Motor Unit Properties and Underlying Determinants in Pathological

Tremor

J. A. Gallego

, J. L. Dideriksen

, A. Holobar

, J. P. Romero

, J. L. Pons

, E. Rocon

and D. Farina

Bioengineering Group, Spanish National Research Council (CSIC), Arganda del Rey, Spain

Department of Neurorehabilitation Engineering, Bernstein Focus Neurotechnology G

ottingen,

Bernstein Center for Computational Neuroscience, University Medical Center G

ottingen,

Georg-August University, G

ottingen, Germany

Faculty of Electrical Engineering and Computer Science, University of Maribor, Maribor, Slovenia

Department of Neurology, University Hospital ‘12 de Octubre’, Madrid, Spain

Keywords:

Motor Unit, Electromyography, Tremor, Essential Tremor, Coherence, Synchronization.

Abstract:

Pathological tremors are accepted to originate from the projection of supraspinal pathological oscillations to

motor neurons that innervate the affected muscles. These oscillations interact with other neural mechanisms

such as reﬂexes, and, together with the mechanical properties of the limb, determine the characteristics of

the tremor. However, much is yet unknown about the exact mechanisms that mediate the different types of

tremor, and how they interact. Given that the neural drive to muscle encodes information about all pathways

that regulate movement, we are investigating the properties of motor unit activities in tremor patients and

the factors that determine them as a means to further our understanding of the disorder. This paper presents

a simulation study that supports our departure hypothesis (that tremor is a common cortical projection to the

motor neuron pool), and ﬁrst experimental evidences on a patient with essential tremor. The latter illustrate that

the predictions derived from the model provide signiﬁcant support for the analysis of real data, and demonstrate

the potential of the analysis techniques here employed.

1 INTRODUCTION

The term pathological tremor encompasses a series of

disorders that originate disabling involuntary oscilla-

tory activity of a body part (Deuschl et al., 1998).

Such oscillatory movement may be ultimately gen-

erated by different mechanisms, such as pathological

oscillations at cortical and subcortical structures–the

underlying cause in most types of tremor–, or a pe-

ripheral neuropathy–a rare type of tremor– (Deuschl

et al., 2001). In addition, due to the intrinsic proper-

ties of the neuromuscular system, other factors such

as the mechanical properties of the limb (muscles)

and spinal and supraspinal afferent loops are also

thought to participate in tremorogenesis (McAuley

and Marsden, 2000).

A motor unit, understood as a motor neuron and

all the ﬁbers it innervates, is the smallest element that

the nervous system can activate (Heckman and Enoka,

2004). The sum of the action potentials ﬁred by all

active spinal motor neurons, which is referred to as

neural drive to muscle, encodes information about de-

scending and afferent neural commands (Farina et al.,

2010). Thus we believe that the elucidation of the

properties of the neural drive to muscle in tremor pa-

tients may help to further what is known about the

pathophysiology of these disorders.

Few studies to date have investigated the speciﬁc

properties of motor unit spike trains in tremor pa-

tients. A characteristic consistently reported in those

available was the presence of paired or tripled dis-

charges, which occurred rhythmically with the tremor

bursts (Das Gupta, 1963) (Dietz et al., 1974) (Elek

et al., 1991) (Baker et al., 1992) (Christakos et al.,

2009). Further, although many works have put for-

ward that motor unit spike trains in tremor patients

exhibit higher synchronization than in healthy sub-

jects (Dietz et al., 1974) (McAuley and Marsden,

2000) (Christakos et al., 2009) (Elble and Deuschl,

2009), no study has speciﬁcally addressed this obser-

vation. Importantly, all of the previous works were

constrained by the limited number of units concur-

rently identiﬁed, typically less than 5–6 per contrac-

tion (Stashuk et al., 2004), a drawback that arises

133

A. Gallego J., L. Dideriksen J., Holobar A., P. Romero J., L. Pons J., Rocon E. and Farina D..

Motor Unit Properties and Underlying Determinants in Pathological Tremor.

