138P Journal of Physiology (1994) 479.P
that glutamate is an important excitatory transmitter at
synapses on to trigeminal motoneurones and also serve to
Glutamate-mediated synaptic interactions in the identify the types of synaptic interactions which may be
rat trigeminal motor nucleus mediated by glutamate.
Hsiu-Wen Yang, Kwabena Appenteng, John Curtis, REFERENCES
Ming-Yuan Min and Sikha Saha* Curtis, J.C., Min, M.-Y. & Appenteng, K. (1994). J. Physiol. 476.P,
Departments of Physiology and *Cardiovascular Studies, 75P.
University of Leeds, Leeds LS2 9NQ Saha, S., Appenteng, K. & Batten, T.F.C. (1992). J. Physiol. 446, 378P.
We have used both eletrophysiological and
ultrastructural methods to determine if glutamate may act
as an excitatory transmitter at synapses on to trigeminal
motoneurones. The electrophysiological experiments Short latency cerebellar responses to alveolar nerve
consisted of recording minature EPSP activity from stimulation in the cat
trigeminal motoneurones and assessing the effect on this A. Taylor and R. Durbaba
activity of application of 6-cyano-7-nitroquinoxaline-2-3-
dione (CNQX), an antagonist of the AMPA (a-amino-3- Sherrington School of Physiology, UMDS, St Thomas's
hydroxy-5-methyl-4-isoxazolepropionic acid) sub-type of Hospital Campus, London SEJ 7EH
the glutamate receptor. The preparation used was a tissue From studies in the ferret it has been suggested that part
slice preparation (500 /am thickness) taken from animals of the projection of alveolar nerve afferents to the cerebellar
aged 8 days. The slices were maintained at 25 °C in an cortex is via direct, unrelayed first-order fibres and in this
interface chamber where they were perfused with ACSF respect resembles the vestibular mossy fibre input (Taylor &
containing tetrodotoxin (0-6 /SM). The whole-cell patch Elias, 1984; Elias et al. 1987). Electrophysiological studies
recording method was used to obtain voltage recordings have now been extended to the cat to compare the shortest
from trigeminal motoneurones. The electrodes were filled latency responses to alveolar nerve stimulation with the
with a potassium gluconate solution and the criteria for longer latency ones which are undoubtedly relayed.
acceptance of recordings were essentially as described by Data were obtained from ten cats surgically prepared
Curtis et al. (1994). The miniature EPSP activity seen under halothane anaesthesia and transferred to a-chloralose
following perfusion with tetrodotoxin was completely (50 mg kg' i.v.). Silver wire electrodes were implanted to
abolished following addition of 10 /M CNQX to the bathing stimulate inferior alveolar (IAN) and maxillary (MaxN)
medium (n = 2). This suggests that the excitatory activity in nerves. Arriving afferent volleys were monitored through
this preparation may be primarily mediated by glutamate an electrode stereotaxically implanted in the trigeminal
acting via AMPA receptors. nerve at its entry to the brain. Recordings were made with
We have used post-embedding immunogold labelling at glass-coated tungsten microelectrodes from the cerebellar
the E.M. level to quantify the incidence and types of cortex and from the brainstem in the region of the
synapses formed by glutamate-immunoreactive boutons in trigeminal motor nucleus.
