ita
c s
to TMS with nociceptive stimuli that produce prominent changes
intracortical mechanisms and the changes in them associated with
motor cortex plasticity (Reis et al., 2008). These include (i) paired-
pulse TMS protocols to study inhibitory and excitatory interneuro-
nal connectivity within a specific region (such as short interval
intracortical inhibition; SICI) and interactions between various
cortical regions, within and across cerebral hemispheres; (ii) repet-
itive TMS (rTMS) to induce cortical plasticity; (iii) paired associa-
tive stimulation (PAS) of the motor cortex and of a peripheral
axon discharges its excitability changes. In addition, when a neu-
and will not excite the same number of intracortical axons as it
did originally. The excitability of axons accessed by the (stronger)
test stimulus will not (and cannot) be the same as with the uncon-
ditioned control stimulus (Chan et al., 2002), unless the interval
between stimuli is sufficiently long to allow full recovery. Unfortu-
nately, no-one knows what that interval is for the axons of cortical
interneurones. Whether axons of intracortical interneurones
are myelinated or not, single and repetitive discharges will
change axonal excitability through ionic mechanisms and the
Clinical Neurophysiology 121 (2010) 121–123
lab
ro
.e lnerve to probe enduring changes in somato-motor connectivity,in spinal reflex pathways (Kofler et al., 2008). [There were differ-
ences in the extent and time course of the changes in on-going
EMG and the MEP, but perhaps this is not unexpected with two
such different methods of measuring motoneurone output (see
Pierrot-Deseilligny and Burke (2005)).]
Recently there has been a profusion of refinements focusing on
rone discharges there is an obligatory after-hyperpolarization be-
fore it regains baseline excitability. Activation of the axons of
intracortical interneurons will lead to obligatory phases of refrac-
toriness and supernormality following single stimuli, and axons
will hyperpolarise following trains of stimuli. With double or
repetitive stimuli, a stimulus of exactly the same strength cannotEditorial
Caveats when studying motor cortex exc
of movement using transcranial magneti
There is little doubt that complex mechanisms underlie the
function of the human spinal cord, and that the use of physiological
approaches to dissect individual circuits, their supraspinal control
and the changes that occur in disease requires carefully controlled
experiments using well validated techniques. The cerebral cortex
represents a quantum leap in complexity, and comparable studies
on the cortical control of movement are correspondingly more
problematic. In 2005, we wrote: ‘‘Over recent years, reappraisal of
the role of direct cortico-motoneuronal projections in higher primates
including humans has led to the view that the control of movement re-
sides in the motor cortical centres that drive spinal motoneurone pools
to produce the supraspinally crafted movement. This view belies the
complex interneuronal machinery that resides in the spinal cord. It is
a thesis of this book that the final movement is only that part of the
supraspinally derived programme that the spinal cord circuitry deems
appropriate. While the capacity of the spinal cord to generate or sus-
tain even simple movements, particularly in human subjects, is lim-
ited, the influence that it plays in shaping the final motor output
should not be underestimated.” (Pierrot-Deseilligny and Burke,
2005).
The development of transcranial magnetic stimulation (TMS) to
stimulate the primary motor cortex (M1) through the scalp in in-
tact co-operative human subjects allows corticospinal function to
be studied in health and disease, and has led to considerable
advances in motor control physiology (Rothwell, 1997), rapid (per-
haps overly rapid) take-up by researchers with varying experience
in neurophysiological studies, and cautionary warning from some
proponents (e.g. Ziemann, 2009). Changes in the motor evoked
potential (MEP) must occur when there are changes in spinal cord
circuitry unless those changes are purely presynaptic: this is well
illustrated in a recent study demonstrating inhibition of the MEP
Contents lists avai
Clinical Neu
journal homepage: www1388-2457/$36.00  2009 International Federation of Clinical Neurophysiology. Publish
doi:10.1016/j.clinph.2009.10.009bility and the cortical control
timulation
and (iv) transcranial direct current stimulation (tDCS) of M1 to
manipulate cortical excitability directly. However, a number of is-
sues needs to be kept in mind when interpreting the results of TMS
studies that use the MEP as the test potential. The present authors
believe that disregard for these caveats is a major cause for the var-
iability in reported findings, and that many studies in the litera-
ture, including some in this journal, may need to be re-evaluated.
The issues may be complex and difficult to control, but it would
be naïve to assume that the processes underlying TMS are simple,
particularly when it is appreciated that interneuronal connectivity
within motor cortex is assessed using changes in the EMG potential
of a remote limb muscle whose motoneurones are subjected to
many segmental and supraspinal influences other than the direct
cortico-motoneuronal input.