DOI: 10.5220/0004679101330138

In Proceedings of the International Congress on Neurotechnology, Electronics and Informatics (DeNeuro-2013), pages 133-138

ISBN: 978-989-8565-80-8

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

from state of the art technologies for recording in-

tramuscular electromyograms (EMG). To circumvent

this limitation, in our study we employe a novel

technique that decomposes the multichannel surface

EMG into constituent motor unit spike trains. This

technique, called Convolution Kernel Compensation

(CKC) (Holobar and Zazula, 2007) (Holobar et al.,

2009), has been largely employed to identify motor

unit spike trains in healthy subjects (Holobar et al.,

2010), and deals effectively with the speciﬁc ﬁring

properties observed in tremor patients (Holobar et al.,

2012).

This work presents an overview of our current

work on the investigation of tremor properties based

on the analysis of motor unit spike trains. We hypoth-

esize that, in tremor patients, the characteristics of

motor unit ﬁring patterns are determined by the pres-

ence of a strong synaptic input related to (i.e at the fre-

quency of) the tremor, which is commonly projected

from supraspinal centers to the entire motor neuron

pool. Thus, here we review a study with a multiscale

model of pathological tremor in which we tested this

departure hypothesis (Gallego et al., 2011), and com-

pare these results with data obtained from one patient

with essential tremor (ET), the most common type of

tremor, and that originates at the cerebellothalamo-

cortical loops (Benito-Le

on and Louis, 2006).

2 Methods

2.1 Computational Model

We employed a multiscale model of a pair of antag-

onist muscles to simulate tremor (Dideriksen et al.,

2011). This model comprised a realistic representa-

tion of a motor neuron pool (Fuglevand et al., 1993),

which served to continuously estimate motor unit ﬁr-

ings during dynamic contractions. The net synaptic

input to each motor neuron integrated a descending

voluntary drive and a descending tremor component,

both of which were commonly projected to the entire

motor neuron pool, and afferent input from muscle

spindles and Golgi tendon organs. EMG was simu-

lated employing a model of multilayer cylindrical vol-

ume conductor that comprised anisotropic muscle tis-

sue and isotropic bone (Farina and Merletti, 2004).

Joint dynamics and afferent inputs were calculated

with a model that accounted for the viscoelastic prop-

erties of muscles. The muscles simulated were the

ﬁrst dorsal interosseus (FDI) and its antagonist, the

second palmar interosseus. Model parameters were

set to those in in (Dideriksen et al., 2011). Simu-

lations were performed at four different contraction

levels (0, 5, 10 and 20 % of the maximum voluntary

contraction [MVC]), and with three different imposed

tremor frequencies (5, 8 and 11 Hz); see (Gallego

et al., 2011) for details.

2.2 Patient and Protocol

We present data for one female ET patient (79 years

old) with bilateral postural and kinetic tremor of mod-

erate severity, recruited at Hospital Universitario “12

de Octubre,” Madrid, Spain. The patient, who was

on medication (propanolol, 120 mg/day) during the

recordings, did not exhibit head or trunk tremor. The

Ethical Committee at the hospital approved the ex-

perimental protocol, and the patient signed a written

informed consent to participate.

The recordings were carried out while the patient

was sitting on an armchair in a dimly illuminated

room. Tremor was triggered by asking the patient to

outstretch both hands, with the palms down and the

ﬁngers slightly outspread, while the forearm was fully

supported. Data were recorded for 4 min.