the trigeminal motor nucleus. The immunolabelling was as Responses to IAN stimulation in the cerebellum were
described previously by Saha et al. (1992), but with two detected in the granular layer of the paravermal regions of
modifications. One was that 20 nm colloidal gold particles lobule V and adjacent lobule VI. They consisted of a brief
were used instead of 10 nm particles, and the other involved positive-negative spike followed in 0-25 ms by a negative
the quantification of labelling. The density of background focal synaptic potential. The earliest such complex occurred
labelling was determined by counting the numbers of gold with a latency of 1'65 ms. The trigeminal nerve volley
particles in a frame of fixed size. The frame was placed at occurred at 0-75 ms whilst periodontal mechanoreceptor
thirty randomly selected sites within a grid and the mean cells in the mesencephalic trigeminal nucleus fired at
density, and standard deviation, of labelling calculated 1P05 ms. The afferent volley reached the principal
from the counts obtained. The numbers of gold particles in a trigeminal sensory nucleus at 1P02 ms. Cells in this region
random selection of terminals forming synaptic contact responding to tooth pressure were fired by IAN or MaxN
within the motor nucleus were determined and the areas of stimulation with a delay of 2-0-2-4 ms. In one experiment
the individual terminal measured to allow determination of three such cells were fired antidromically by stimulation of
the density of gold particles. Boutons which contained gold the cerebellar cortex at the point of maximum response
particles at a density of greater than the mean+ 2-6 evoked by MaxN stimulation. The antidromic delay was
standard deviations of the background level were considered 0-6-0-75 ms, so that the expected latency of the cerebellar
to be glutamate immunoreactive. So far 196 boutons have cortical response by this route would be 2-6-3-15 ms,
been examined and 21 % judged to be glutamate- i.e. 0-95-1P50 ms longer than the minimum observed latency
immunoreactive. Glutamate-immunoreactive boutons to nerve stimulation. Thus it appears that these projecting
formed axo-somatic and axo-dendritic and received axo- cells could not have been the origin of the earliest cortical
axonic contacts from other boutons. Taken together, the responses. The latter must either be due to directly
electrophysiological and ultrastructural evidence suggests projecting afferents or to a relay through some other
Journal of Physiology (1994) 479.P 139P
unknown cell group capable of responding very rapidly. The before 24P7 %, mean after 21P5 %, n = 7). In general, baclofen
complexes consisting of a spike and focal synaptic potential, caused a hyperpolarization of the motoneurone (mean
evoked in the granular layer can be shown to occur in an all- 33 mV) and the change in potential reached statistical
or-nothing way with graded stimulation. They are significance in three cases (unpaired t test). In four out of
interpreted as glomerular potentials (Taylor et al. 1987). seven cases where the membrane potential did not change
by more than 4 mV following drug administration, the
EPSP amplitude was reduced by a mean of 47 2 % (range
Elias, S.A., Taylor, A. & Somjen, J. (1987). Proc. Roy. Soc. B. 231, 29 0-64 7 %). The lack of effect of baclofen on the level of
199-216. paired pulse potentiation suggests that the predominant
Tay-lor, A. & Elias, S.A. (1984). Brain Behav. & Evol. 25,157-165.
Taylor, A., Elias, S.A. & Sormjen, G. (1987). Proc. Roy. Soc. B. 231, action of baclofen is on postsynaptic GABAB receptors. The
217-230. results also suggest that GABAB receptors may be absent
from the terminals of spindle afferents. This leaves open the
possibility that presynaptic GABA actions on spindle
afferent terminals may be mediated by GABAA receptors.
Action of baclofen on compound EPSPs recorded
in trigeminal motoneurones in the anaesthetized REFERENCES
rat Grimwood, P.D. & Appenteng, K. (1993). J. Physiol. 459, 463P.
Grimwood, P.D., Appenteng, K. & Curtis, J.C. (1992). J. Phtysiol. 455,
P.D. Grimwood 641-662.
Department of Physiology, University of Leeds, Leeds LS2 Luo, P. & Li, J. (1991). Braint Res. 559, 267-275.
Saha, S., Appenteng, K. & Batten, T.F.C. (1991). Brain Res. 561,
We have previously reported high levels of transmission
failure for EPSPs elicited by single spindle afferents in
single trigeminal motoneurones in pentobarbitone
anaesthetized rats (Grimwood et al. 1992). Using paired Spectral analysis of acoustic myogram during
pulse testing we have shown that these unitary EPSPs can exhausting isometric contraction of human tibialis
undergo potentiation and that the potentiation can be anterior
ascribed to an increase in release probability (Grimwood &
Appenteng, 1993). Therefore, changes in the level of paired I.E. Takamjani
pulse paired potentiation can be used to identify changes in Institute of Physiology, The University, Glasgow G12 8QQ
presynaptic function. GABA immunoreactive boutons have
been reported to form 28 % of all boutons within the The acoustic myogram (AAIG) is a low frequency
trigeminal motor nucleus and to form axo-axonic contacts, vibration emitted by active skeletal muscle, which is usually
as well as axo-dendritic and axo-somatic contacts (Saha et processed like EMG. Rectified-integrated AMG increases
al. 1991). In addition, axo-axonic contacts have been with force up to 80 % of maximal voluntary contraction
identified as the terminals of labelled spindle afferents (Luo (MVC) before it declines with further force increases (Orizio
& Li, 1991). Therefore, one possibility is that GABAergic et al. 1989; Takamjani & Baxendale, 1993). We have
interneurones are responsible for producing the high levels examined the power spectrum of AMIG during fatiguing
of transmission failure at synapses of spindle afferents on to isometric contractions of tibialis anterior.
