Intracortical inhibitory circuits and sensory input: a study with
transcranial magnetic stimulation in humans
Carlo Trompetto, Alessandro Buccolieri, Giovanni Abbruzzese
*
Department of Neurological Sciences and Vision, University of Genoa, Via de Toni, 16132 Genoa, Italy
Received 25 July 2000; received in revised form 25 October 2000; accepted 2 November 2000
Abstract
We compared intracortical inhibition (ICI) following paired transcranial magnetic stimulation (TMS) (interstimulus
interval, ISI: 3 ms) and the inhibition of motor evoked potentials (MEPs) to TMS induced by stimulation of the median
nerve (ISI: 200 ms) in six normal subjects. MEPs evoked by focal TMS were recorded in the relaxed opponens pollicis
muscle and the size of the conditioned responses was expressed as a percentage of the size of the non-conditioned
responses. Maximal ICI, ranging from 4 to 40%, and inhibition after median nerve stimulation, ranging from 11 to 68%,
were signi®cantly correlated. The results suggest that both phenomena are possibly mediated by the same g amino-
butyric acid-ergic (GABAergic) inhibitory circuits and that afferent inputs to the cortex may contribute to their physio-
logical activation. q 2001 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Transcranial magnetic stimulation; Motor evoked potentials; Intracortical inhibition; Peripheral stimulation; Inhibitory
circuits
Using a paired-pulse transcranial magnetic stimulation
(TMS) protocol, it has been shown that a conditioning
subthreshold stimulus over the motor cortex suppress the
motor potential (MEP) evoked in the intrinsic hand muscles
by a subsequent suprathreshold stimulus given at intersti-
mulus intervals (ISIs) of 1±6 ms [6]. The mechanisms
responsible for this suppression, (intracortical inhibition,
ICI), are thought to be cortical in origin and g aminobutyric
acid-ergic (GABAergic) circuits are supposed to be impli-
cated [6,15]. Chen at al. [3] studying amputees found that
the amount of ICI was reduced contralaterally to the ampu-
tation, suggesting that reduction of GABAergic inhibition
participate in plastic changes after amputation in humans.
This observation supports the hypothesis that motor cortical
plasticity may be in¯uenced by changes in peripheral feed-
back [16]. Recently, it has been demonstrated that MEPs to
TMS are reduced 200±1000 ms after a conditioning stimu-
lation of the contralateral median nerve because of
decreased motor cortex excitability induced by peripheral
activation [4]. It might be important, therefore, to ®nd out if
neural circuits involved in ICI are sensitive to afferent
inputs.
In this study we compared the amount of ICI and that of
MEP inhibition following median nerve stimulation. As the
two phenomena are both inhibitory and thought to be corti-
cal in origin, the ®nding of a correlation between them in
individual subjects would suggest a common neural
mechanism and shed new light on physiological activation
of the neural circuits tested by paired TMS.
Six neurologically normal, right handed subjects, ®ve
male and one female, aged 46 ^ 18 years (mean ^ SD)
were investigated. Informed consent was obtained from all
the subjects and the Local Ethical Committee approved the
experimental protocol. Subjects were seated in a comforta-
ble reclining chair and electromyographic (EMG) activity
was recorded from the right abductor pollicis brevis (APB)
muscle using Ag-AgCl surface electrodes placed over the
muscle belly, with the reference over the metacarpophalan-
geal joint. The EMG signals were ampli®ed ( £ 2000), and
®ltered (bandwidth 50±2000 Hz, 23 dB), captured on a
computer and converted by an analog-to-digital interface
at a sampling rate of 5 kHz for further analysis. In the ®rst
recording session, we tested ICI using several conditioning
stimulus intensities, in order to ®nd the intensity of the
conditioning stimulus inducing the maximal ICI [14]. In
the second session, performed in the same day, we studied
the modi®cation of motor cortex excitability following
median nerve stimulation. TMS was performed with a
Neuroscience Letters 297 (2001) 17±20
0304-3940/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved.
PII: S0304-3940(00)01648-7
www.elsevier.com/locate/neulet
* Corresponding author. Tel.: 1 39-10-353-7040; fax: 1 39-10-
353-8631.
E-mail address: giabbr@csita.unige.it (G. Abbruzzese).

®gure-of-eight coil (external diameter: 9.5 cm) placed over
the optimal position of the left scalp for evoking motor
responses from the right APB muscle, with the handle of
the coil pointing backwards. Stimulus intensity was adjusted
to produce test MEPs of about 1.5 mV peak-to-peak ampli-
tude in the relaxed APB muscle. The subjects were relaxed
throughout the experiments and audio-visual EMG monitor-
ing checked muscle relaxation. The resting trials in which
background EMG activity was present in the prestimulus
time period were rejected off-line. Relaxed motor threshold
(MTh) was determined during complete electrical silence
and de®ned as the stimulus intensity required to evokes
responses of 50±100 mV peak-to-peak amplitude in ®ve
out of ten successive trials. Paired pulse TMS was
performed with two Magstim 200 (Magstim Co. Ltd.,
Dyfed, UK) magnetic stimulators connected to the same
stimulating coil through a Bistim module. Intracortical inhi-
bition was recorded using the method described by Kujirai
et al. [6]. Each recorded epoch lasted 200 ms, of which 100
ms preceded the stimulus. The ISI of 3 ms was investigated.