Hand tremor at the most affected side (right) was

recorded with a multichannel EMG electrode grid (13

x 5 electrodes, 8 mm inter-electrode distance, LISiN–

OT Bioelettronica, Torino, Italy) placed over the ex-

tensor digitorum communis; a moistened bracelet at-

tached to the wrist served as common reference. Con-

currently, we recorded electroencephalographic activ-

ity (EEG) from 32 positions at the somatosensory-

cortex (AFz, F3, F1, Fz, F2, F4, FC5, FC3, FC1,

FCz, FC2, FC4, FC6, C5, C3, C1, Cz, C2, C4, C6,

CP5, CP3, CP1, CPz, CP2, CP4, CP6, P3, P1, Pz, P2,

and P4 according to the 10-20 system) with passive

Au electrodes; the common potential of the two ear-

lobes was used as reference, and Az as ground. EMG

signals were ampliﬁed (EMGUSB, OT Bioelettron-

ica, Torino, Italy), band-pass ﬁltered (10–750 Hz),

and sampled at 2,048 Hz by a 12-bit A/D converter;

EEG signals were ampliﬁed (gUSBamp, g.Tec gmbh,

Graz, Austria), band-pass (0.1–60 Hz) and notch (50

Hz) ﬁltered, and sampled at 256 Hz by a 16 bit A/D

converter.

2.3 Data Processing and Analysis

The analysis focused on two aspects. First we as-

sessed, with the model, how the tremor drive com-

monly projected from supraspinal centers to the entire

motor neuron pool would be transmitted to the output

of such motor neuron population. We expected that

linear sampling would arise after the spike trains of a

few motor neurons were considered together, as ob-

served for the voluntary drive in healthy subjects (Ne-

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

134

33 11 55 10

0.02

0.04

0.06

0.08

0.1

number of motor unitsnumber of motor units

1.05

0.95

0.90

0.85

A B

coherence

Figure 1: Coherence between the supraspinal tremor oscillations and the CSTs, as function of the number of motor units

considered in the CST. Data in (A) corresponds to the model, in (B) to the patient. The plot represents the mean ± SD for

all possible combinations (in the case of the model considering 1 out of 10 motor units recruited). In (B) the dashed line

represents the signiﬁcance threshold (P < 0.05); in the case of (A) it is now visible (it was 0.145 for P < 0.05).

gro and Farina, 2011). Then, we investigated whether

the patient data replicated the behavior expected from

the model. In a second phase, we compared qual-

itatively the simulated and observed (in the patient)

properties of motor unit spike trains, in terms of ﬁring

statistics and synchronization between the activities

of different motor units.

Linear sampling of the (supraspinal) tremor in-

put by the motor neuron pool was investigated by

calculating the coherence between groups of motor

unit spike trains (referred to as composite spike trains

[CST]) and the tremor drive. We also assessed how

coherence varied as more motor units were included

in the CST in order to verify the hypothesis that, in

patients, the tremor is a common cortical input to

the motor neuron pool; if the projection were com-

mon, the coherence at the tremor frequency should

increase until reaching a plateau (saturation) for a few

motor units (Negro and Farina, 2011), meaning that

the motor neuron pool had linearized the transmis-

sion of the tremor drive. In the case of the model,

the tremor drive was directly available for the com-

putations, while for the patient it was assessed from

the raw EEG. In both cases, coherence was estimated

in 1-s disjoint Hann windows (Halliday et al., 1995);

in the case of the patient, windows contaminated with

artefacts were carefully removed. Signiﬁcance was

estimated following (Rosenberg et al., 1989).

Motor unit behavior was assessed by computing

histograms of their inter-spike intervals (ISI), in or-

der to have a global representation of their statis-

tical properties. Special attention was paid to the

presence of paired or tripled discharges (Das Gupta,

1963) (Dietz et al., 1974) (Elek et al., 1991) (Baker

et al., 1992) (Christakos et al., 2009). Further, we

assessed the synchronization between pairs of motor

units to investigate whether synchronization is greater

in tremor patients than in healthy counterparts. In the

model, we expected that abnormally large synchro-

nization would arise due to the presence of a com-

mon synaptic input related to tremor. In the case of

patients, it would be suggestive of the presence of a

strong common input (Kirkwood and Sears, 1978). A

standard metric based on the computation of cross-

correlograms between pairs of motor unit spike trains,

the Common Input Strength index (CIS) (Nordstrom

et al., 1992) was employed. Signiﬁcant synchroniza-

tion was inferred from the cumulative sum of the

cross-correlogram (Ellaway, 1978). For the calcu-

lation, we followed the methodology in (Dideriksen

et al., 2011) (Gallego et al., 2011).