trigeminal motoneurones. The aim of this work has been to Ethical approval was obtained. Experiments were
determine whether GABAB receptors are present at performed in eight healthy male volunteers who made
synapses on to trigeminal motoneurones and, if so, contractions initially at 40, 60, 80 and 100 % MIVC in a
determine if the receptors are located pre- or post- randomized sequence. At least 3 days of recovery were
synaptically. We have used the GABAB agonist baclofen to allowed between experiments. AMIG was detected by a
activate GABAB receptors and used paired pulse testing to contact sensor (HP21050A, bandwidth 0 02-2000 Hz)
identify pre- or postsynaptic changes in transmission. strapped over tibialis anterior. EMG electrodes were placed
The experiments were performed on rats which were immediately beside the microphone. Signals were then
anaesthetized with pentobarbitone (initial dose 60 mg kg-') amplified, filtered between 2-100 Hz and 2-500 Hz
and paralysed with gallamine triethiodide and artificially respectively and stored on a magnetic tape for later
ventilated as described by Grimwood et al. (1992). analysis. AMIG and EMIG at the beginning, midpoint and
Intracellular recordings were made from masseter synergist end of each contraction were digitized at 250 Hz and
motoneurones of compound EPSPs elicited by electrical 1000 Hz respectively, via a CED 1401 interface and subjected
to FFT analysis to establish their frequenecy content.
stimulation of the masseter nerve (2-3 times threshold)
every 15 s, with alternate single and paired pulses (10 ms Table 1. Mean of median frequency (+ S.E.M.) of the AMIGC
interspike interval). Administration of baclofen i.v. AIVC (%) 40 60 80 100
(0-88-2-40 mg kg-') had no significant effect on the level of Begi 6-9 + 03 Hz 76 + 04 Hz 79 + 03 Hz 86 + 05 Hz
paired pulse potentiation, either for individual units Endl 66+08Hz 6+05Hz 45+03Hz 45+023Hz
(unpaired t test) or for the pooled data (paired t test, mean
140P Journal of Physiology (1994) 479.P
The fatiguing contractions caused compression at all movement without any change in N20, P20 peaks as in
contraction levels. The EMIG median frequencies fell from normal subjects. Preliminary data on radial SEPs from the
108 to 67 Hz at 40 % MVC and from 93 to 59 Hz at AIVC. As affected hand showed enhanced amplitude of the P25 peak
shown in Table 1, similar observations were made for AMIG. (17 %). These results indicate that the human somatosensory
The AMCG frequency spectrum did not change during the cortex deprived of peripheral input is still capable of
40 % AlVC contractions, but at higher forces, there was a processing an afferent volley and could also undergo
shift towards the lower range. dynamic changes during movement although the activity is
less than that of the normal side. Changes in amplitudes of
the cortical responses could be attributed to the adaptive
Ebrahimni-Takamnjani, I. & Baxendale, R.H. (1993). Proc. XXXII Conlg. modifications which may have taken place in the
IUIPS, Glasgow, 284 85/P. somatosensory pathway of the affected side.
Elrahimi-Takamjani, I. & Baxendale, R.H. (1994). J. Physiol. 477.P,
58P. Supported by the AMedical Research Council. V.S.W. was
Oi-izio, C., Peiini, R. & Veicsteinas, A. (1989). Eur. J. Appi. Physiol. suppoirted by the Association of Commonwealth UJniversities.
AMerzenich, AMAI., Kaas, J.H., Wall, J.T., Sur, Al., Nelson, R.J. &
Felleman, D.J. (1983). iNeuros(ience 10, 639-665.
