Exp Brain Res (2008) 191:57–66
DOI 10.1007/s00221-008-1495-5
RESEARCH ARTICLE
Cortical excitability changes following grasping exercise
augmented with electrical stimulation
Gergely I. Barsi Æ Dejan B. Popovic Æ
Ina M. Tarkka Æ Thomas Sinkjær Æ Michael J. Grey
Received: 8 April 2008 / Accepted: 8 July 2008 / Published online: 29 July 2008
Ó Springer-Verlag 2008
Abstract Rehabilitation with augmented electrical stim-
ulation can enhance functional recovery after stroke, and
cortical plasticity may play a role in this process. The
purpose of this study was to compare the effects of three
training paradigms on cortical excitability in healthy sub-
jects. Cortical excitability was evaluated by analysing the
input–output relationship between transcranial magnetic
stimulation intensity and motor evoked potentials (MEPs)
from the flexor muscles of the fingers. The study was per-
formed with 25 healthy volunteers who underwent 20-min
simulated therapy sessions of: (1) functional electrical
stimulation (FES) of the finger flexors and extensors, (2)
voluntary movement (VOL) with sensory stimulation, and
(3) therapeutic FES (TFES) where the electrical stimulation
augmented voluntary activation. TFES training produced a
significant increase in MEP magnitude throughout the
G. I. Barsi (&)
D. B. Popovic
T. Sinkjær
M. J. Grey
Center for Sensory-Motor Interaction (SMI),
Department of Health Science and Technology,
Aalborg University, Fredrik Bajers Vej 7 D-3,
9220 Aalborg, Denmark
e-mail: barsi@hst.aau.dk; barsi@miba.auc.dk
D. B. Popovic
Faculty of Electrical Engineering, University of Belgrade,
Belgrade, Serbia
I. M. Tarkka
Brain Research and Rehabilitation Center Neuron,
Kuopio, Finland
M. J. Grey
Departments of Neuroscience and Pharmacology and Exercise
and Sport Sciences, University of Copenhagen, Blegdamsvej 3b,
The Panum Institute 22.3, 2200 Copenhagen, Denmark
stimulation range, suggesting an increase in cortical excit-
ability. In contrast, neither the FES nor voluntary movement
alone had such an effect. These results suggest that the
combination of voluntary effort and FES has greater
potential to induce plasticity in the motor cortex and that
TFES might be a more effective approach in rehabilitation
after stroke than FES or repetitive voluntary training alone.
Keywords Plasticity
Motor cortical excitability

Human
Flexor Digitorum profundus

Transcranial magnetic stimulation
Introduction
Cortical plasticity is now widely accepted to play a fun-
damental role in motor learning and neurorehabilitation.
Changes in cortical excitability can be used as an indicator
of cortical plasticity and has been shown to be associated
with motor re-learning (e.g. Jenkins and Merzenich 1987;
Castro-Alamancos et al. 1992; Di Piero et al. 1992; Chae
and Yu 1999; Ridding and Rothwell 1999). The excitability
of the cortical projections to a muscle involved in learning
of a new motor task is increased during the learning pro-
cess (Muellbacher et al. 2001). However, increased
excitability can also be caused by decreased inhibitory
activity, as suggested by Jacobs and Donoghue (1991).
Inhibitory circuitry may also have an important role in
plasticity by regulating correlative activity of cortical cells
based on spatial distribution (Ferster 2004).
A growing body of research has demonstrated that
repetitive practice of a motor task combined with positive
feedback leads to reorganisation of the motor cortex (for
recent reviews see Nudo 2003, 2007). Repetition of even a
simple movement can produce changes in the motor
123
58
response evoked by transcranial magnetic stimulation
(TMS), leading to a transient reorganisation in motor
connectivity (Classen et al. 1998). Changes in cortical
excitability are further increased when performing tasks
that require greater skill. Moreover, active involvement in
task performance leads to a substantial increase in cortical
excitability compared to non-skilful or passive training
(Perez et al. 2004).
Electrical stimulation of the periphery has been shown
to lead to cortical reorganisation (Ridding and Rothwell
1999; Rossini and Pauri 2000) and/or change in cortical
excitability (Khaslavskaia et al. 2002; Knash et al. 2003).
Several recent studies have shown that peripheral electrical
stimulation can facilitate motor learning (e.g. McDonnell
and Ridding 2006) and lead to improvements in function
following disability (e.g. Powell et al. 1999).
Functional electrical stimulation (FES) has been used as
a neuroprosthetic for several decades (for review see
Peckham and Knutson 2005). Recently, functional and
clinical improvements have been observed with the thera-
peutic application of FES whereby the stimulation was
used to augment residual voluntary activation after stroke
(Popovic et al. 2003; Gritsenko and Prochazka 2004;
Popovic et al. 2004) and spinal cord injury (Popovic et al.
2006; Thrasher et al. 2006). Therapeutic FES (TFES), the
combination of FES with voluntary training, synonymous
with Functional Electrical Therapy (Popovic et al. 2002),
has produced changes in excitability of area of the motor
cortex responsible for control of tibialis anterior (Kido
Thompson and Stein 2004; Khaslavskaia and Sinkjær
2005), and the effect is greater than that of either electrical
stimulation or voluntary movement when performed sep-
arately. This suggests that TFES might be effective because
it strengthens corticospinal circuitry.
