in physiological tremor (e.g., [3]) or con-
∗ Corresponding author at: BioMag Laboratory, P.O. Box 340, FIN-00029 HUS, Finland. Fax +358 9 47175781. E-mail address: katja.airaksinen@helsinki.fi (K. Airaksinen).
tribute to dominant peripheral resonance caused by the mechanical properties of the oscillating limb [4]. Nevertheless, synchronous central input may not be necessary for finger physiological tremor [5]. Physiological tremor can be enhanced by posture; this postural tremor appears during static maintenance of position against gravity [6]. Postural tremor is found in healthy controls, and is induced in the fingers with static tension of the wrist extensor muscles. Postural tremor differs from resting tremor which appears in a relaxed limb and attenuates with limb movement, and from kinetic tremor emerging during voluntary movement, as well as from isometric tremor which appears when the limb is contracted against a constant resistance [1]. Cortico-muscular coherence (CMC) is estimated between simultaneously measured magnetoencephalography (MEG) or electroencephalography (EEG) and electromyography (EMG)
http://dx.doi.org/10.1016/j.neulet.2015.06.034 0304-3940/© 2015 Published by Elsevier Ireland Ltd.
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
G Model
ARTICLE IN PRESS
NSL 31395 1–5
K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
2
Table 1 The mean peak frequencies and maximum amplitudes of CMC and CKC (±standard deviation). Also the pooled CMC and CKC are shown. The mean peak frequency (Hz) ± SD CMC
The mean max amplitude ± SD
CKC
Left hand
Right hand
Left hand
Right hand
21.9 ± 7.9 p = 0.26
24.8 ± 7.4
18.0 ± 5.6 p = 0.049
23.6 ± 6.0
Pooled CMC
Pooled CKC
23.3 ± 7.6 p = 0.013
20.7 ± 6.3
recordings. CMC at the 12–33 Hz frequency band has been found between motor cortex and different muscles during isometric contraction in healthy subjects (e.g., [7–9]). CMC is thought to reflect 59 the interplay between cortex and muscle in motor coordination and 60 to reflect central drive to the spinal motoneuron pool (e.g., [10]). 61 More recent studies have emphasized both the efferent descend62 ing drive from the motor cortex to the muscle as well as ascending 63 afferent drive from the muscles to the cortex in producing the CMC 64 [11,12]. However, more detailed information about its physiologi65 cal and pathophysiological mechanisms is lacking. 66 Conway et al. [7] and Salenius et al. [8] have observed coherence 67Q4 also at frequencies outside the beta band. The EMG spectra showed 68 peaks roughly at 10 Hz (alpha), 20 Hz (beta), and 40–50 Hz (gamma 69 band), and significant coherence was present at all three bands. 70 However, alpha and gamma band coherence peaks were occasional 71 findings, only present in minority of subjects. Alpha band coher72 ence has been associated with Parkinson’s disease (PD) tremor [13], 73 whereas gamma band coherence has been linked with very strong 74 muscle contractions [14]. 75 Corticokinematic coherence (CKC) was recently introduced as 76 an alternative method for the functional mapping of the motor 77 cortex [15]. CKC is estimated between MEG or EEG and a three78 axis accelerometer signal recording the kinematics of movement. 79 When subjects performed self-paced repetitive finger movements, 80 a coherence peak was observed at the movement frequency and 81 its first harmonic. The source of the CKC lies either close to the 82 “hand knob” of the primary motor cortex or slightly posterior to it in 83 the primary sensory cortex. As passive finger movements induce a 84 similar CKC, it may reflect somatosensory, probably proprioceptive 85 input to the sensorimotor cortex [16]. 86 Sensitive accelerometers detect the slight postural tremor 87 observed in healthy individuals. [2,4,17]. If postural hand tremor is 88 involved in modification of CMC, the comparison between CMC and 89 CKC during a hold task could potentially reveal a common cortical 90 component. We measured CMC and CKC during wrist dorsiflexion, 91 generating a sustained contraction of arm muscles in healthy con92 trols. To our knowledge, no previous studies have compared CMC 93 94 and CKC during static muscle contraction. 57 58
95
96 97
2. Methods Ten healthy subjects (6 male and 4 female) aged between 22 and 58 years (mean age 30 years) were instructed to extend their wrist
CMC
CKC
Left hand
Right hand
Left hand
Right hand
0.088 ± 0.05 p = 0.953
0.111 ± 0.06
0.062 ± 0.04 p = 0.14
0.090 ± 0.04
Pooled CMC
Pooled CKC
0.099 ± 0.05 p = 0.059
0.075 ± 0.04
(dorsiflexion) for one minute at a time. The task was repeated five times with a pause of 20 s between each trial. The aim was to produce a sustained isometric contraction by using submaximal force. In a preliminary trial before the actual task the subjects extended their wrist with a maximal force to establish a baseline for the contraction force. The subjects performed the task first with the right and then with the left hand. To obtain an EMG of the muscles activated in the motor task, a bipolar surface electrode was attached to the forearm on top of the extensor carpi radialis muscle. To record finger tremor related to muscle tension, an accelerometer (ADXL335 iMEMS Accelerometer, Analog Devices Inc., Norwood, MA, USA) was attached on the nail of the index finger in both hands. The accelerometers measured finger acceleration in three orthogonal directions. The subjects were seated under a 306-channel neuromagnetometer (Elekta Neuromag® , Elekta Oy, Helsinki, Finland) to measure the brain activity during the task. The head position with respect to the MEG helmet was recorded continuously using the vendor-supplied head position indicator system. The recording passband was 0.03–330 Hz and the sampling rate 1012 Hz for MEG, EMG and accelerometers. The EMG signal was not rectified. The spatiotemporal signal space separation (tSSS) [18] method implemented in MaxFilterTM software (Elekta Oy) with an 8-s time window and a subspace correlation limit of 0.9 [19] was applied to the MEG data to suppress the external and close-by artifacts before data analysis. Based on the singular value decomposition of the three orthogonal accelerometer channels, we extracted the signal corresponding to most significant singular value and used the resulting signal in subsequent analysis. MEG and anatomical MRI were co-registered using the MRILab software (Elekta Oy). The matching provided a coordinate transformation between the head coordinates and the MRI coordinates. MR images were segmented to distinguish the brain from the surrounding tissues to provide volume for possible sources using the FreeSurfer software (MGH, Boston, MA) [20] or Seglab software (Elekta Oy). Power spectral density (PSD) of MEG, accelerometer and EMG signals were estimated for each subject using Welch’s method with 50% overlapping 1024-point Hanning windows, resulting in a frequency resolution of approximately 1 Hz. The sensor-level coherence was estimated between the EMG or accelerometer signal and each MEG gradiometer pair in turn. Also the coherence between the EMG and the accelerometer sig-
Table 2 The mean coordinates of x, y and z for CMC and CKC in the upper part of the table. In the lower part are the mean differences between CMC and CKC coordinates and the confidence intervals of these differences. The mean x-coordinate (mm) CMC
The mean y-coordinate (mm) CKC
35.5 ± 9.0 37.2 ± 8.2 p = 0.345 The mean difference −1.65 ± 5.2 95% confidence interval of the difference Upper Lower 0.8 −4.1
The mean z-coordinate (mm)
CMC
CKC
CMC
CKC
11.9 ± 8.7 p = 0.807
12.0 ± 8.1
98.7 ± 6.0 p = 0.701
98.4 ± 5.9
−0.05 ± 4.4 Lower −2.1
0.24 ± 3.3 Upper 2.0
Lower −1.3
Upper 1.8
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
G Model NSL 31395 1–5
ARTICLE IN PRESS K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
3
Q5 Fig.1. Localization of CMC and CKC. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Q6 A: The localization of CMC (left) and CKC (right) in same subject as in this Figure overlaid on the MR image of the subject. B: The mean (±1 standard deviation) values of x and y (left) and x and z (right) coordinates of all subjects. CMC in red and CKC in blue.