For the patient, we identiﬁed motor unit spike

trains from the decomposition of the multichannel

EMG using the CKC technique (Holobar et al., 2012),

as mentioned above. Motor unit spike trains were ver-

iﬁed by an experienced operator, and motor neurons

that were not active for a signiﬁcant proportion of the

trial (<65 %) or identiﬁed with great accuracy (height

of spike trains compared to baseline jitter ≥26 dB)

were discarded for the analysis. In the case of the

model, motor unit spike trains were directly available.

Results are reported as mean ± SD.

3 RESULTS

For the patient, the total number of identiﬁed motor

units that satisﬁed the criteria presented above were

5. The number of high quality 1-s EEG windows was

93. These data were employed for the subsequent ex-

perimental analyses, except where mentioned other-

wise.

MotorUnitPropertiesandUnderlyingDeterminantsinPathologicalTremor

135

3.1 Common Supraspinal Input

For all simulated conditions (amount of voluntary

contraction and imposed tremor frequency), the co-

herence at the tremor frequency was largely signiﬁ-

cant even when 1 single motor unit was considered

(grand mean for all conditions 0.953 ± 0.027; the

conﬁdence limit was 0.145 for P < 0.05), as expected

from the implementation of the model (Dideriksen

et al., 2011) (Gallego et al., 2011). Even for a such a

large coherence with CSTs comprising 1 motor unit,

when assessing the relationship between coherence

and number of motor units in the CST (see Fig. 1A),

we observed that there was an exponential trend that

reached a plateau when ∼4 motor neurons were con-

sidered (visual inspection). This suggests that the ex-

istence of a common synaptic input related to tremor

could be inferred, even in the concurrent presence of

a voluntary drive, by observing the trend of the co-

herence peak at the tremor frequency as function of

the number of motor units sampled, as previously ob-

served for healthy subjects (Negro and Farina, 2011).

As to the patient, we obtained signiﬁcant coher-

ence at the tremor frequency (5.750 Hz) between the

contralateral cortical activity (largest at CP3) even

when 1 motor unit was sampled (coherence 0.045;

the conﬁdence limit was 0.032 for P < 0.05). As ex-

pected, the coherence was considerably smaller than

for the model given the presence of interneurons in

the descending pathways, which distorted the trans-

mission process due to their nonlinear transfer func-

tion (Gerstner and Kistler, 2002). Interestingly, the

analysis of corticospinal coherence as function of the

number of motor units also indicated that there was

an exponential trend (see Fig 1B), and thus suggested

that the descending tremor drive was a common input

at the tremor frequency.

3.2 Motor Unit Firing Properties and

Synchronization

In the model, the analysis of motor unit behavior

yielded that motor unit ﬁring characteristics were

largely inﬂuenced by motor neuron size (which de-

termines its recruitment threshold (Fuglevand et al.,

1993)) and by the strength of the voluntary drive and

the frequency of the tremor. As motor neuron size in-

creased, at low contraction levels, motor neurons ex-

hibited less likelihood to ﬁre paired and tripled dis-

charges. On the contrary, at strong voluntary con-

tractions small motor units exhibited a ﬁring pattern

that resembled that observed during voluntary con-

tractions in the absence of tremor, and larger motor

neurons ﬁred paired and tripled discharges depend-

0 5 10 20

-2

voluntary contraction (% MVC)

CIS (pps)

5 Hz

8 Hz

11 Hz

Figure 2: Motor unit synchronization as computed with the

CIS. The plot illustrates the mean ± SD synchronization

for all possible pairs of motor neurons (considering 1 out of

each 10 simulated) for 4 contraction levels and 3 different

tremor frequencies (see the legend).