The present study was motivated by the promising
clinical results that have been observed with therapeuti-
cally applied FES. Our aim was to investigate the changes
in cortical excitability produced by 20-min training ses-
sions in healthy volunteers where grasp was generated by
(1) FES alone, (2) volitional muscle activity augmented
with FES (simulated TFES), and (3) volitional (VOL)
muscle activity alone. We hypothesised that TFES training
would increase the excitability of the finger flexor area of
the motor cortex, and that the increase would be greater
than that induced by voluntary training or FES training.
Changes in cortical excitability were determined by mea-
suring the input–output relationship (stimulus-response
curve) between transcranial magnetic stimuli and their
corresponding motor evoked potentials (MEPs) (Devanne
et al. 1997; Ridding and Rothwell 1997). Changes in the
inhibitory circuitry were determined by measuring the
duration of the cortical silent period (SP). The SP refers to
123
Exp Brain Res (2008) 191:57–66
the suppressed electromyography (EMG) activity that fol-
lows a MEP evoked by TMS in actively contracting
muscle. The duration of the SP is used as an estimate of
excitability of cortical inhibitory circuits (Cantello et al.
1992; Chen 2000; Siebner and Rothwell 2003; Seitz et al.
2004).
Methods
Subjects
This study was performed in two experiments with a total
of 25 healthy adults (18 male, 7 female; 21 right-handed, 4
left-handed; aged 21–38 years) with no known history of
neuromuscular disorder. All participants completed a
safety screen questionnaire (modified from Keel JC, July
2000), and provided informed written consent prior to their
participation. All experimental procedures were approved
by the local ethics review board (Nordjyllands Amt, project
number VN-20010029MCH) and the experiments were
conducted in accordance with the Declaration of Helsinki.
Apparatus and instrumentation
The participants were seated comfortably in a chair with
the right forearm positioned mid-way between pronation
and supination and supported by a padded armrest. The
elbow was held at *120°, and the wrist was supported.
The fingers were free to move, allowing grasping and
releasing of a 50-cl water bottle for hand grasp exercises.
This position was maintained throughout the experiment.
Electrical muscle stimulation
The stimulation was applied with a commercial electrical
stimulator (ActigripÒ, Neurodan A/S, Denmark) using
disposable self-adhesive surface electrodes (3.2 cm diam-
eter, round) with the cathodes positioned over the
respective motor points of finger flexor/extensor muscles:
flexor digitorum profundus (FDP) and extensor digitorum
communis (EDC). A common anode (4 cm 9 6.4 cm,
oval) was placed on the lateral surface of the forearm just
proximal to the wrist. The electrode positions were care-
fully selected to ensure that finger extension/flexion
movement was as natural as possible. The stimulation
pattern was designed to mimic activity of fingers typical of
slow grasping and releasing. The pulse duration, frequency,
and amplitude (current) were set to minimise discomfort
during stimulation yet produce a finger flexion/extension
movement that would allow the subject to grasp and hold a
50-cl bottle of water without voluntary assistance.
Exp Brain Res (2008) 191:57–66
Electromyography
Surface EMG was used to record muscle activation.
Monopolar surface electrodes in bipolar montage were
placed over the FDP and EDC muscles of the right forearm.
The placement of the electrodes over each muscle was
following the recommendations of the SENIAM project
(Hermens et al. 2000). The EMG signals were amplified,
band-pass filtered between 1 and 1,000 Hz, sampled at
4 kHz, and stored on a computer for off-line analysis.
Transcranial magnetic stimulation
Transcranial magnetic stimulation was used to assess the
excitability of the cortical projections to the target muscle.
Magnetic stimuli were delivered with a Magstim Rapid2
and Magstim 200 stimulator (The Magstim Company,
Dyfed, UK). The stimuli were delivered with a custom-
made 90 mm double coil (batwing design; type no. 15411,
Magstim company, Dyfed, UK) placed on the scalp over
the motor cortical area of the left hemisphere where acti-
vation of the right FDP muscle was obtained. The coil was
oriented parallel to midline, with the handle pointing
backwards. The optimal coil placement was defined as the
location where the stimulation produced a maximal MEP in
the FDP muscle. When the coil placement was found, the
head was stabilised with a padded support and the coil was
fixed with respect to the head to ensure that the same area
of the cortex was stimulated throughout the study.
The stimulus-response curve assessments were con-
ducted with the finger flexors relaxed (experiment 1,
n = 14 subjects) and while holding a 5% maximum vol-
untary contraction (experiment 2, n = 15 subjects). The
primary reason for experiment 2 (5% finger flexion) was to
obtain SP information for full stimulus-response curve
recordings (see Sect. ’’Analysis’’). In both cases magnetic
stimuli were delivered at pseudo-random stimulus intensi-
ties (15–99% of maximal stimulator output, MSO) and
random time intervals between 1–1.5 s (experiment 1: 160
pulses, 6% increments; Magstim Rapid2) and 3.5–4 s
(experiment 2: 80–100 pulses, 5% increments; Magstim
200). When TMS was applied during muscle contraction,
visual feedback of the subjects’ force was provided to them
with an oscilloscope.
Peripheral nerve stimulation
Peripheral nerve stimulation was used to normalise the
magnetic evoked potentials so that intra-subject compari-
sons could be made between MEPs measured during
different test sessions. A compound action potential in FDP
was elicited with monopolar electrical stimulation of the
59
median nerve. The EMG responses representing the direct
motor stimulation (M-wave) were monitored as the stim-
ulation intensity was increased from a subliminal level
until there was no further increase in the peak-to-peak
amplitude of the M-wave with increasing intensity. The
stimulation amplitude was then increased by approximately
15%. This supra-maximal stimulation level was used to
elicit a maximal M-wave (Mmax) to which the MEPs were
normalised.