141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
nals were determined. The 50% overlapping 1024-point Hanning windows were used. The number of averaged segments used for coherence calculations varied (range 336–618; mean 556.3 ± 53.7). The time segments were the same for both EMG and accelerometer data. The location of the CMC/CKC maximum peak was then determined by visual inspection from a selection of 15 gradiometer pairs over the sensorimotor cortex contralateral to the activated hand. The coherence spectra were then inspected. Peak amplitudes and frequencies were determined from the planar gradiometer pair showing the largest response over the sensorimotor cortex. Timeshifted coherence (coherence calculated by shifting EMG data 3 s in relation to the MEG data) was used for defining the statistical significance level between 5 and 45 Hz. This time shift destroys any true MEG-EMG coherence in the data. The significance level was set at 99% of the time-shifted coherence [13]. Source-space coherence was estimated using the Dynamic Imaging of Coherence Sources (DICS) method in Beamformer software (Elekta Oy). We selected 6-Hz frequency bands based on the sensor-level CMC and CKC coherence peaks, for coherence estimation and evenly distributed grid of locations covering the brain volume with a resolution of 3 mm for source space. For each location, only the direction with maximum current power was used. After the beamformer scanning, we estimated a virtual electrode signal, i.e. a time domain estimate of source activity, from the location of maximum coherence; the same beamformer weights were used to extract the virtual electrode signal and only results within the previously selected 6-Hz frequency band were used in analysis. We then computed the final CMC and CKC results using these virtual electrode signals.
Statistical calculations were done with SPSS (SPSS for Windows version 13.0, SPSS Inc., Chicago, Illinois, United States). Correlations were calculated with Spearman’s rho. The Wilcoxon Signed Ranks Test was used for comparison of two dependent groups. The significance and confidence intervals for coordinate differences were calculated with One-Sample T-Test. The level of significance was set at 0.05. The results are reported as mean ± standard deviation when appropriate. For statistical calculation the absolute values of the x-coordinates were used. 3. Results Significant CMC was found in 19 out of 20 hemispheres; CKC was seen in every hemisphere. The 19 hemispheres presenting both CMC and CKC were used for analyses. The frequency bands of the CMC and CKC peaks overlapped at least in part in 15 out of 19 hemispheres (Supplementary Fig. S1). In those cases the frequency distributions of the CMC and CKC resembled each other strongly. One subject had most prominent CMC and CKC peaks at 10 Hz (Supplementary Fig. S1). As the coherence values between the hemispheres did not differ, we pooled the CMC and CKC of both hemispheres together. Their mean peak frequency was 23 Hz for CMC and 21 Hz for CKC (p = 0.013) (see Table 1). The peak frequencies of CMC and CKC correlated significantly (rs = 0.681, N = 19, p = 0.001, two tailed). Supplementry material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.neulet.2015.06.034 The pooled mean maximum CMC amplitude was 0.1 ± 0.05 and mean maximum CKC amplitude 0.08 ± 0.04; the difference was non-significant (p = 0.059) (see Table 1). Also maximum CMC
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
171 172 173 174 175 176 177 178 179
180
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
G Model
ARTICLE IN PRESS
NSL 31395 1–5
K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
4
Coherence
A
B
0.1
CMC CMC, sign. CKC CKC, sign.
0.05
0 1
Coherence
C
D
peripheral coherence peripheral coherence, sign.
0.8
0.6
0.4
0.2
0
5
10
15
20
25
30
35 Hz
5
10
15
20
25
30
35
Hz
Fig. 2. A and B: Cortico-muscular (CMC, red line; significance level thin horizontal red line) and corticokinematic (CKC, blue dashed line; significance level horizontal blue dotted line) coherences from both hemispheres of one subject. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) C and D: Peripheral coherence between EMG and accelerometer (black line, horizontal dotted black line for significance level). A and C left hand activation, B and D right hand activation. CKC and CMC in the 15–25 Hz range are notably similar. The main peak of peripheral coherence occurs around 10 Hz. Note different y-axis scales for central and peripheral coherence graphs. The time-shifted coherence between 5 and 45 Hz is used for the significance levels.