Adaptive changes in the somatosensory cortex after
peripheral nerve injury in man
V.S. Weerasinghe, E.M. Sedgwick, E. Glasspool and
T.B. Docherty Ultrastructural immunohistochemical study of a
Clinical Neurological Sciences, University of Southampton neural pathway from the spinal dorsal horn
anid Southampton General Hospital, Southampton S09 (lamina I) to the nucleus tractus solitarii in the
4XY anaesthetized rat
Cortical sensory representational maps are capable of T.F.C. Batten, P.N. McWilliam, F. Esteves* and
reorganization after various peripheral perturbations. It has D. Lima*
been shown that after peripheral nerve transection or digit
amputation in adult primates cortical representation of the Institute for Cardiovascular Research, Research School of
affected area shows expression of expanded receptive fields Medicine, University of Leeds, Leeds LS2 9JT and
of the adjacent areas (Merzenich et al. 1983). If the human *Institute of Histology & Embryology, Faculty of Medicine,
sensory cortex is subject to the same plastic changes, scalp 4200 Oporto, Portugal
recorded somatosensory evoked potentials (SEPs) should Neuronal tracing studies at light microscopic (LM) level
show similar modifications after peripheral nerve injury. have shown a projection to the nucleus tractus solitarii
Carpal tunnel syndrome provides a model in which the (NTS) from lamina I of the dorsal horn of the rat. Neurones
projection area of median nerve would be expected to show involved were classified into three morphological types:
reduced responsiveness and the SEPs from radial and ulnar fusiform, flattened and pyramidal (Esteves et al. 1993). A
nerves would be enhanced. We have studied the SEPs LAI immunohistochemical analysis (Lima et al. 1993) has
r ecorded from the scalp when the median nerve was shown that many of the fusiform cells in lamina I contain
stimulated at the wrist proximal to the site of compression immunoreactivity (IR) for y-aminobutyric acid (GABA).
in nine patients (age range = 33-67 years) with unilateral Inhibitory projections from these nociceptive cells may
Carpal tunnel syndrome diagnosed by peripheral nerve contribute to effects on the cardiovascular system resulting
conduction studies. Hospital ethical committee approval from peripheral stimuli (Abram et al. 1983; MicMahon et al.
was obtained and the patients gave informed consent for the 1992). In this study we have examined whether lamina I
study. fusiform cells with an NTS projection contain GABA-IR.
SEPs were recorded using standard methods from a scalp Injections of cholera toxin b-subunit (CTb) or biotinylated
ariay of twenty-one electrodes with the subjects at rest and dextran (bDEX) were made into the NTS of adult rats
while making fractionated finger movements. Contralateral (n = 3) under halothane anaesthesia. After 3-4 days
frontal and parietal cortical waves (N20, P20, P25, N30, P45 survival, rats were re-anaesthetized with 35 % chloral
and N60) were present in the affected side although they hydrate (1 ml kg-' i.P.) and perfused with a 2 % para-
were of decreased amplitudes. N20 amplitude was 26 % and formaldehyde and 2 % glutaraldehyde fixative. Segments of
P25, 27 % less than that of the unaffected side, even spinal cord were postfixed overnight and then sectioned at
allowing for the higher amplitude found from the dominant 75 atm on a Vibratome. The tracer in the retrogradely
hand in normal subjects. There was no significant change in labelled lamina I cells was visualized immuno-
the latency. Isopotential contour maps showed spatial histochemically with diaminobenzidine. After drawing the
distribution of the peak amplitudes with smaller areas and cells from the LAI, sections were osmium post-fixed and
lower peaks on the affected side than that of the normal embedded in Epon for electron microscopy (EMu). Serial
side. P25 and N30 peaks were attenuated during finger ultra-thin sections through the stained neurones were then
Journal of Physiology (1994) 479.P 141P
processed for immunogold labelling using antibodies to conditions. Three patients with clinically total bilateral loss
GABA, glutamate, substance P (SP) and calcitonin gene- of labyrinthine functions showed, in comparison to the
related peptide (CGRP). normal subjects, strikingly lower gains and much longer
Of eighteen NTS-projecting fusiform cells so far phase lags. These results suggest that our new driven-
examined, none were found to contain GABA-IR, although helmet, in combination with the scleral search coil
many other unstained fusiform cells in lamina I were technique, is effective in the assessment of the VOR at
GABA-IR. Terminals with GABA-IR, together with relatively high frequencies where visual, proprioceptive
glutamate, SP and CGRP-IR terminals probably (and mental) influences are minimized.