Experimental procedure
Three types of training were investigated: (1) FES alone to
produce finger flexion/extension movement, (2) TFES
where electrical stimulation augmented voluntary move-
ment, and (3) voluntary movement (VOL) with low-
intensity electrical stimulation to elicit cutaneous sensa-
tion. Thus the difference between FES and TFES training
was voluntary effort, and the difference between TFES and
VOL training was the intensity of electrical stimulation. To
ensure no carry-over effects occurred from one session to
another the three different training paradigms were applied
on different days with at least 24 h between each session.
All subjects participated each of the three training sessions,
each lasting 20 min and presented in a randomised order.
To assess changes in cortical excitability, transcranial
magnetic stimuli and peripheral electrical stimuli were
applied before and after each training session.
Training paradigms
The FES in this study was designed only for fingers. The
stimulation pattern was designed to replicate typical hand
open, grasp, and release phases: (1) fingers extension, (2)
fingers flexion (open and close of the hand around the
object), and (3) fingers extension automatically followed
by stimulation off (release of the object and relax) (Popovic
et al. 2004). These operations were controlled by the sub-
ject via a push-button with the left hand. The extension
phases lasted for 1.5 s, and the duration of the sustained
flexion (grasp) was controlled by the subject and varied
between 2 and 3 s. On some subjects, the finger flexion-
extension movement was recorded with a goniometer
(S720 Miniature Joint Angle ShapeSensorTM, Measurand
Inc., Canada). The goniometer was placed on the proximal
interphalangeal joint of the index finger. Figure 1 shows an
example goniometer recording of the grasp phase (finger
extension followed by finger flexion) from a single subject
in all three training paradigms. During (1) FES and (2)
TFES the following parameters were used: f = 50 pulse/s,
T = 200 ls, Iflexor = 6–13 mA, and Iextensor = 7–16 mA,
and in (3) VOL Iflexor = 3–5 mA and Iextensor = 4–5 mA.
123
60
Fig. 1 An example of grasp kinematics for a typical subject. The
goniometer was placed on the proximal interphalangeal joint of the
right index finger. The ordinate is adjusted such that 0° represents full
extension and 90° is full flexion. The vertical dotted line indicates
movement initiation. The plots show 20 trials of the grasp phase from
one subject during electrical stimulation (FES), therapeutic functional
electrical stimulation (TFES), and voluntary training with sensory
level stimulation (VOL). Some trial-to-trial variability is evident
within all three training paradigms, yet the movement was always
within similar range
Analysis
Data were excluded from the analysis if they did not meet
strict inclusion criteria. For experiment 1, where the subject
was instructed to relax, trials were rejected from further
analysis if the background EMG activity was [20 lV in a
50-ms window before the stimulation. For experiment 2,
where the subject held a 5% MVC contraction, trials were
rejected from the analysis if the mean background force
before the TMS pulse was 20% above or below the target
123
Exp Brain Res (2008) 191:57–66
level. Furthermore, the full data set was rejected when
more than 50% of total trials failed to meet the inclusion
criteria. With these criteria, 4 out of 42 (nexp1 = 14, 3
sessions) data sets belonging to three subjects in experi-
ment 1, and 9 of 45 (nexp2 = 15, 3 sessions) data sets
belonging to six subjects in experiment 2 were rejected. As
a result, 9 of the 25 subjects recruited for the study were
excluded from analysis. Four subjects participated in both
experiment 1 and 2, resulting in nexp1 = 11 and nexp2 = 9.
Three peripheral nerve stimuli were given at a supra-
maximal stimulation level to determine the maximal M-
wave (Mmax). The M-wave onset was defined as the time
when the rectified EMG signal rose above 3% of the
maximal amplitude and its offset was defined as when it
fell beyond 3% again for at least 3 ms. The Mmax was
quantified as the integral of the rectified EMG response.
Mmax was obtained once for before and once after the
training.
Motor threshold (MT) was defined as the TMS stimu-
lation intensity that evoked five out of ten MEPs above
50 lV (Ridding and Rothwell 1995) for experiment 1 and
200 lV (Rosenkranz and Rothwell 2004) for experiment 2
in the finger flexor muscles. The MEP onset latency was
defined as the time from the TMS stimulus to the first
deflection of MEP in the EMG signal. The SP was defined
as the relative silent period (Stetkarova et al. 1994) cal-
culated as the time elapsed from MEP onset to the return of
uninterrupted background EMG activity, defined as the
time where the rectified EMG activity rose above the mean
of the rectified pre-stimulus EMG activity. The MEP onset
latency and SP duration were determined by visual
inspection with the examiner blinded to the subject and
exercise condition.
Motor evoked potentials were quantified as the integral
of the rectified EMG response in a 15–60 ms window
following the magnetic stimulation and were normalised to
the Mmax. MEPs were plotted against the varied stimulus
intensities. The stimulus-response curves were constructed
by modelling the relation between stimulus intensities and
MEPs with a Boltzmann-like sigmoid equation:
MEPmax  MEPmin
MEPðIÞ ¼ MEPmin þ I50 I
1þe S
The equation has four parameters: the maximum plateau
value (MEPmax), the base offset (MEPmin), the stimulus
intensity at the inflection point (I50), and the slope (S) at the
inflection point. The inverse of the slope parameter (1/S) is
directly proportional to the maximum steepness of the
function. These parameters can be interpreted as estimates
of background EMG activity (MEPmin), maximum elicited
response (MEPmax), and the transition between them (S,
I50) in relation to stimulus intensity (I) (Devanne et al.