199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
amplitude and maximum CKC amplitude correlated significantly (rs = 0.575, N = 19, p = 0.010, two tailed). The coordinates of the source locations of CMC and CKC did not differ significantly (see Table 2, Fig. 1) within different hemispheres. The difference between CMC and CKC x-, y, and z-coordinates was calculated. One-Sample T-Test did not reveal significant difference between the remainders of x-, y, and z-coordinates. The 95% confidence intervals of the difference are shown in Table 2. The tremor frequency (the peak frequency of accelerometer power spectrum between 6 and 14 Hz) did not correlate with CMC (rs = −0.178, N = 19, p = 0.465, two tailed) or CKC (rs = 0.045, N = 19, p = 0.854, two tailed) peak frequencies. The ratio between EMG variation during task and maximum wrist extension did not correlate with the CMC (rs = 0.004, N = 19, p = 0.986, two tailed) or CKC (rs = 0.120, N = 19, p = 0.624, two tailed) frequency. The coherence in the 5–11 Hz band was inspected on sensor signals. In channels having the strongest coherence, significant CMC was found in 12 hemispheres and CKC in 11 hemispheres in the 5–11 Hz band. Seven subjects had significant peaks within 1 Hz both in CMC and CKC in the 5–11 Hz band. When the CMC and CKC were compared between individuals, the coherence spectra had considerable variability. This variability was less prominent when the right and left hemisphere of a given subject were compared (Supplementary Fig. 1). To evaluate the joint properties of accelerometer and EMG signals, we estimated the coherence between EMG and accelerometer signals. This peripheral coherence peaked between 7 and 12 Hz in all but one hand. In one hand the maximum peak occurred at 22 Hz but a coherence peak at 12 Hz was present as well. Another peak around 20–30 Hz was seen in six subjects. The mean peak
coherence amplitudes were clearly stronger than the CMC or CKC values (Fig. 2). The peripheral coherence in right and left hand did not differ; the mean peak frequencies were 10.3 ± 4.0 Hz for the left and 9.4 ± 1.2 Hz for the right hand (p = 0.22), the mean peak amplitudes were 0.6 ± 0.2 and 0.7 ± 0.1 respectively (p = 0.21). The peak frequency or peak amplitude of the peripheral coherence did not correlate with the corresponding values of CMC or CKC. The peripheral coherence frequency correlated with the postural tremor frequency estimated from the accelerometer (rs = 0.489, N = 19, p = 0.034, two tailed). 4. Discussion In addition to detecting voluntary finger movements [15,16], the accelerometer was sensitive enough to detect the small postural tremor of our healthy subjects (Supplementary Figs. S2 and S3). The MEG/EEG signals are known to be coherent with the motor unit spike trains or the EMG signal during static contractions. Assuming a linear model, the acceleration signals are a low-pass filtered version of the summation of all motor unit spike trains. Therefore, it is not surprising that a significant coherence between the brain and accelerometer signals may be observed. The accelerometer is easier to use than EMG and its signal is not as sensitive to recording position than EMG which also has limitations, such as band-pass behavior and spectral properties that depend on the type of recording. Thus CKC may provide an attractive alternative for studying motor control. Repeatability of CMC between sessions is only moderate [21]; further study will reveal, for example, if intersession repeatability of CKC is better. Supplementry material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.neulet.2015.06.034
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
230 231 232 233 234 235 236 237 238 239
240
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
G Model NSL 31395 1–5
ARTICLE IN PRESS K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
300
The peak frequencies and maximum amplitudes of CMC and CKC correlated significantly, and their maximum amplitudes did not differ. The beamformer source localization results indicate that CMC and CKC maxima colocalize at the cortex in close proximity of the central sulcus. As CMC and CKC resemble each other closely, and the kinematics during the static hold task are generated by postural tremor, we suggest that CMC is related to postural tremor. The CMC and CKC spectra were similar but not identical; this is not surprising as the postural tremor probably relates to activity in several muscles whereas we recorded EMG only from the ECR muscle. Similar to previous studies, our results show a large individual variation in CMC frequency and magnitude. The distributions of CMC and CKC spectra, however, were remarkably similar within subjects (Fig. 1; Supplementary Fig. S1). This suggests that although the drive between the cortex and the muscles is relatively individually organized, the selected strategy is the same for both hemispheres. The accelerometer power spectra revealed a low-frequency peak in nine out of ten subjects between 5 and 11 Hz (Supplementary Fig. S3) indicating the presence of small-amplitude postural tremor, in line with previous reports [2,4,17]. Although the peak coherence amplitude, the main target of our study, occurred at 18–22 Hz for both CMC and CKC, coherence exceeding the noise level in 5–11 Hz band was also present in 11–12 hemispheres in sensor level coherence. It may be assumed that the peripheral and cortical activities (CMC and CKC) should have coherence at the frequency most prominent in the peripheral signals. However, Parkinsonian [22] and essential [23] tremors as well as Parkinsonian tremor imitated by healthy subjects [22] induce CMC at the tremor frequency and its first harmonic. The same holds true for CKC induced by voluntary movements [16]. Therefore it is not surprising that the coherence between postural tremor and cortical signals occurs both at 5–11 Hz and at 10–22 Hz. Peripheral mechanical-reflex component of postural tremor, not related to cortical activation, also contributes strongly to the accelerometer signals at 5–8 Hz (e.g., [24]). The CMC has been extensively used in studies of Parkinson’s disease. The first suggestion of a link between postural tremor and CMC at 12–18 Hz was made on the basis of an increased CMC in a group of PD patients displaying postural tremor [24]. The relationship with CMC and postural tremor should be kept in mind when making inferences of CMC modifications by various therapies of PD.
301
Financial disclosures
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
302 303 304 305 306 307 308 309
310
311 312
Katja Airaksinen received research funding from Finnish Brain Foundation and the Finnish Parkinson Foundation. Other authors reported no biomedical financial interests or potential conflicts of interests. This work was supported by the Academy of Finland (grant n.o. 122,725) and by the SalWe Research Program for Mind and Body (Tekes – the Finnish Funding Agency for Technology and Innovation grant 1104/10). Acknowledgement We want to thank Dr. Veikko Jousmäki for providing the accelerometer device.
5
References
313
[1] H.W. Berendse, T. van Laar, Tremor, in: E. Wolters Ch, T. van Laar, H.W. Berendse (Eds.), Parkinsonism and Related Disorders, VU University Press, Amsterdam, 2007, pp. 309–325. [2] S. Morrison, K.M. Newell, Postural and resting tremor in the upper limb, Clin. Neurophysiol. 111 (2000) 651–663.