ing on several factors (such as the strength of the

voluntary drive, tremor frequency, and motor neuron

size). Higher tremor frequencies appeared to facili-

tate the transition from rhythmic motor unit ﬁring (en-

trained with the tremor) to a tonic ﬁring pattern. Ac-

cordingly, the ISI histograms of the simulated motor

unit spike trains followed both unimodal and bimodal

(with one peak reﬂecting the paired and tripled dis-

charges) distributions. Interestingly, motor unit syn-

chronization as estimated with the CIS was remark-

ably larger (grand mean for all conditions 6.745 ±

1.870 pps) than when simulating voluntary contrac-

tions in the absence of tremor (mean CIS for the 4

contraction levels 0.203 ± 0.159 pps). Fig. 2 illus-

trates the inﬂuence of strength of voluntary contrac-

tion and tremor frequency on motor unit synchroniza-

tion as estimated with the CIS.

The patient data followed an unimodal, slightly

skewed distribution with a peak at ∼50 ms (joint data

for the 5 identiﬁed motor units, i.e. cumulative ISI

histogram). Unexpectedly, this value was not related

to tremor frequency. However, the motor units often

ﬁred paired and tripled discharges (ISI 20–80 ms), as

previously reported in a study on ET patients (Elek

et al., 1991). Nine out of 10 pairs of motor unit spike

trains exhibited signiﬁcant correlation, and the mean

CIS for the last 2 min of the trial was 1.61 ± 1.48

pps. This value was remarkably larger than previ-

ously reported for healthy subjects performing vol-

untary contractions (≤0.7 pps (Keen and Fuglevand,

2004)), which indicates that the motor neuron pool

receives a strong common synaptic input.

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

136

4 DISCUSSION

This paper showed, based on a multiscale simulation

model and data on one representative tremor patient,

that the supraspinal oscillations that mediate tremor

are a common cortical projection to the motor neuron

pool, as indirectly suggested in previous works com-

puting EEG-EMG coherence. Furthermore it puts

forward, based on physiologically plausible simula-

tions, that the presence of a common tremor drive to

the motor neuron pool may cause abnormally large

motor unit synchronization, as largely hypothesized

for tremor. Experimental results in one representa-

tive ET patient suggest that this abnormally large syn-

chronization may be systematically found in patients,

which would indirectly prove the existence of a strong

common synaptic input related to tremor.

It must be noticed that the coherence values found

in the model were notably larger than those observed

in the patient. We believe that the reason for this is

(at least) twofold. First, in the model the tremor was a

narrowband colored noise that was directly projected

to all motor neurons in the pool, which implies that it

neglects the spectral distortion that interneurons intro-

duce in the transmission of supraspinal oscillations,

given their nonlinear response (Gerstner and Kistler,

2002) (Negro and Farina, 2011). Second, in the model

the central (supraspinal) tremor was directly avail-

able, while for the case of the patient we had to as-

sess its contribution from the EEG. Indeed, the am-

plitude spectrum of the EEG recorded in this patient

did not shown a peak at the tremor frequency, oppo-

site to the clear contribution observed in the case of

the model. Remarkably, low coherence values are re-

ported in other works studying EEG-EMG coherence

in tremor patients (Raethjen et al., 2007) (Volkmann

et al., 1996) (Timmermann et al., 2003).

As to motor neuron synchronization, for the

model we showed that if a strong common tremor in-

put was projected to the entire motor neuron popula-

tion, the motor unit spike trains exhibited abnormally

synchronization. Interestingly, synchronization was

remarkably large for all tremor frequencies and lev-

els of concurrent voluntary contraction. In the case

of the patient, we observed that motor unit synchro-

nization as computed with the CIS was more than

two times larger than previously reported for healthy

counterparts. Thus, given the well-established causal

relationship between synchronization and common

synaptic inputs (Kirkwood and Sears, 1978) we con-

sider this as a further proof of the existence of a com-

mon cortical input to the motor neuron pool related to

tremor.