1997; Knash et al. 2003). The parameters were estimated
Exp Brain Res (2008) 191:57–66
by fitting this equation to the MEP data with a Marquardt–
Levenberg non-linear least squares algorithm (Press et al.
1986).
Statistical analysis
The MT and the range of magnetic stimulation intensity
necessary to produce a full stimulus-response curve from
no response to maximum response were variable between
subjects. To overcome this issue the stimulation intensities
where normalised for each subject by setting the intensity
at 80% of the MT (expressed in %MSO) to 0% normalised
stimulus intensity (NSI) and then setting the stimulation
intensity where the stimulus-response curve reached the
upper plateau (expressed in %MSO) to 100% NSI.
The upper plateau was reached at the lowest intensity
where the difference between the stimulus-response curve
amplitude and MEPmax was not [0.5% of the maximum
height of the curve:
ðMEPmax  MEPmin Þ  0:005  MEPmax  SRðI Þ:
The MEPs were plotted against NSI. All statistical tests
involving MEPs were conducted on this data set.
MEP onset latency and SP duration data were normally
distributed. MEP and Mmax data were lognormally dis-
tributed and significance is reported based on log-
transformed values. All statistical tests were considered
significant at an alpha level of 0.05. Results are reported as
mean ± 1 standard deviation.
In order to determine if there was a training effect on the
excitability of the cortical projections to the target muscle
(MEP data), the cortical inhibitory circuits (MEP onset
latency and SP data), and the periphery (Mmax data) one-way
repeated measures analysis of variance (rmANOVA; time)
were used for each training paradigm (FES, TFES, VOL). A
two-way rmANOVA (session 9 time) with Scheffé’s mul-
tiple comparison was used to test for differences between the
effects of the three training paradigms. MEP data were
tested separately for experiment 1 and 2. MEP onset latency
and SP data were tested only for experiment 2. Mmax data
were pooled together for experiment 1 and 2.
Results
Figure 1 illustrates the hand grasp kinematics for a repre-
sentative subject undergoing the FES, TFES, and VOL
training sessions. While there is some trial-to-trial vari-
ability within all three training paradigms, the finger
flexion/extension excursion was always within similar
range. Figure 2 shows an example of unprocessed flexor
EMG record from one subject at three different stimulation
intensities before and after a single training session in
61
experiment 2, in this case TFES. The MEP magnitude
increased both with stimulation intensity and as a result of
the training. The MEP onset latency decreased and the SP
increased with stimulation intensity only. Figure 3 illus-
trates how the stimulus-response curves were constructed
from the recorded MEPs before and after training from the
same subject. Across all subjects the mean goodness of fit
of the Boltzmann-sigmoid curves to the MEP data was
r2 = 0.89 ± 0.08.
Excitability of the cortical projections to the flexor
muscles
In experiment 1 across all subjects (n = 11) the TFES
training produced a mean increase of 15% in the MEP area
across all stimulation intensities (61.62 ± 49.96%MEPmax
to 77.23 ± 70.41%MEPmax; P = 0.002) in the flexor
muscles (Fig. 4a). FES training showed a similar trend with
an increase in MEP magnitude of 8% (69.84 ±
54.08%MEPmax to 78.47 ± 64.23%MEPmax; P = 0.079),
but it did not reach significance. In contrast, VOL training
decreased the MEP magnitude by 7% (64.89 ± 55.43%
MEPmax to 57.47 ± 51.00%MEPmax; P = 0.007). The
two-way rmANOVA revealed a significant session effect
(P \ 0.001) and interaction effect (P \ 0.001), with no
significant effect of time (P = 0.291). Scheffé’s multiple
comparison test indicated that MEP magnitudes were sig-
nificantly bigger for both TFES and FES training compared
with MEP magnitudes with VOL training (P \ 0.05).
In experiment 2 (Fig. 4b) across all subjects (n = 9)
TFES training produced a 19% increase in the magnitude
of the FDP MEP (63.50 ± 37.14%MEPmax to 82.92 ±
57.25%MEPmax; P \ 0.001). Neither FES training
(56.81 ± 36.62%MEPmax to 60.76 ± 45.10%MEPmax;
P = 0.628) nor VOL training (64.06 ± 37.59%MEPmax to
65.67 ± 49.52%MEPmax; P = 0.115) produced significant
changes in the FDP MEP magnitude. Similar to experiment
1, the two-way rmANOVA revealed significant session
(P = 0.009) and interaction (P \ 0.001) effects with no
significant time effect (P = 0.097). The Scheffé’s multiple
comparison test indicated that MEP magnitudes for TFES
training were significantly greater than for FES training
(P \ 0.05).
MEP onset latency and silent period
In experiment 2 all subjects (n = 9) were asked to keep their
finger flexor muscles contracted at a constant level of 5%
MVC during TMS data acquisition. MEP onset latencies and
SP were analysed to investigate modulation of cortical
inhibitory circuits due to training. Both the MEP onset
latency and SP showed linear relationship with TMS stim-
ulation intensity (MEP onset latency: r2 = 0.99 ± 0.003,
123
62 Exp Brain Res (2008) 191:57–66

Fig. 2 Raw EMG recording

example of TFES training in
one subject from experiment 2.