314
[3] A.R. Mehta, J.-S. Brittain, P. Brown, The selective influence of rhythmic cortical versus cerebellar transcranial stimulation on human physiological tremor, J. Neurosci. 34 (2014) 7501–7508. [4] J. Raethjen, F. Pawlas, M. Lindemann, R. Wenzelburger, G. Deuschl, Determinants of physiologic tremor in a large normal population, Clin. Neurophysiol. 111 (2000) 1825–1837. [5] C.A. Vernooij, R.F. Reynolds, M. Lakie, A dominant role for mechanical resonance in physiological finger tremor revealed by selective minimization of voluntary drive and movement, J. Neurophysiol. 109 (2013) 2317–2326. [6] G. Deuschl, J. Raethjen, M. Lindemann, P. Krack, The pathophysiology of tremor, Muscle Nerves 24 (2001) 716–735. [7] B.A. Conway, D.M. Halliday, S.F. Farmer, U. Shahani, P. Maas, A.I. Weir, J.R. Rosenberg, Synchronization between motor cortex and spinal motoneuronal pool during the performance of a maintained motor task in man, J. Physiol. 489 (1995) 917–924. [8] S. Salenius, K. Portin, M. Kajola, R. Salmelin, R. Hari, Cortical control of human motoneuron firing during isometric contraction, J. Neurophysiol. 77 (1997) 3401–3405. [9] J. Gross, P.A. Tass, S. Salenius, R. Hari, H.-J. Freund, A. Schnitzler, Cortico-muscular synchronization during isometric muscle contraction in humans as revealed by magnetoencephalography, J. Physiol. 527 (2000) 623–631. [10] R. Hari, S. Salenius, Rhytmical corticomotor communication, NeuroReport 10 (1999) R1–R10. [11] C.L. Witham, C.N. Riddle, M.R. Baker, S.N. Baker, Contributions of descending and ascending pathways to corticomuscular coherence in humans, J. Physiol. 589 (2011) 3789–3800. [12] M. Lim, J.S. Kim, M. Kim, C.K. Chung, Ascending beta oscillation from finger muscle to sensorimotor cortex contributes to enhanced steady-state isometric contraction in humans, Clin. Neurophysiol. 125 (2014) 2036–2045. [13] S. Salenius, S. Avikainen, S. Kaakkola, R. Hari, P. Brown, Defective cortical drive to muscle in Parkinson’s disease and its improvement with levodopa, Brain 125 (2002) 491–500. [14] P. Brown, S. Salenius, J.C. Rothwell, R. Hari, Cortical correlate of the piper rhythm in humans, J. Neurophysiol. 80 (1998) 2911–2917. [15] M. Bourguignon, X. De Tiege, M.O. de Beeck, B. Pirotte, P. Van Bogaert, S. Goldman, R. Hari, V. Jousmaki, Functional motor-cortex mapping using corticokinematic coherence, NeuroImage 55 (2011) 1475–1479. [16] H. Piitulainen, M. Bourguignon, X. De Tiège, R. Hari, V. Jousmäki, Corticokinematic coherence during active and passive finger movements, Neuroscience 238 (2013) 361–370. [17] R.J. Elble, Characteristics of physiologic tremor in young and elderly adults, Clin. Neurophysiol. 114 (2003) 624–635. [18] S. Taulu, J. Simola, Spatiotemporal signal space separation method for rejecting nearby interference in MEG measurements, Phys. Med. Biol. 51 (2006) 1759–1768. [19] M. Medvedovsky, S. Taulu, R. Bikmullina, A. Ahonen, R. Paetau, Fine tuning the correlation limit of spatio-temporal signal space separation for magnetoencephalography, J. Neurosci. Methods 177 (2009) 203–211. [20] Freesurfer image analysis suite: http://surfer.nmr.mgh.harvard.edu/ 2014(2013). [21] M. Pohja, S. Salenius, R. Hari, Reproducibility of cortex-muscle coherence, NeuroImage 26 (2005) 764–770. [22] B. Pollok, H. Makhloufi, M. Butz, J. Gross, L. Timmermann, L. Wojtecki, A. Schnitzler, Levodopa affects functional brain networks in Parkinsonian resting tremor, Mov. Disorders 24 (2009) 91–98. [23] A. Schnitzler, C. Münks, M. Butz, L. Timmermann, J. Gross, Synchronized brain network associated with essential tremor as revealed by magnetoencephalography, Mov. Disorders 24 (2009) 1629–1635. [24] J.N. Caviness, H.A. Shill, M.N. Sabbagh, V.G.H. Evidente, J.L. Hernandez, C.H. Adler, Corticomuscular coherence is increased in the small postural tremor of Parkinson’s disease, Mov. Disorder 21 (2006) 492–499.
319
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
315 316 317 318
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
ARTICLE IN PRESS
NSL 31395 1–5
Neuroscience Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Research article
1
Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction
2
3
4
Q1
5
Katja Airaksinen a,b,∗ , Tuuli Lehti a , Jussi Nurminen a , Jarkko Luoma a , Liisa Helle c , Samu Taulu c,d,e , Eero Pekkonen f , Jyrki P. Mäkelä a a
BioMag Laboratory, HUS Medical Imaging Center, University of Helsinki and Helsinki University Hospital, Finland Clinical Neurosciences, Neurology, University of Helsinki and Helsinki Univer-sity Hospital, Finland c Elekta Oy, Helsinki, Finland d Institute for Learning and Brain Sciences, University of Washington, Seattle, USA e Department of Physics, University of Washington, Seattle, USA f Department of Neurology, University of Helsinki and Helsinki University Hospital, Finland
6
b
7 8 9 10 11 12
h i g h l i g h t s
13 14
• Corticokinematic coherence (CKC) can be measured during isometric hand activity. • Cortico-muscular coherence (CMC) and CKC of the postural tremor resemble each other. • The sources of CMC and CKC of postural tremor colocalize in the cortex.
