The process of swallowing comprises an ordered sequence of
sensory and motor events that transport food from mouth to
stomach, whilst ensuring protection of the airway. It is
functionally divided into three stages: oral, pharyngeal and
oesophageal (Miller, 1982), reflecting the anatomical
structures involved. The central neural control of swallowing
is also divided, with cortical centres in conjunction with
afferent feedback from the musculature acting to initiate
and modulate the volitional swallow (Sumi, 1969; Car,
1970; Hockman, Beiger & Weerasuriya, 1979; Jean, 1990;
Martin & Sessle, 1993), whilst the brainstem swallowing
centre (Jean, 1990) generates the sequenced events of
reflex swallowing via the V, IX, X and XII cranial nerve
nuclei. The interaction between each of these elements is
responsible for a normal swallow in healthy subjects, whilst
their disruption will lead to swallowing dysfunction (Wiles,
1991).
The importance of suprabulbar influences in regulating
swallowing has been established in animal studies, through
disruption of cortical swallowing regions by lesioning,
anaesthesia or cooling (Sumi, 1969, 1972b; Hockman et al.
1979; Martin & Sessle, 1993), and disturbing the normal
swallowing pattern. Animal data have also demonstrated
Journal of Physiology (1998), 506.3, pp.857—866
857
Sensorimotor modulation of human cortical swallowing
pathways
Shaheen Hamdy*†, Qasim Aziz*, John C. Rothwell †, Anthony Hobson*
and David G. Thompson*
*University of Manchester Gastroenterology Unit, Hope Hospital, Salford M6 8HD and
†MRC Human Movement and Balance Unit, Institute of Neurology, Queen Square,
London WC1N 3BG, UK
(Received 7 February 1997; accepted after revision 3 October 1997)
1. Transcranial magnetic stimulation over motor areas of cerebral cortex in man can activate
short latency bilateral cortical projections to the pharynx and oesophagus. In the present
paper we investigate the interaction between pathways from each hemisphere and explore
how activity in these pathways is modulated by afferent feedback from the face, pharynx
and oesophagus.
2. Comparison of unilateral and bilateral stimulation (using interstimulus intervals (ISIs) of 1,
5 or 10 ms between shocks) showed spatial summation of responses from each hemisphere at
an ISI of 1 ms, indicating that cortical efferents project onto a shared population of target
neurones. Such summation was not evident at ISIs of 5 or 10 ms. There was little evidence
for transcallosal inhibition of responses from each hemisphere, as described for limb muscles.
3. Single stimuli applied to the vagus nerve in the neck or the supraorbital nerve, which alone
produce intermediate (onset 20—30 ms) and long (50—70 ms) latency reflex responses in the
pharynx and oesophagus, were used to condition the cortical responses. Compared with rest,
responses evoked by cortical stimulation were facilitated when they were timed to coincide
with the late part of the reflex. The onset latency was reduced during both parts of the
reflex response. No facilitation was observed with subthreshold reflex stimuli.
4. Single electrical stimuli applied to the pharynx or oesophagus had no effect on the response
to cortical stimulation. However, trains of stimuli at frequencies varying from 0·2 to 10 Hz
decreased the latency of the cortically evoked responses without consistently influencing
their amplitudes. The effect was site specific: pharyngeal stimulation shortened both
pharyngeal and oesophageal response latencies, whereas oesophageal stimulation shortened
only the oesophageal response latencies.
5. Cortical swallowing motor pathways from each hemisphere interact and their excitability is
modulated in a site-specific manner by sensory input. The latter may produce a mixture of
excitation and inhibition at both brainstem and cortical levels.
6592
Keywords: Cerebral cortex, Deglutition, Magnetic stimulation

that repetitive stimulation of either hemisphere can evoke a
swallow, that stimulation of both cerebral hemispheres,
simultaneously, can enhance the number of swallows evoked
compared with unilateral stimulation (Sumi, 1969), and that
concurrent afferent excitation can facilitate swallowing in an
intensity- and frequency-dependent manner (Sumi, 1969;
Miller, 1972; Beiger & Hockman, 1976; Jean & Car, 1979).
We have shown that the cortical pathways to human
swallowing musculature can be studied by applying trans-
cranial magneto-electric stimuli to the motor cortex whilst
recording the electromyographic (EMG) responses evoked
from the oro-pharynx and upper oesophagus (Aziz et al.
1994; Hamdy et al. 1996). In addition, we have demonstrated
that these pathways are present on both hemispheres,
display marked inter-hemispheric asymmetry independent
of handedness (Aziz, Rothwell, Hamdy, Barlow & Thompson,
1996; Hamdy et al. 1996), and can be facilitated by prior
stimulation of cranial nerve afferents, which evoke reflex
pharyngo-oesophageal EMG responses (Aziz, Rothwell,
Barlow & Thompson, 1995; Hamdy, Aziz, Rothwell, Hobson,
Barlow & Thompson, 1997). It remains uncertain, however,
how the cortical centres on each hemisphere interact, or how
the facilitation of the cortical swallowing pathways is
modulated by afferent stimulation.