On the left, one representative
EMG trace from after training
(black line) is superimposed on
one representative EMG trace
before training (grey line) at the
intensities of 20, 40, and 60% of
maximal stimulator output. The
TMS stimulus is given at time
0 s (marked by the dashed line).
The end of the cortical silent
period is marked by dash-dot
lines (single dot—before
training, double dot—after
training). In this recording, the
silent period decreased slightly
after training, but the decrease
was not statistically significant
(P = 0.27). On the right, the
plots show the same curves as
on the left, but the time axis is
zoomed in from 15 to 60 ms
after stimulus (marked by the
black rectangle on the left side).
The onset of the MEPs (marked
by dotted lines) decreases by
increasing stimulation intensity

SP: r2 = 0.97 ± 0.021). The MEP onset latency decreased
and the SP increased with increasing stimulation intensities.
Across all subjects, the FES and TFES training caused the
mean MEP onset latency to increase by 0.29 ms
(15.55 ± 1.69 ms to 15.84 ± 2.64 ms; P = 0.004) and
0.33 ms (15.8 ± 1.35 ms to 16.14 ± 1.51 ms; P \ 0.001)
on average (Fig. 5a). VOL had no effect on the mean MEP
onset latency (16.03 ± 1.42 ms to 16.03 ± 1.44 ms;
P = 0.424). The two-way rmANOVA revealed a significant
session effect (P \ 0.001), time effect (P \ 0.001) and
interaction effect (P = 0.012). The Scheffé’s multiple
comparison test indicated that MEP latencies were signifi-
cantly longer for FES training compared with MEP latencies
with TFES and VOL training (P \ 0.05). The average SP in
FES (134.33 ± 48.91 ms to 138.01 ± 45.77 ms; P =
0.113), TFES (139.37 ± 45.57 ms to 137.04 ± 48.65 ms;
P = 0.578), and VOL (139.24 ± 42.18 ms to 135.84 ±
45.63 ms; P = 0.151) were all similar (Fig. 5b). The two-
way rmANOVA revealed no significant session effect
(P = 0.477), time effect (P = 0.893), or interaction effect
(P = 0.064). The SP was not affected by any of the three
training paradigms.
Peripheral M-wave
Peripheral electrical stimulation evoked M-wave areas
were decreased after FES and TFES by 15% (17.08 ±
7.63 mV
ms to 14.81 ± 6.89 mV
ms; P = 0.027) and
17% (17.08 ± 7.65 mV
ms to 14.57 ± 7.13 mV
ms;
P = 0.003) respectively. VOL also showed similar
trend (7% decrease; 16.49 ± 6.71 mV
ms to 15.37 ±
5.8 mV
ms; P = 0.252), but it was not statistically

123
Exp Brain Res (2008) 191:57–66 63

Discussion

In this study we investigated if a single session of FES

training, TFES training, or voluntary grasp training could
produce an increase in the excitability of the cortical pro-
jections to the finger flexor muscles. TFES training did
indeed produce an increase in excitability in relaxed and
contracted finger flexors. In contrast, FES and VOL train-
ing had no effect when measured during contraction, and
VOL training actually produced a statistically significant
decrease in cortical excitability when measured with
relaxed finger flexors. These results suggest that the com-
bination of voluntary training with augmented electrical
stimulation can produce greater changes in cortical excit-
ability than voluntary training or FES alone.
The SP duration was linearly related to the magnetic
stimulation intensity, as shown previously (Cantello et al.
1992; Inghilleri et al. 1993). The SP measures in the present
study were also in agreement with previously reported
range of SP (e.g. Taylor et al. 1997; Ho et al. 1998). None of
the training paradigms had an effect on the duration of the
Fig. 3 Example of the stimulus-response (SR) curve fit for the same cortical SP indicating that cortical inhibitory circuits were
representative subject as in Fig. 2. Calculated MEP areas are marked unaffected by the training. This result, in combination with
with circles and Xs for before and after TFES training respectively. the increased MEP observed with training suggests that
The SR-curves (grey—before, black—after) are fit to all correspond- facilitatory and inhibitory circuitry were selectively influ-
ing points (circles or Xs)
enced by the different training interventions.
significant. The two-way rmANOVA revealed a significant A very small (*0.3 ms), but statistically significant
time effect (P = 0.005), with no significant effect of ses- increase in MEP onset latency was observed after FES and
sion (P = 0.756) or interaction (P = 0.563). TFES training sessions. While the onset latency increases
Fig. 4 Group results of the
training effects on MEP area
measured from the finger
flexors. Group average data
after training (black line, filled
circles) are superimposed on
group average data before
training (grey line, empty
circles), error bars represent
standard error. MEP area is
expressed in % of the baseline
MEPmax and it is plotted against
% of normalised stimulus
intensity. Columns (from left to
right) stand for functional
electrical stimulation (FES),
therapeutic functional electrical
stimulation (TFES), and
voluntary movement (VOL)
training. Panel a shows the
results of experiment 1 (relaxed
finger flexors, n = 11), panel b
shows the results of experiment
2 (contracted finger flexors,
n = 9). The P-values indicate
significant training effects

123
64 Exp Brain Res (2008) 191:57–66

Fig. 5 Group results of the

training effects on MEP onset
latency (LAT) and silent period
(SP) duration measured from
the finger flexors. Group
average data after training
(black line, filled circles) are
superimposed on group average
data before training (grey line,
empty circles), error bars
represent standard error. LAT
and SP are expressed in
milliseconds and they are
plotted against % of maximal
stimulation intensity. Columns
(from left to right) stand for
functional electrical stimulation
(FES), therapeutic functional
electrical stimulation (TFES),
and voluntary movement (VOL)
training. Panel a shows the
effects on LAT, panel b shows
the effects on SP (experiment 2,
n = 9). The P-values indicate
significant training effects

as the stimulating coil is moved away from the optimal
position (Fuhr et al. 1991), we took extraordinary precau-
tion to fix the coil position and orientation relative to the
head throughout the experiments. MEP onset latency is
also affected by fatigue. For example, Taylor et al. (1999)
reported a mean increase in onset latency of 0.7 ms during
a sustained fatiguing contraction of the biceps brachii
muscle, and Andersen et al. (2003) reported a 1 ms
increase after a sustained fatiguing contraction of the
abductor digiti minimi which was followed by a fast
recovery (*30 s) after completion of the task. Our data
shows longer lasting ([5 min) increase of MEP onset
latency, yet it is unlikely that this small increase in latency
affected the outcome of the study.