15 16 17 18
a r t i c l e
19 32
i n f o
a b s t r a c t
20
Article history: Received 17 February 2015 Received in revised form 7 May 2015 Accepted 17 June 2015 Available online xxx
21 22 23 24 25 26
31
Keywords: Cortico-muscular coherence Corticokinematic coherence Magnetoencepha-lography Postural tremor
33
1. Introduction
27 28 29 30
34Q3 35 36 37 38 39 40 41
Q2
Corticokinematic coherence (CKC), i.e., coherence calculated between MEG and an accelerometer signal recording movement kinematics can be used for functional mapping of the sensorimotor cortex. Cortical sources of CKC, induced by both voluntary and passive movements, localize at the proximity of sensorimotor cortex. We tested the CKC during a static contraction to compare it with simultaneously measured cortico-muscular coherence (CMC) estimated between MEG and surface EMG to study the role of postural tremor in CMC in ten healthy volunteers. CKC was detectable also during this static task. CKC and CMC spectra had similar power distributions, and sources of CMC and CKC colocalized at the cortex in close proximity of the central sulcus. During the static hold task, the accelerometer signal originates from the postural tremor. The similarities between CMC and CKC indicate that postural tremor is related to CMC in healthy subjects. © 2015 Published by Elsevier Ireland Ltd.
Physiological tremor is considered a general property of the neuromuscular system. It occurs during normal conditions, and is increased e.g., by stress [1]. Physiological tremor originates from complex interactions between neural events and limb-related mechanical properties [2]. However, it is unclear whether the neural or the mechanical properties are most prominent factors in the tremor. Synchronous oscillations in the central nervous system may play an important role in physiological tremor (e.g., [3]) or con-
∗ Corresponding author at: BioMag Laboratory, P.O. Box 340, FIN-00029 HUS, Finland. Fax +358 9 47175781. E-mail address: katja.airaksinen@helsinki.fi (K. Airaksinen).
tribute to dominant peripheral resonance caused by the mechanical properties of the oscillating limb [4]. Nevertheless, synchronous central input may not be necessary for finger physiological tremor [5]. Physiological tremor can be enhanced by posture; this postural tremor appears during static maintenance of position against gravity [6]. Postural tremor is found in healthy controls, and is induced in the fingers with static tension of the wrist extensor muscles. Postural tremor differs from resting tremor which appears in a relaxed limb and attenuates with limb movement, and from kinetic tremor emerging during voluntary movement, as well as from isometric tremor which appears when the limb is contracted against a constant resistance [1]. Cortico-muscular coherence (CMC) is estimated between simultaneously measured magnetoencephalography (MEG) or electroencephalography (EEG) and electromyography (EMG)
http://dx.doi.org/10.1016/j.neulet.2015.06.034 0304-3940/© 2015 Published by Elsevier Ireland Ltd.
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
G Model
ARTICLE IN PRESS
NSL 31395 1–5
K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
2
Table 1 The mean peak frequencies and maximum amplitudes of CMC and CKC (±standard deviation). Also the pooled CMC and CKC are shown. The mean peak frequency (Hz) ± SD CMC
The mean max amplitude ± SD
CKC
Left hand
Right hand
Left hand
Right hand
21.9 ± 7.9 p = 0.26
24.8 ± 7.4
18.0 ± 5.6 p = 0.049
23.6 ± 6.0
Pooled CMC
Pooled CKC
23.3 ± 7.6 p = 0.013
20.7 ± 6.3
recordings. CMC at the 12–33 Hz frequency band has been found between motor cortex and different muscles during isometric contraction in healthy subjects (e.g., [7–9]). CMC is thought to reflect 59 the interplay between cortex and muscle in motor coordination and 60 to reflect central drive to the spinal motoneuron pool (e.g., [10]). 61 More recent studies have emphasized both the efferent descend62 ing drive from the motor cortex to the muscle as well as ascending 63 afferent drive from the muscles to the cortex in producing the CMC 64 [11,12]. However, more detailed information about its physiologi65 cal and pathophysiological mechanisms is lacking. 66 Conway et al. [7] and Salenius et al. [8] have observed coherence 67Q4 also at frequencies outside the beta band. The EMG spectra showed 68 peaks roughly at 10 Hz (alpha), 20 Hz (beta), and 40–50 Hz (gamma 69 band), and significant coherence was present at all three bands. 70 However, alpha and gamma band coherence peaks were occasional 71 findings, only present in minority of subjects. Alpha band coher72 ence has been associated with Parkinson’s disease (PD) tremor [13], 73 whereas gamma band coherence has been linked with very strong 74 muscle contractions [14]. 75 Corticokinematic coherence (CKC) was recently introduced as 76 an alternative method for the functional mapping of the motor 77 cortex [15]. CKC is estimated between MEG or EEG and a three78 axis accelerometer signal recording the kinematics of movement. 79 When subjects performed self-paced repetitive finger movements, 80 a coherence peak was observed at the movement frequency and 81 its first harmonic. The source of the CKC lies either close to the 82 “hand knob” of the primary motor cortex or slightly posterior to it in 83 the primary sensory cortex. As passive finger movements induce a 84 similar CKC, it may reflect somatosensory, probably proprioceptive 85 input to the sensorimotor cortex [16]. 86 Sensitive accelerometers detect the slight postural tremor 87 observed in healthy individuals. [2,4,17]. If postural hand tremor is 88 involved in modification of CMC, the comparison between CMC and 89 CKC during a hold task could potentially reveal a common cortical 90 component. We measured CMC and CKC during wrist dorsiflexion, 91 generating a sustained contraction of arm muscles in healthy con92 trols. To our knowledge, no previous studies have compared CMC 93 94 and CKC during static muscle contraction. 57 58
95
96 97
2. Methods Ten healthy subjects (6 male and 4 female) aged between 22 and 58 years (mean age 30 years) were instructed to extend their wrist
CMC
CKC
Left hand
Right hand
Left hand
Right hand
0.088 ± 0.05 p = 0.953
0.111 ± 0.06
0.062 ± 0.04 p = 0.14
0.090 ± 0.04
Pooled CMC
Pooled CKC
0.099 ± 0.05 p = 0.059
0.075 ± 0.04