ACKNOWLEDGEMENTS

This work has been funded by the EU Commission

through grants EU-FP7-2007-224051 (TREMOR)

and EU-FP7-2011-287739 (NeuroTREMOR).

REFERENCES

Baker, J. R., Davey, N. J., Ellaway, P. H., and Friedland,

C. L. (1992). Short-term synchrony of motor unit dis-

charge during weak isometric contraction in parkin-

son’s disease. Brain, 115 Pt 1:137–154.

Benito-Le

on, J. and Louis, E. D. (2006). Essential tremor:

emerging views of a common disorder. Nature clinical

practice. Neurology, 2(12):666–78;.

Christakos, C. N., Erimaki, S., Anagnostou, E., and Anasta-

sopoulos, D. (2009). Tremor-related motor unit ﬁring

in parkinson’s disease: implications for tremor gene-

sis. J Physiol, 587(Pt 20):4811–4827.

Das Gupta, A. (1963). Paired response of motor units dur-

ing voluntary contraction in parkinsonism. J Neurol

Neurosurg Psychiatry, 26:265–268.

Deuschl, G., Bain, P., and Brin, M. (1998). Consensus state-

ment of the movement disorder society on tremor. ad

hoc scientiﬁc committee. Mov Disord, 13 Suppl 3:2–

23.

Deuschl, G., Raethjen, J., Lindemann, M., and Krack, P.

(2001). The pathophysiology of tremor. Muscle &

Nerve, 24:716–735.

Dideriksen, J., Enoka, R., and Farina, D. (2011). A model

of the surface electromyogram in pathological tremor.

IEEE Trans Biomed Eng.

Dietz, V., Hillesheimer, W., and Freund, H. J. (1974). Corre-

lation between tremor, voluntary contraction, and ﬁr-

ing pattern of motor units in parkinson’s disease. J

Neurol Neurosurg Psychiatry, 37(8):927–937.

Elble, R. J. and Deuschl, G. (2009). An update on essential

tremor. Neurol Neurosci Reports, 9:273–7.

Elek, J. M., Dengler, R., Konstanzer, A., Hesse, S.,

and Wolf, W. (1991). Mechanical implications of

paired motor unit discharges in pathological and vol-

untary tremor. Electroencephalogr Clin Neurophysiol,

81(4):279–283.

Ellaway, P. H. (1978). Cumulative sum technique and

its application to the analysis of peristimulus time

histograms. Electroencephalogr Clin Neurophysiol,

45(2):302–304.

Farina, D., Holobar, A., Merletti, R., and Enoka, R. M.

(2010). Decoding the neural drive to muscles from

the surface electromyogram. Clin Neurophysiol,

121(10):1616–1623.

Farina, D. and Merletti, R. (2004). Estimation of av-

erage muscle ﬁber conduction velocity from two-

dimensional surface emg recordings. J Neurosci

Methods, 134(2):199–208.

Fuglevand, A. J., Winter, D. A., and Patla, A. E. (1993).

Models of recruitment and rate coding organization in

motor-unit pools. J Neurophysiol, 70(6):2470–2488.

MotorUnitPropertiesandUnderlyingDeterminantsinPathologicalTremor

137

Gallego, J. A., Dideriksen, J. L., Farina, D., Rocon, E.,

Holobar, A., and Pons, J. L. (2011). A modelling

study on transmission of the central oscillator in

tremor by a motor neuron pool. In Proc. Annual Int

Engineering in Medicine and Biology Society,EMBC

Conf. of the IEEE, pages 2037–2040.

Gerstner, W. and Kistler, W. M. (2002). Spiking neuron

models. Single neurons, populations, plasticity. Cam-

bridge University Press.