Effect of motor training on cortical excitability
The VOL training paradigm used in the present study did
not increase excitability of the cortical projections to the
finger flexor muscles. Grasping and releasing are trivial
hand functions for healthy young adults and, therefore, it
would be highly unlikely if the subjects learned a new motor
skill. Similarly, our FES training paradigm was insufficient
to produce an increase in excitability. Our subjects were
instructed to relax during FES training to passively endure
the electrical stimulation. These results correspond to the
finding of Perez et al. (2004), who showed that cortical
excitability increased after short-term skilful training, yet
remained unchanged after non-skilled or passive training.
However, the afferent input from the electrical stimulation
of the periphery can influence the excitability of corre-
sponding motor areas (e.g. Ridding et al. 2000; Knash et al.
2003; Khaslavskaia and Sinkjær 2005).
The primary goal of the present study was to investigate
if one session of upper-limb TFES produced measurable
changes in cortical excitability of the target muscles, and
thus our training paradigms were designed to mimic the
therapy applied for the affected upper-limb after stroke.
The effect of more refined electrical stimulation, closer to
our TFES, was studied by Kido Thompson and Stein
(2004). They applied electrical stimulation to the common
peroneal nerve during the swing phase of walking in
healthy volunteers (such as in FES to alleviate drop-foot
symptoms). Their intervention facilitated the MEPs of the
tibialis anterior, which they interpreted as evidence for
increased cortical excitability. Their data, as well as ours,
showed no changes in SP duration. It is likely that cortical
facilitatory and inhibitory circuits respond differently to
FES if the stimulation is presented as an augmentation of
voluntary activation as opposed to FES alone.
Methodological considerations
The participants had control over the initiation of the muscle
stimulation sequence during the training sessions. Before the
first training session the subjects practised the stimulation

123
Exp Brain Res (2008) 191:57–66
and exercise. Here, some emphasis was placed on the pace of
the training and consistency in the movements. While the
actual experiment was self-paced, all subjects performed at a
similar pace resulting in *130–140 movement cycles dur-
ing the 20-min exercise period. Similarly, the movement
range was alike in all sessions (see Fig. 1).
In contrast with the older population of stroke survivors
who undergo rehabilitation training, our subjects were
young healthy adults. Age affects both motor performance
(e.g. Smith et al. 1999) and motor cortex excitability
(Oliviero et al. 2006), thus elderly stroke survivors might
be expected to have different responses to our training
paradigms. A specific clinical study is needed to investigate
this possibility further.
Due to technical failure, we used different magnetic
stimulators for the two experiments. The Magstim 200 used
in experiment 2 has a longer recharge time than does the
Magstim Rapid2, thus the inter-stimulus interval had to be
increased. To maintain a similar assessment time between
the two experiments, the number of pulses used to con-
struct the stimulus-response curves was decreased, and the
range of stimulation intensities was tailored to the indi-
vidual responses of the subjects while the increments were
reduced by 1%. Despite the change of magnetic stimula-
tors, the same magnetic coil was used in both experiments.
Nevertheless, the two stimulators do produce different
magnetic fields (Magstim Rapid2—biphasic field, Magstim
200—monophasic field) and the Magstim 200 produces a
stronger field than does the Magstim Rapid2 for the same
percent of maximum stimulator output. The different field
strength between the two stimulators was overcome by
normalising the stimulation intensities. Most importantly,
the MEPs produced in the two experiments should be
compared separately; it is not meaningful to compare the
MEPs directly between the two experiments. These dif-
ferences have no bearing on the outcome of the study.
Pilot experiments were conducted prior to the study in
which the training lasted 30 min. The 30-min training
session produced very clear indications of fatigue. For
example, the intensity of the electrical stimulation con-
trolling the fingers had to be increased by 1–3 mA during
the experiment to provide the necessary force output to
compensate for the decreasing muscle force (Winslow et al.
2003). With the revised 20-min training protocol partici-
pants did not report fatigue and the initial stimulation
intensity was sufficient throughout the interventions.
While we did not specifically monitor fatigue during the
experiments, the decrease in M-wave in some of the training
sessions with electrical stimulation suggests that fatigue was
still a factor (Sacco et al. 1998; Pitcher and Miles 2002). To
verify this possibility, we analysed the power spectrum
density of the pre-trigger FDP EMG from experiment 2. In
this case, the power spectrum density shifted to lower
65
frequencies after TFES and VOL (in both cases P \ 0.001,
one-way rmANOVA), and showed similar, but statistically
insignificant, trend after FES (P = 0.16). This suggests the
muscles may have been fatigued following the training.