(dorsiflexion) for one minute at a time. The task was repeated five times with a pause of 20 s between each trial. The aim was to produce a sustained isometric contraction by using submaximal force. In a preliminary trial before the actual task the subjects extended their wrist with a maximal force to establish a baseline for the contraction force. The subjects performed the task first with the right and then with the left hand. To obtain an EMG of the muscles activated in the motor task, a bipolar surface electrode was attached to the forearm on top of the extensor carpi radialis muscle. To record finger tremor related to muscle tension, an accelerometer (ADXL335 iMEMS Accelerometer, Analog Devices Inc., Norwood, MA, USA) was attached on the nail of the index finger in both hands. The accelerometers measured finger acceleration in three orthogonal directions. The subjects were seated under a 306-channel neuromagnetometer (Elekta Neuromag® , Elekta Oy, Helsinki, Finland) to measure the brain activity during the task. The head position with respect to the MEG helmet was recorded continuously using the vendor-supplied head position indicator system. The recording passband was 0.03–330 Hz and the sampling rate 1012 Hz for MEG, EMG and accelerometers. The EMG signal was not rectified. The spatiotemporal signal space separation (tSSS) [18] method implemented in MaxFilterTM software (Elekta Oy) with an 8-s time window and a subspace correlation limit of 0.9 [19] was applied to the MEG data to suppress the external and close-by artifacts before data analysis. Based on the singular value decomposition of the three orthogonal accelerometer channels, we extracted the signal corresponding to most significant singular value and used the resulting signal in subsequent analysis. MEG and anatomical MRI were co-registered using the MRILab software (Elekta Oy). The matching provided a coordinate transformation between the head coordinates and the MRI coordinates. MR images were segmented to distinguish the brain from the surrounding tissues to provide volume for possible sources using the FreeSurfer software (MGH, Boston, MA) [20] or Seglab software (Elekta Oy). Power spectral density (PSD) of MEG, accelerometer and EMG signals were estimated for each subject using Welch’s method with 50% overlapping 1024-point Hanning windows, resulting in a frequency resolution of approximately 1 Hz. The sensor-level coherence was estimated between the EMG or accelerometer signal and each MEG gradiometer pair in turn. Also the coherence between the EMG and the accelerometer sig-
Table 2 The mean coordinates of x, y and z for CMC and CKC in the upper part of the table. In the lower part are the mean differences between CMC and CKC coordinates and the confidence intervals of these differences. The mean x-coordinate (mm) CMC
The mean y-coordinate (mm) CKC
35.5 ± 9.0 37.2 ± 8.2 p = 0.345 The mean difference −1.65 ± 5.2 95% confidence interval of the difference Upper Lower 0.8 −4.1
The mean z-coordinate (mm)
CMC
CKC
CMC
CKC
11.9 ± 8.7 p = 0.807
12.0 ± 8.1
98.7 ± 6.0 p = 0.701
98.4 ± 5.9
−0.05 ± 4.4 Lower −2.1
0.24 ± 3.3 Upper 2.0
Lower −1.3
Upper 1.8
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
G Model NSL 31395 1–5
ARTICLE IN PRESS K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
3
Q5 Fig.1. Localization of CMC and CKC. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Q6 A: The localization of CMC (left) and CKC (right) in same subject as in this Figure overlaid on the MR image of the subject. B: The mean (±1 standard deviation) values of x and y (left) and x and z (right) coordinates of all subjects. CMC in red and CKC in blue.
141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
nals were determined. The 50% overlapping 1024-point Hanning windows were used. The number of averaged segments used for coherence calculations varied (range 336–618; mean 556.3 ± 53.7). The time segments were the same for both EMG and accelerometer data. The location of the CMC/CKC maximum peak was then determined by visual inspection from a selection of 15 gradiometer pairs over the sensorimotor cortex contralateral to the activated hand. The coherence spectra were then inspected. Peak amplitudes and frequencies were determined from the planar gradiometer pair showing the largest response over the sensorimotor cortex. Timeshifted coherence (coherence calculated by shifting EMG data 3 s in relation to the MEG data) was used for defining the statistical significance level between 5 and 45 Hz. This time shift destroys any true MEG-EMG coherence in the data. The significance level was set at 99% of the time-shifted coherence [13]. Source-space coherence was estimated using the Dynamic Imaging of Coherence Sources (DICS) method in Beamformer software (Elekta Oy). We selected 6-Hz frequency bands based on the sensor-level CMC and CKC coherence peaks, for coherence estimation and evenly distributed grid of locations covering the brain volume with a resolution of 3 mm for source space. For each location, only the direction with maximum current power was used. After the beamformer scanning, we estimated a virtual electrode signal, i.e. a time domain estimate of source activity, from the location of maximum coherence; the same beamformer weights were used to extract the virtual electrode signal and only results within the previously selected 6-Hz frequency band were used in analysis. We then computed the final CMC and CKC results using these virtual electrode signals.
Statistical calculations were done with SPSS (SPSS for Windows version 13.0, SPSS Inc., Chicago, Illinois, United States). Correlations were calculated with Spearman’s rho. The Wilcoxon Signed Ranks Test was used for comparison of two dependent groups. The significance and confidence intervals for coordinate differences were calculated with One-Sample T-Test. The level of significance was set at 0.05. The results are reported as mean ± standard deviation when appropriate. For statistical calculation the absolute values of the x-coordinates were used. 3. Results Significant CMC was found in 19 out of 20 hemispheres; CKC was seen in every hemisphere. The 19 hemispheres presenting both CMC and CKC were used for analyses. The frequency bands of the CMC and CKC peaks overlapped at least in part in 15 out of 19 hemispheres (Supplementary Fig. S1). In those cases the frequency distributions of the CMC and CKC resembled each other strongly. One subject had most prominent CMC and CKC peaks at 10 Hz (Supplementary Fig. S1). As the coherence values between the hemispheres did not differ, we pooled the CMC and CKC of both hemispheres together. Their mean peak frequency was 23 Hz for CMC and 21 Hz for CKC (p = 0.013) (see Table 1). The peak frequencies of CMC and CKC correlated significantly (rs = 0.681, N = 19, p = 0.001, two tailed). Supplementry material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.neulet.2015.06.034 The pooled mean maximum CMC amplitude was 0.1 ± 0.05 and mean maximum CKC amplitude 0.08 ± 0.04; the difference was non-significant (p = 0.059) (see Table 1). Also maximum CMC
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
171 172 173 174 175 176 177 178 179
180
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
G Model
ARTICLE IN PRESS
NSL 31395 1–5
K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
4
Coherence
A
B
0.1
CMC CMC, sign. CKC CKC, sign.