The conditioning stimulus intensity was progressively
reduced (in steps of 5% of the stimulator output) from rest
MTh value to intensity inducing no effects on the test MEPs.
At least eight conditioned and eight non-conditioned trials
were collected in each subject, for each conditioning stimu-
lus intensity. Conditioned and non-conditioned stimuli were
randomly intermixed and given every 10 s. To test the
modulation of motor cortex excitability by peripheral stimu-
lation, the median nerve was stimulated at the wrist with
standard surface electrodes (cathode proximal) using 0.2 ms
square wave electrical pulses delivered by a Grass S88
stimulator (Grass Instruments Co., Quincy, MS.) at an inten-
sity just above thumb twitch threshold. Median nerve stimu-
lation was timed to precede TMS of 200 ms. Each recorded
epoch lasted 600 ms, of which 100 ms preceded the stimu-
lus. At least ten non-conditioned (TMS alone) and ten
conditioned trials (median nerve stimulation (TMS) were
collected in each subject. Conditioned and non-conditioned
stimuli were randomly intermixed and given every 10 s.
The size (area of the recti®ed EMG signal) of individual
responses was measured off-line. The mean size of the
conditioned responses was calculated and expressed as a
percentage of the mean size of the non-conditioned
responses. The effects of median nerve stimulation or
magnetic conditioning stimulus on test MEP amplitudes
were analysed with the unpaired t-test. Correlation between
the two phenomena was tested with the Pearson correlation
coef®cient. Differences were considered signi®cant if
P , 0:05.
A signi®cant ICI was present in all the subjects (unpaired
t-test: P , 0:01). The conditioning stimulus intensity indu-
cing the maximum of ICI ranged from 64 to 80% of relaxed
MTh (mean 71.5 ^ SD 6.5). Maximal ICI ranged from 4%,
in subject 1, to 40%, in subject 6 (Table 1). The modulation
of test MEPs following median nerve stimulation ranged
from 11%, in subject 1, to 122%, in subject 6 (Table 1).
In ®ve subjects (1±5) median nerve stimulation induced a
signi®cant MEP inhibition (unpaired t-test: P , 0:05).
Subject 6 did not show any MEP inhibition following
median nerve stimulation, but a slight, though not signi®-
cant, facilitation (Table 1). There was a signi®cant correla-
tion between the maximum of ICI and MEP inhibition
following median nerve stimulation in each subject. Subject
1, who showed the larger ICI, had also the more intense
MEP inhibition following median nerve stimulation (Fig.
1A). Similarly, subject 6, showing the less pronounced
ICI, did not show any MEP inhibition to median nerve
stimulation (Fig. 1B). Therefore, in each subject, the ratio
between MEP inhibition following median nerve stimula-
tion and ICI (`inhibition ratio') was always positive (mean
2.14 ^ SD 0.63). Fig. 2 shows the correlation between ICC
and MEP inhibition following median nerve stimulation in
the six subjects (Pearson correlation coef®cient: r ˆ 0:95;
P , 0:005).
This study demonstrates that ICI with paired-pulse TMS
and MEP inhibition following median nerve stimulation are
signi®cantly correlated in normal subjects. It is dif®cult to
draw de®nite conclusions from a correlation study in a small
sample. In particular, it should be noticed that ICI was
present in one subject (no. 6) even if a peripheral stimulus
was not able to produce any inhibition. Furthermore the
inhibition produced by paired-pulse TMS was always higher
than the inhibition after median nerve stimulation. In spite
of these differences, the two phenomena share some
common characteristics: ®rst, obviously, both are inhibitory
phenomena; second, they are supposed to be cortical in
C. Trompetto et al. / Neuroscience Letters 297 (2001) 17±2018
Table 1
Percentage values of inhibitory phenomena in individual subjects
Subjects Age (years) Maximal ICI (%) MEP inhibition following
MN stimulation (%)
Inhibition ratio
a
1 51 4 11 2.75
2 64 8 12 1.5
3 27 16 26 1.62
4 36 22 41 1.86
5 29 33 68 2.06
6 67 40 122 3.05
Mean ^ SD 45.7 ^ 17.6 20.5 ^ 14.0 46.7 ^ 42.6 2.14 ^ 0.6
a
Ratio between MEP inhibition by median nerve (MN) stimulation and maximal ICI.