However, the duration of pre-trigger EMG was only 50 ms
(200 data points at 4 kHz sampling rate) and this is probably
insufficient to reliably determine the level of fatigue. Thus
we cannot exclude fatigue as a confounding factor and the
effect of muscle fatigue on the MEP in this context requires
further investigation.
Implications for rehabilitation
Electrical stimulation of a motor nerve through surface
electrodes produces a muscle contraction and a barrage of
cutaneous and proprioceptive afferent feedback. These
diverse, yet highly correlated inputs to the CNS may play a
crucial role in multimodal integration, especially when
they coincide with voluntary movement or the intent to
move. This interaction may enhance CNS plasticity and
facilitate motor rehabilitation after a CNS injury.
Based on the results of our study we suggest that the
combination of voluntary effort and FES has a greater
potential to induce motor cortex plasticity than does elec-
trical stimulation or voluntary training alone. Therefore,
the present study suggests that the therapeutic application
of FES may be a more effective approach to stroke reha-
bilitation than either FES or repetitive voluntary training.
Acknowledgements The authors wish to thank the participants of
the study and the personnel of the Centre of Sensory-Motor Interac-
tion, Aalborg University, for their help during the study. This study
was supported by grants from the Danish National Research Foun-
dation and Det Obelske Familiefond.
References
Andersen B, Westlund B, Krarup C (2003) Failure of activation of
spinal motoneurons after muscle fatigue in healthy subjects
studied by transcranial magnetic stimulation. J Physiol 551:345–
356
Cantello R, Gianelli M, Civardi C, Mutani R (1992) Magnetic brain
stimulation: the silent period after the motor evoked potential.
Neurology 42:1951–1959
Castro-Alamancos M, Garcia-Segura L, Borrell J (1992) Transfer of
function to a specific area of the cortex after induced recovery
from brain damage. Eur J Neurosci 4:853–863
Chae J, Yu D (1999) Neuromuscular stimulation for motor relearning
in hemiplegia. Crit Rev Phys Med Rehabil Med 11:279–297
Chen R (2000) Studies of human motor physiology with transcranial
magnetic stimulation. Muscle Nerve Suppl 9:26–32
Classen J, Liepert J, Wise SP, Hellett M, Cohen LG (1998) Rapid
plasticity of human cortical movement representation induced by
practice. J Neurophysiol 79:1117–1123
Devanne H, Lavoie BA, Capaday C (1997) Input-output properties
and gain changes in the human corticopinal pathway. Exp Brain
Res 114:329–338
123
66
Di Piero V, Chollet F, Mac Carthy P, Lenzi G, Frackowiak R (1992)
Motor recovery after acute ischaemic stroke: a metabolic study. J
Neurol Neurosurg Psychiatr 55:990–996
Ferster D (2004) Neuroscience. Blocking plasticity in the visual
cortex. Science 303:1619–1621
Fuhr P, Cohen LG, Roth BJ, Hallet M (1991) Latency of motor
evoked potentials to focal transcranial stimulation varies as a
function of scalp position stimulated. Electroencephalogr Clin
Neurophysiol 81(2):81–89
Gritsenko V, Prochazka A (2004) A functional electric stimulation-
assisted exercise therapy system for hemiplegic hand function.
Arch Phys Med Rehabil 85:881–885
Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G (2000) Develop-
ment of recommendations for SEMG sensors and sensor
placement procedures. J Electromyogr Kinesiol 10(5):361–374
Ho KH, Nithi K, Mills KR (1998) Covariation between human
intrinsic hand muscles of the silent periods and compound
muscle action potentials evoked by magnetic brain stimulation:
evidence for common inhibitory connections. Exp Brain Res
122:433–440
Inghilleri M, Berardelli A, Cruccu G, Manfredi M (1993) Silent
period evoked by transcranial stimulation of the human cortex
and cervicomedullary junction. J Physiol 466:521–534
Jacobs KM, Donoghue JP (1991) Reshaping the cortical motor map
by unmasking latent intracortical connections. Science
251(4996):944–947
Jenkins WM, Merzenich MM (1987) Reorganisation of neocortical
representations after brain injury: a neurophysiological model of
the bases of recovery from stroke. In: Seil FJ, Herbert E, Carlson
B (eds) Progress in brain research, vol 71. Elsevier, Amsterdam,
p 249
Khaslavskaia S, Sinkjaer T (2005) Motor cortex excitability following
repetitive electrical stimulation of the common peroneal nerve
depends on the voluntary drive. Exp Brain Res 162:497–502
Khaslavskaia S, Ladouceur M, Sinkjaer T (2002) Increase in tibialis
anterior motor cortex excitability following repetitive electrical
stimulation of the common peroneal nerve. Exp Brain Res
145(3):309–315
Kido Thompson A, Stein RB (2004) Short-term effects of functional
electrical stimulation on motor-evoked potentials in ankle flexor
and extensor muscles. Exp Brain Res 159:491–500
Knash ME, Kido A, Gorassini M, Chan KM, Stein RB (2003)
Electrical stimulation of the human common peroneal nerve
elicits lasting facilitation of cortical motor-evoked potentials.
Exp Brain Res 153:366–377
McDonnell MN, Ridding MC (2006) Afferent stimulation facilitates
performance on a novel motor task. Exp Brain Res 170:109–115
Muellbacher W, Ziemann U, Boroojerdi B, Cohen L, Hallett M
(2001) Role of the human motor cortex in rapid motor learning.