0.05
0 1
Coherence
C
D
peripheral coherence peripheral coherence, sign.
0.8
0.6
0.4
0.2
0
5
10
15
20
25
30
35 Hz
5
10
15
20
25
30
35
Hz
Fig. 2. A and B: Cortico-muscular (CMC, red line; significance level thin horizontal red line) and corticokinematic (CKC, blue dashed line; significance level horizontal blue dotted line) coherences from both hemispheres of one subject. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) C and D: Peripheral coherence between EMG and accelerometer (black line, horizontal dotted black line for significance level). A and C left hand activation, B and D right hand activation. CKC and CMC in the 15–25 Hz range are notably similar. The main peak of peripheral coherence occurs around 10 Hz. Note different y-axis scales for central and peripheral coherence graphs. The time-shifted coherence between 5 and 45 Hz is used for the significance levels.
199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
amplitude and maximum CKC amplitude correlated significantly (rs = 0.575, N = 19, p = 0.010, two tailed). The coordinates of the source locations of CMC and CKC did not differ significantly (see Table 2, Fig. 1) within different hemispheres. The difference between CMC and CKC x-, y, and z-coordinates was calculated. One-Sample T-Test did not reveal significant difference between the remainders of x-, y, and z-coordinates. The 95% confidence intervals of the difference are shown in Table 2. The tremor frequency (the peak frequency of accelerometer power spectrum between 6 and 14 Hz) did not correlate with CMC (rs = −0.178, N = 19, p = 0.465, two tailed) or CKC (rs = 0.045, N = 19, p = 0.854, two tailed) peak frequencies. The ratio between EMG variation during task and maximum wrist extension did not correlate with the CMC (rs = 0.004, N = 19, p = 0.986, two tailed) or CKC (rs = 0.120, N = 19, p = 0.624, two tailed) frequency. The coherence in the 5–11 Hz band was inspected on sensor signals. In channels having the strongest coherence, significant CMC was found in 12 hemispheres and CKC in 11 hemispheres in the 5–11 Hz band. Seven subjects had significant peaks within 1 Hz both in CMC and CKC in the 5–11 Hz band. When the CMC and CKC were compared between individuals, the coherence spectra had considerable variability. This variability was less prominent when the right and left hemisphere of a given subject were compared (Supplementary Fig. 1). To evaluate the joint properties of accelerometer and EMG signals, we estimated the coherence between EMG and accelerometer signals. This peripheral coherence peaked between 7 and 12 Hz in all but one hand. In one hand the maximum peak occurred at 22 Hz but a coherence peak at 12 Hz was present as well. Another peak around 20–30 Hz was seen in six subjects. The mean peak
coherence amplitudes were clearly stronger than the CMC or CKC values (Fig. 2). The peripheral coherence in right and left hand did not differ; the mean peak frequencies were 10.3 ± 4.0 Hz for the left and 9.4 ± 1.2 Hz for the right hand (p = 0.22), the mean peak amplitudes were 0.6 ± 0.2 and 0.7 ± 0.1 respectively (p = 0.21). The peak frequency or peak amplitude of the peripheral coherence did not correlate with the corresponding values of CMC or CKC. The peripheral coherence frequency correlated with the postural tremor frequency estimated from the accelerometer (rs = 0.489, N = 19, p = 0.034, two tailed). 4. Discussion In addition to detecting voluntary finger movements [15,16], the accelerometer was sensitive enough to detect the small postural tremor of our healthy subjects (Supplementary Figs. S2 and S3). The MEG/EEG signals are known to be coherent with the motor unit spike trains or the EMG signal during static contractions. Assuming a linear model, the acceleration signals are a low-pass filtered version of the summation of all motor unit spike trains. Therefore, it is not surprising that a significant coherence between the brain and accelerometer signals may be observed. The accelerometer is easier to use than EMG and its signal is not as sensitive to recording position than EMG which also has limitations, such as band-pass behavior and spectral properties that depend on the type of recording. Thus CKC may provide an attractive alternative for studying motor control. Repeatability of CMC between sessions is only moderate [21]; further study will reveal, for example, if intersession repeatability of CKC is better. Supplementry material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.neulet.2015.06.034
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
230 231 232 233 234 235 236 237 238 239
240
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
G Model NSL 31395 1–5
ARTICLE IN PRESS K. Airaksinen et al. / Neuroscience Letters xxx (2015) xxx–xxx
300
The peak frequencies and maximum amplitudes of CMC and CKC correlated significantly, and their maximum amplitudes did not differ. The beamformer source localization results indicate that CMC and CKC maxima colocalize at the cortex in close proximity of the central sulcus. As CMC and CKC resemble each other closely, and the kinematics during the static hold task are generated by postural tremor, we suggest that CMC is related to postural tremor. The CMC and CKC spectra were similar but not identical; this is not surprising as the postural tremor probably relates to activity in several muscles whereas we recorded EMG only from the ECR muscle. Similar to previous studies, our results show a large individual variation in CMC frequency and magnitude. The distributions of CMC and CKC spectra, however, were remarkably similar within subjects (Fig. 1; Supplementary Fig. S1). This suggests that although the drive between the cortex and the muscles is relatively individually organized, the selected strategy is the same for both hemispheres. The accelerometer power spectra revealed a low-frequency peak in nine out of ten subjects between 5 and 11 Hz (Supplementary Fig. S3) indicating the presence of small-amplitude postural tremor, in line with previous reports [2,4,17]. Although the peak coherence amplitude, the main target of our study, occurred at 18–22 Hz for both CMC and CKC, coherence exceeding the noise level in 5–11 Hz band was also present in 11–12 hemispheres in sensor level coherence. It may be assumed that the peripheral and cortical activities (CMC and CKC) should have coherence at the frequency most prominent in the peripheral signals. However, Parkinsonian [22] and essential [23] tremors as well as Parkinsonian tremor imitated by healthy subjects [22] induce CMC at the tremor frequency and its first harmonic. The same holds true for CKC induced by voluntary movements [16]. Therefore it is not surprising that the coherence between postural tremor and cortical signals occurs both at 5–11 Hz and at 10–22 Hz. Peripheral mechanical-reflex component of postural tremor, not related to cortical activation, also contributes strongly to the accelerometer signals at 5–8 Hz (e.g., [24]). The CMC has been extensively used in studies of Parkinson’s disease. The first suggestion of a link between postural tremor and CMC at 12–18 Hz was made on the basis of an increased CMC in a group of PD patients displaying postural tremor [24]. The relationship with CMC and postural tremor should be kept in mind when making inferences of CMC modifications by various therapies of PD.