We therefore conducted the following series of studies, in
healthy human volunteers, to explore (i) how bilateral
stimulation of the cortical hemispheres would influence the
cortically evoked swallow response, and (ii) whether the
facilitation of the cortical swallowing pathways by afferent
pathway stimulation displays intensity or frequency
dependence.
METHODS
Subjects
All participants in the studies were healthy adult volunteers
recruited from personnel affiliated with the research units involved
in the project. None gave any history of swallowing problems. The
project was formally approved by Salford & Trafford Ethics
Committee. Subjects were given details of the experimental
protocol and all gave full written informed consent before the
studies were conducted.
Magnetic stimulation
Magnetic stimulation of the cerebral cortex and cranial nerves was
performed using two commercially available magnetic stimulators
(Magstim 200, MAGSTIM Company Limited, Whitland, Dyfield,
UK).
Cortical stimulation. Focal magnetic stimulation of one
hemisphere was performed using a 70 mm (outer diameter) figure
of eight coil oriented at an angle of 45 deg to the parasagittal plane,
tangential to the scalp surface and with the anterior edge of the
bifurcation positioned over the site of interest.
Focal bilateral stimulation was performed using two identical
magnetic stimulators and two figure of eight coils, one positioned
over each hemisphere. Both stimulators were attached to a timing
device (Bistim Module, MAGSTIM Company Limited), the output
of which was programmed to discharge the stimulators at intervals
varying from 1—10 ms.
Diffuse stimulation of both hemispheres was performed using a
single magnetic stimulator connected to a 90 mm (outer diameter)
circular coil, which, when discharged over the vertex of the
cranium, provided diffuse stimulation of the cortex beneath the
coil.
Cranial nerve stimulation. This was performed using a magnetic
stimulator connected to a small (50 mm outer diameter) figure of
eight coil, which provided focal stimulation of an area of tissue
approximately one square centimetre beneath its bifurcation.
Stimulation of the vagus nerve was performed by discharging the
coil over the right side of neck, 2 cm below the angle of the jaw, at
the anterior border of the sternocleidomastoid muscle (Aziz et al.
1994, 1995).
Stimulation of the trigeminal nerve was performed by positioning
the centre of the figure of eight coil over the right supra-orbital
branch of the trigeminal nerve (Hamdy et al. 1997). This was
chosen because it is a purely afferent branch, and because its
excitation evokes a bilateral blink reflex (Kimura, Powers & Van
Allen, 1969), thereby confirming that effective stimulation had
occurred.
Combined cranial nerve and cortical stimulation. This was
performed using the two magnetic stimulators connected to the
timing device, programmed to discharge both stimulators at
intervals varying from 10—100 ms.
Electrical stimulation
Electrical stimulation of the pharynx and oesophagus was
performed using two pairs of bipolar platinum ring electrodes built
into a 3 mm diameter intraluminal catheter (Gaeltec, Dunvegan,
Isle of Skye, UK). The electrode pairs were sited 5 and 12 cm from
the distal tip of the catheter, the proximal electrode pair being used
to stimulate the mucosa of the pharynx and the distal pair to
stimulate the mucosa of the upper oesophagus. Each electrode pair
had an inter-electrode distance of 1 cm and each was connected to
an electrical stimulator device (Stimulator Model DS7, Digitimer,
Welwyn Garden City, Hertfordshire, UK) via a trigger generator
(Neurolog System, Digitimer), which delivered trains of stimuli
(pulse duration, 0·1 ms; voltage, 280 V) repetitively, at frequencies
of 0·2—10 Hz.
Combined sensory and cortical stimulation. This was performed
by connecting both the magnetic stimulator and the electrical
stimulator to a timing device (Neurolog System, Digitimer),
programmed to discharge the magnetic stimulator 100 ms after the
last pulse of a train of twenty-five electrical stimuli for each
frequency studied.
Electromyographic recording
The swallow muscle groups chosen for study were the pharyngeal
muscles and the striated muscle of the upper oesophagus.
EMG responses were detected using a second intraluminal catheter
identical to that used for electrical stimulation. The proximal
electrode pair was used to detect pharyngeal EMG responses and
the distal electrode pair to detect oesophageal EMG responses.
Each electrode pair was connected to a pre-amplifier (CED 1902,
Cambridge Electronic Design, Cambridge, UK) with filter settings
of 5—2000 Hz. Response signals were then collected through a
S. Hamdy, Q. Aziz, J. C. Rothwell, A. Hobson and D. G. Thompson
J. Physiol. 506.3
858