Exp Brain Res 136:431–438
Nudo RJ (2003) Functional and structural plasticity in motor cortex:
implication for stroke recovery. Phys Med Rehabil Clin N Am
14(1 Suppl):857–876
Nudo RJ (2007) Postinfarct cortical plasticity and behavioral
recovery. Stroke 38(2 Suppl):840–845
Oliviero A, Profice P, Tonali PA, Pilato F, Saturno E, Dileone M,
Ranieri F, Di Lazzaro V (2006) Effects of ageing on motor
cortex excitability. Neurosci Res 55:74–77
Peckham PH, Knutson JS (2005) Functional electrical stimulation for
neuromuscular applications. Annu Rev Biomed Eng 7:327–360
Perez MA, Lungholt BKS, Nyborg K, Nielsen JB (2004) Motor skill
training induces changes in the excitability of the leg cortical
area in healthy humans. Exp Brain Res 159:197–205
Pitcher JB, Miles TS (2002) Alterations in corticospinal excitability
with imposed vs. voluntary fatigue in human hand muscles.
J Appl Physiol 92:2132–2138
123
Exp Brain Res (2008) 191:57–66
Popovic MB, Popovic DB, Sinkjaer T, Stefanovic A, Schwirtlich L
(2002) Restitution of reaching and grasping promoted by
functional electrical therapy. Artif Organs 26(3):271–275
Popovic MB, Popovic DB, Sinkjaer T, Stefanovic A, Schwirtlich L
(2003) Clinical evaluation of functional electrical therapy in
acute hemiplegic subjects. J Rehabil Res Dev 40:443–453
Popovic DB, Popovic MB, Sinkjaer T, Stefanovic A, Schwirtlich L
(2004) Therapy of paretic arm in hemiplegic subjects augmented
with a neural prosthesis: a cross-over study. Can J Physiol
Pharmacol 82:749–756
Popovic MR, Thrasher TA, Adams ME, Takes V, Zivanovic V,
Tonack MI (2006) Functional electrical therapy: retraining
grasping in spinal cord injury. Spinal Cord 44(3):143–151
Powell J, Pandyan D, Granat M, Cameron M, Stott DJ (1990)
Electrical stimulation of wrist extensors in poststroke hemiple-
gia. Stroke 30:1384–1389
Press WH, Flannery BP, Teukolsky SA, Vetterling WT (1986)
Numerical recipes. Cambridge University Press, Cambridge
Ridding MC, Rothwell JC (1995) Reorganisation in human motor
cortex. Can J Physiol Pharmacol 73(2):218–222
Ridding MC, Rothwell JC (1997) Stimulus/response curves as a
method of measuring motor cortical excitability in man.
Electroencephalogr Clin Neurophysiol 105:340–344
Ridding MC, Rothwell JC (1999) Afferent input and cortical
organisation: a study with magnetic stimulation. Exp Brain Res
126:536–544
Ridding MC, Brouwer B, Miles TS, Pitcher JB, Thompson PD (2000)
Changes in muscle responses to stimulation of the motor cortex
induced by peripheral nerve stimulation in human subjects. Exp
Brain Res 131:135–143
Rossini PM, Pauri F (2000) Neuromagnetic integrated methods
tracking human brain mechanisms of sensorimotor areas ‘plastic’
reorganisation. Brain Res Rev 33:131–154
Rosenkranz K, Rothwell JC (2004) The effect of sensory input and
attention on the sensorimotor organization of the hand area of the
human motor cortex. J Physiol 561(Pt 1):307–320
Sacco P, Newberry R, McFadden L, Brown T, McComas AJ (1998)
Depression of human electromyographic activity by fatigue of
synergistic muscle. Muscle Nerve 20:710–717
Seitz RJ, Bütefisch CM, Kleiser R, Hömberg V (2004) Reorganisation
of cerebral circuits in human ischemic brain disease. Restor
Neurol Neurosci 22(3–5):207–229
Siebner HR, Rothwell JC (2003) Transcranial magnetic stimulation:
new insights into representational cortical plasticity. Exp Brain
Res 148:1–16
Smith CD, Umberger GH, Manning EL, Slevin JT, Wekstein DR,
Schmitt FA, Markesbery WR, Zhang Z, Gerhardt GA, Kryscio
RJ, Gash DM (1999) Critical decline in fine motor hand
movements in human ageing. Neurology 53(7):1458–1461
Stetkarova I, Leis AA, Stokic DS, Delapasse JS, Tarkka IM (1994)
Characteristics of the silent period after transcranial magnetic
stimulation. Am J Phys Med Rehabil 73:98–102
Taylor JL, Allen GM, Butler JE, Gandevia SC (1997) Effect of
contraction strength on responses in biceps brachii and adductor
pollicis to transcranial magnetic stimulation. Exp Brain Res
117:472–478
Taylor JL, Butler JE, Gandevia SC (1999) Altered responses of
human elbow flexors to peripheral-nerve and cortical stimulation
during a sustained maximal voluntary contraction. Exp Brain
Res 127:108–115
Thrasher TA, Flett HM, Popovic MR (2006) Gait training regimen for
incomplete spinal cord injury using functional electrical stimu-
lation. Spinal Cord 44(6):351–367
Winslow J, Jacobs PL, Tepavac D (2003) Fatigue compensation
during FES using surface EMG. J Electromyogr Kinesiol
13:555–568