301
Financial disclosures
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
302 303 304 305 306 307 308 309
310
311 312
Katja Airaksinen received research funding from Finnish Brain Foundation and the Finnish Parkinson Foundation. Other authors reported no biomedical financial interests or potential conflicts of interests. This work was supported by the Academy of Finland (grant n.o. 122,725) and by the SalWe Research Program for Mind and Body (Tekes – the Finnish Funding Agency for Technology and Innovation grant 1104/10). Acknowledgement We want to thank Dr. Veikko Jousmäki for providing the accelerometer device.
5
References
313
[1] H.W. Berendse, T. van Laar, Tremor, in: E. Wolters Ch, T. van Laar, H.W. Berendse (Eds.), Parkinsonism and Related Disorders, VU University Press, Amsterdam, 2007, pp. 309–325. [2] S. Morrison, K.M. Newell, Postural and resting tremor in the upper limb, Clin. Neurophysiol. 111 (2000) 651–663.
314
[3] A.R. Mehta, J.-S. Brittain, P. Brown, The selective influence of rhythmic cortical versus cerebellar transcranial stimulation on human physiological tremor, J. Neurosci. 34 (2014) 7501–7508. [4] J. Raethjen, F. Pawlas, M. Lindemann, R. Wenzelburger, G. Deuschl, Determinants of physiologic tremor in a large normal population, Clin. Neurophysiol. 111 (2000) 1825–1837. [5] C.A. Vernooij, R.F. Reynolds, M. Lakie, A dominant role for mechanical resonance in physiological finger tremor revealed by selective minimization of voluntary drive and movement, J. Neurophysiol. 109 (2013) 2317–2326. [6] G. Deuschl, J. Raethjen, M. Lindemann, P. Krack, The pathophysiology of tremor, Muscle Nerves 24 (2001) 716–735. [7] B.A. Conway, D.M. Halliday, S.F. Farmer, U. Shahani, P. Maas, A.I. Weir, J.R. Rosenberg, Synchronization between motor cortex and spinal motoneuronal pool during the performance of a maintained motor task in man, J. Physiol. 489 (1995) 917–924. [8] S. Salenius, K. Portin, M. Kajola, R. Salmelin, R. Hari, Cortical control of human motoneuron firing during isometric contraction, J. Neurophysiol. 77 (1997) 3401–3405. [9] J. Gross, P.A. Tass, S. Salenius, R. Hari, H.-J. Freund, A. Schnitzler, Cortico-muscular synchronization during isometric muscle contraction in humans as revealed by magnetoencephalography, J. Physiol. 527 (2000) 623–631. [10] R. Hari, S. Salenius, Rhytmical corticomotor communication, NeuroReport 10 (1999) R1–R10. [11] C.L. Witham, C.N. Riddle, M.R. Baker, S.N. Baker, Contributions of descending and ascending pathways to corticomuscular coherence in humans, J. Physiol. 589 (2011) 3789–3800. [12] M. Lim, J.S. Kim, M. Kim, C.K. Chung, Ascending beta oscillation from finger muscle to sensorimotor cortex contributes to enhanced steady-state isometric contraction in humans, Clin. Neurophysiol. 125 (2014) 2036–2045. [13] S. Salenius, S. Avikainen, S. Kaakkola, R. Hari, P. Brown, Defective cortical drive to muscle in Parkinson’s disease and its improvement with levodopa, Brain 125 (2002) 491–500. [14] P. Brown, S. Salenius, J.C. Rothwell, R. Hari, Cortical correlate of the piper rhythm in humans, J. Neurophysiol. 80 (1998) 2911–2917. [15] M. Bourguignon, X. De Tiege, M.O. de Beeck, B. Pirotte, P. Van Bogaert, S. Goldman, R. Hari, V. Jousmaki, Functional motor-cortex mapping using corticokinematic coherence, NeuroImage 55 (2011) 1475–1479. [16] H. Piitulainen, M. Bourguignon, X. De Tiège, R. Hari, V. Jousmäki, Corticokinematic coherence during active and passive finger movements, Neuroscience 238 (2013) 361–370. [17] R.J. Elble, Characteristics of physiologic tremor in young and elderly adults, Clin. Neurophysiol. 114 (2003) 624–635. [18] S. Taulu, J. Simola, Spatiotemporal signal space separation method for rejecting nearby interference in MEG measurements, Phys. Med. Biol. 51 (2006) 1759–1768. [19] M. Medvedovsky, S. Taulu, R. Bikmullina, A. Ahonen, R. Paetau, Fine tuning the correlation limit of spatio-temporal signal space separation for magnetoencephalography, J. Neurosci. Methods 177 (2009) 203–211. [20] Freesurfer image analysis suite: http://surfer.nmr.mgh.harvard.edu/ 2014(2013). [21] M. Pohja, S. Salenius, R. Hari, Reproducibility of cortex-muscle coherence, NeuroImage 26 (2005) 764–770. [22] B. Pollok, H. Makhloufi, M. Butz, J. Gross, L. Timmermann, L. Wojtecki, A. Schnitzler, Levodopa affects functional brain networks in Parkinsonian resting tremor, Mov. Disorders 24 (2009) 91–98. [23] A. Schnitzler, C. Münks, M. Butz, L. Timmermann, J. Gross, Synchronized brain network associated with essential tremor as revealed by magnetoencephalography, Mov. Disorders 24 (2009) 1629–1635. [24] J.N. Caviness, H.A. Shill, M.N. Sabbagh, V.G.H. Evidente, J.L. Hernandez, C.H. Adler, Corticomuscular coherence is increased in the small postural tremor of Parkinson’s disease, Mov. Disorder 21 (2006) 492–499.
319
Please cite this article in press as: K. Airaksinen, et al., Cortico-muscular coherence parallels coherence of postural tremor and MEG during static muscle contraction, Neurosci. Lett. (2015), http://dx.doi.org/10.1016/j.neulet.2015.06.034
315 316 317 318
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