1. TMS is not really focal, and the ability to activate discrete
brain regions or target specific neural connections is only relative.
In different stimulus protocols, TMS will activate a number of sys-
tems and have effects that depend on a balance of opposing ac-
tions. With standard round coils, the induced current in the brain
flows from an annulus underneath the coil, some 8–12 cm in diam-
eter. The effective stimulation point can be any site on the circum-
ference of the coil. With coils wound in a figure-of-8 shape, the
magnetic flux is strongest at the crossing of the 8, so that the effec-
tive stimulus is twice as great there (Rothwell, 1997). However,
while this produces a stronger stimulus at that point, it is not really
focal, and this is relevant when one appreciates that subthreshold
stimulation (such as would occur away from the crossing) can have
significant effects on cortical excitability (see below).
2. It is remarkable that most studies and reviews of TMS do not
address precisely which elements in M1 are stimulated by TMS:
the stimulus activates axons, not cell bodies, and whenever an
le at ScienceDirect
physiology
sevier .com/locate /c l inphed by Elsevier Ireland Ltd. All rights reserved.

Gandevia, 2004), possibly even as distal as the brachial plexus. Un-
ysiosodium–potassium pump, and this will occur whether the activity
is natural or due to some conditioning stimulus paradigm (Vagg
et al., 1998; Bostock et al., 2003; Kiernan et al., 2004).
3. If cortical excitability changes, the threshold for the test MEP
maychange, and so toowould the subthreshold conditioning stimuli
that are referenced to the threshold for the testMEP. Somepublished
studies have not recalibrated the threshold for the test MEP during
experiments specifically designed to change cortical excitability.
However, even if they did, it remains an assumption that the excit-
ability of all relevant interneuronal inhibitory and facilitatory cir-
cuits would maintain the same relationship to the threshold for
the testMEPwhen that threshold is changed. This is particularly rel-
evantwhen there is pathology: if the threshold for the testMEPwere
higher, the conditioning stimulus would be higher and SICI would
differ from controls (as has been reported), merely because of this.
4. The corticospinal volley underlying the MEP is not a single
volley, and the latency (and complexity) of motoneuronal dis-
charge will probably depend on summation of repetitive volleys
set up by the single stimulus in the motoneurone and on the back-
ground excitability of the motoneurone (Rothwell, 1997). If a con-
ditioning stimulus causes motoneurones to discharge on different I
waves in the test volley, the interval between I waves (1.5 ms) pro-
vides an opportunity for other circuits, spinal and supraspinal, to
affect the discharge. In addition, if an effect is exerted on a specific
I wave, changes in cortical excitability could cause a change in the
I-wave composition of the test volley, independent of any change
in cortical connectivity. Importantly, control experiments con-
ducted under one set of conditions may not be valid under other
conditions.
5. The corticospinal influences on the motoneurone pool will
depend on both the direct monosynaptic cortico-motoneuronal
projection (if there is a direct monosynaptic projection) and on
any indirectly relayed di- and oligosynaptic corticospinal projec-
tions to the target motoneurone pool (Pierrot-Deseilligny, 2002;
Pierrot-Deseilligny and Burke, 2005). [Here it is worth noting that
there is no evidence for propriospinal projections to motoneurones
innervating the intrinsic muscles of the hand, though there are for
more proximal upper limb muscles, some of which may lack a
monosynaptic corticospinal input.] This issue has two implica-
tions: (a) given the complexities of the descending volley, the indi-
rect inputs will not be recognised in the MEP unless special
techniques are used to identify them; (b) whenever the MEP is
used as the test potential to demonstrate a change in cortical excit-
ability, the resulting MEP will depend on not just the excitability of
the spinal motoneurone pool but also on the transmission to it of
indirect corticospinal influences. It is necessary but not sufficient
to demonstrate that a stimulus paradigm has no effect on spinal
motoneurone excitability as tested using H-reflexes (or possibly
F waves). However, the H-reflex may not be the ideal measure to
test spinal motoneurone excitability in this situation because,
particularly at rest, the motoneurone pools excited by TMS and
Ia-afferent inputs may be different, at least for extensor carpi radi-
alis (Morita et al., 1999; Nielsen et al., 1999). There may be a num-
ber of reasons for this, e.g., the propriospinally transmitted
component of the corticospinal volley will affect the MEP much
more than the H-reflex (see above), and presynaptic inhibition
and post-activation depression can affect the Ia volley but not
the corticospinal volley (Pierrot-Deseilligny and Burke, 2005).