Halliday, D. M., Rosenberg, J. R., Amjad, A. M., Breeze,

P., Conway, B. A., and Farmer, S. F. (1995). A frame-

work for the analysis of mixed time series/point pro-

cess data–theory and application to the study of physi-

ological tremor, single motor unit discharges and elec-

tromyograms. Prog Biophys Mol Biol, 64(2-3):237–

278.

Heckman, C. J. and Enoka, R. M. (2004). Handbook of

Clinical Neurophysiology, vol. 4, Clinical Neurophys-

iology of Motor Neuron Diseases, chapter Physiology

of the motor neuron and the motor unit, pages 119–

147. Elsevier.

Holobar, A., Farina, D., Gazzoni, M., Merletti, R., and Za-

zula, D. (2009). Estimating motor unit discharge pat-

terns from high-density surface electromyogram. Clin

Neurophysiol, 120(3):551–562.

Holobar, A., Glaser, V., Gallego, J. A., Dideriksen, J. L., and

Farina, D. (2012). Non-invasive characterization of

motor unit behaviour in pathological tremor. J Neural

Eng, 9(5):056011.

Holobar, A., Minetto, M. A., Botter, A., Negro, F., and Fa-

rina, D. (2010). Experimental analysis of accuracy in

the identiﬁcation of motor unit spike trains from high-

density surface emg. IEEE Trans Neural Syst Rehabil

Eng, 18(3):221–229.

Holobar, A. and Zazula, D. (2007). Multichannel blind

source separation using convolution kernel compen-

sation. Signal Processing, IEEE Transactions on,

55(9):4487 –4496.

Keen, D. A. and Fuglevand, A. J. (2004). Common input

to motor neurons innervating the same and different

compartments of the human extensor digitorum mus-

cle. J Neurophysiol, 91(1):57–62.

Kirkwood, P. A. and Sears, T. A. (1978). The synaptic con-

nexions to intercostal motoneurones as revealed by

the average common excitation potential. J Physiol,

275:103–134.

McAuley, J. H. and Marsden, C. D. (2000). Physiological

and pathological tremors and rhythmic central motor

control. Brain, 123:1545–1567.

Negro, F. and Farina, D. (2011). Linear transmission of cor-

tical oscillations to the neural drive to muscles is me-

diated by common projections to populations of mo-

toneurons in humans. J Physiol, 589(Pt 3):629–637.

Nordstrom, M. A., Fuglevand, A. J., and Enoka, R. M.

(1992). Estimating the strength of common input to

human motoneurons from the cross-correlogram. J

Physiol, 453:547–574.

Raethjen, J., Govindan, R. B., Kopper, F., Muthuraman,

M., and Deuschl, G. (2007). Cortical involvement

in the generation of essential tremor. J Neurophysiol,

97(5):3219–3228.

Rosenberg, J. R., Amjad, A. M., Breeze, P., Brillinger,

D. R., and Halliday, D. M. (1989). The fourier ap-

proach to the identiﬁcation of functional coupling be-

tween neuronal spike trains. Prog Biophys Mol Biol,

53(1):1–31.

Stashuk, D. W., Farina, D., and Sgaard, K. (2004). Elec-

tromyography: Physiology, Engineering, and Nonin-

vasive Applications, chapter Decomposition of intra-

muscular emg signals. Wiley–IEEE Press.

Timmermann, L., Gross, J., Dirks, M., J., V., H.J., F., and

Schnitzler, A. (2003). The cerebral oscillatory net-

work of parkinsonian resting tremor. Brain, 126:199–

212.

Volkmann, J., Joliot, M., Mogilner, A., Ioannides, A. A.,

Lado, F., Fazzini, E., Ribary, U., and Llins, R. (1996).

Central motor loop oscillations in parkinsonian resting

tremor revealed by magnetoencephalography. Neurol-

ogy, 46(5):1359–1370.

NEUROTECHNIX2013-InternationalCongressonNeurotechnology,ElectronicsandInformatics

138