Unfortunately, F waves are not really a good measure of motoneu-
rone excitability (Nielsen and Hultborn, 1996; Espiritu et al., 2003;
Lin and Floeter, 2004; Rivner, 2008). Consideration needs to be gi-
ven to whether the conditioning paradigm could have influenced
transmission through propriospinal circuits as well. Despite the
122 Editorial / Clinical Neurophreservations expressed in point #6 below, the only techniques that
will adequately control for this are transcranial electrical stimula-
tion and possibly cervico-medullary stimulation (see next point).der these circumstances, it is conceivable that a change in moto-
neurone excitability might occur without changes in the ‘‘CMEP”.
7. Some authors have reported studies using a standard ‘‘1-mV
MEP” without having first constructed a recruitment curve for the
MEP showing the input–output relationship between stimulus
strength and MEP size (see Ridding and Rothwell (1997)). A ‘‘stan-
dard” 1-mV size could represent a near-maximal MEP for some
subjects, but 50% of maximum for others. This creates enormous
variability across subjects. Recent evidence indicates that the stim-
ulus–response curve should always be measured and MEP size
should be set as a percentage of the maximal M wave, so avoiding
issues due to non-linearity within the spinal motoneurone pool
(A. Lackmy and V. Marchand-Pauvert, unpublished observations).
Acknowledgements
The authors thank Dr Sabine Meunier and Professor Ulf
Ziemann for suggestions and constructive comments on the text,
and acknowledge grant support from NHMRC of Australia.
References
Bostock H, Campero M, Serra J, Ochoa J. Velocity recovery cycles of C fibres
innervating human skin. J Physiol 2003;553:649–63.
Chan JHL, Lin CS-Y, Pierrot-Deseilligny E, Burke D. Excitability changes in human
peripheral nerve axons in a paradigm mimicking paired-pulse transcranial
magnetic stimulation. J Physiol 2002;542:951–61.
Di Lazzaro V, Ziemann U, Lemon RN. State of the art: physiology of transcranial
motor cortex stimulation. Brain Stimulation 2008;1:345–62.
Espiritu MG, Lin CS-Y, Burke D. Motoneuron excitability and the F wave. Muscle
Nerve 2003;27:720–7.
KiernanMC, Lin CS-Y, Burke D. Differences in activity-dependent hyperpolarization
in human sensory and motor axons. J Physiol 2004;558:341–9.
Kofler M, Valls-Solé J, Fuhr P, Schindler C, Zaccaria BR, Saltuari L. Sensory
modulation of voluntary and TMS-induced activation in hand muscles. Exp
Brain Res 2008;188:399–409.
Lin JZ, Floeter MK. Do F-wave measurements detect changes in motor neuron
excitability? Muscle Nerve 2004;30:289–94.
Morita H, Baumgarten J, Petersen N, Christensen LO, Nielsen J. Recruitment of
extensor-carpi-radialis motor units by transcranial magnetic stimulation and
radial-nerve stimulation in human subjects. Exp Brain Res 1999;128:557–62.
Nielsen JB, Hultborn H. Comments: methodological problems of comparing F
responses and H reflexes. Muscle Nerve 1996;19:1347–8.6. Demonstrating that a particular manoeuvre has a cortical lo-
cus of action does not necessarily exclude that the manoeuvre also
produced changes at subcortical and spinal levels, and that these
might have contributed to changes in the MEP. The currently sim-
plest way to do so would involve demonstrating that there are no
demonstrable changes, facilitatory or inhibitory on the MEP to
transcranial electrical stimulation (TES). This is based on the view
that TES activates the corticospinal system directly by excitation of
corticospinal axons. However, this is true only near threshold and
in the pre-activated target muscle (e.g. Di Lazzaro et al., 2008). In
addition, TES may be painful, as may recording the cervico-medul-
lary evoked potential (CMEP; see Taylor and Gandevia (2004)), and
attention, anticipation and discomfort could change ‘‘motor set”.
This is not an insignificant issue given that these factors can change
the size MEP, and TMS and TES trials should therefore be randomly
presented within the same block of trials to exclude differences in
‘‘motor set”. Discomfort aside, the CMEP would constitute an
acceptable control if it excited the corticospinal pathways at the
decussation of the pyramids (as it is believed to do; Ugawa et al.,
1991a,b), because this would then be above the C3–C4 segments
where the interneurones of the propriospinal system are believed
to be located (Pierrot-Deseilligny, 2002). However, when strong
(and presumably when not directed appropriately), this stimulus
can excite motor axons in the peripheral nerve roots (Taylor and
logy 121 (2010) 121–123Nielsen J, Morita H, Baumgarten J, Petersen N, Christensen LO. On the comparability
of H-reflexes and MEPs. Electroenceph Clin Neurophysiol 1999(Suppl. 51):
93–401.