representing primary afferents, synapsed on dendrites of The same technique was applied to oscillate the head at
the stained cells. On the other hand, of sixteen retrogradely 5 Hz during large gaze (eye + head) saccades (about 100 deg).
labelled fusiform cells examined, two contained a high level By matching and subtracting pairs of gaze saccades that
of glutamate-IR, suggesting that the lack of GABA-IR was differed only in the presence or absence of imposed
unlikely to be due to a loss of antigenicity in the cells caused oscillation, we were able to isolate the oscillatory component
by the retrograde labelling. These results argue against a of gaze, eye-in-head and head movements. The changes in
direct inhibitory GABAergic projection from lamina I to the amplitude of the eye-in-head oscillation as a function of
NTS. time (with complementary changes in the oscillation of gaze)
Supported by a British Council Treaty of Windsor grant. yielded a continuous measure of VOR gain. The results
showed that the VOR was, as expected, working quite well
REFERENCES before the gaze saccade, at a gain of about 0 8-0 85, but was
Abram, S.E., Kostreva, D.R., Hopp, F.A. & Kampine, J.P. (1983). almost completely suppressed during the gaze saccade.
Am. J. Physiol. 245, R576-580. Immediately after gaze landed on target, the VOR was
Esteves, F., Lima, D. & Coimbra, A. (1993). Somatosens. Mlotor Res. 10, working again, at an apparently even supranormal gain
203-216. (above 09), resulting in a very flat gaze signal. When we
Lima, D., Avelino, A. & Coimbra, A. (1993). J. Chem. Neuroanat. 6, studied 60 deg saccades, without active head movements,
McMahon, S.E., McWilliam, P.N., Robertson, J. & Kaye, J.C. (1992). imposing 14 Hz head oscillation and using the same
J. Physiol. 452, 224P. matching procedure, we found a partial suppression of the
VOR. Gain was of the order of half the normal value during
the saccade and again there seemed to be a supranormal
VOR gain after the saccade.
Probing of the human vestibulo-ocular reflex by We conclude that high-frequency head oscillations, which
high-frequency, helmet-driven head movements can be conveniently elicited by a helmet carrying an
S. Tabak, J.B.J. Smeets and H. Collewijn oscillating mass, are a very useful tool in studying the VOR
at high resolution.
Department of Physiology, Erasmus University Rotterdam,
Testing the vestibulo-ocular reflex (VOR) by head-
oscillation at relatively high frequencies has potentially two Fusimotor effects of cerebellar outflow in the
advantages: (1) the elimination of visual or proprioceptive anaesthetized cat
contamination of vestibular responses, enabling a cleaner R. Durbaba, A. Taylor, J.F. Rodgers and A.J. Fowle
evaluation of vestibular function; (2) the possibility to probe
the VOR gain continuously with a high time resolution, Sherrington School of Physiology, UMDS, St Thomas's
e.g. to reveal changes during combined head and eye Campus, London SE1 7EH
movements. We produced head rotations in the 2-20 Hz It has long been known that cerebellar stimulation can
range by a new technique, based on a helmet with a torque- cause fusimotor effects in anaesthetized animals (Jansen &
motor, oscillating a mass. Head and eye movements were Matthews, 1962). However, recently Gorassini et al. (1993)
recorded with magnetic sensor-coils in a homogeneous have found that in alert cats local anaesthetic injections in
magnetic field, with the head unrestrained. To assess the the cerebellum, which cause ataxia, do not seem to change
influence of the visual system on the recorded compensatory fusimotor output to the hindlimbs. We have therefore made
eye movements we measured in three conditions: (1) a a further study by stimulating cerebellar paths to determine
stationary visual target; (2) darkness with the subject which are involved in fusimotor control.
imagining the stationary target and (3) a head-fixed target. In four cats surgical preparation was performed under
The results in fifteen healthy subjects were highly halothane anaesthesia which was then replaced by i.v.
consistent. At 2 Hz, VOR gain was near unity; above 2 Hz, a-chloralose (40 mg kg') with i.v. supplements of ketamine
VOR gain started to decrease, but this trend reversed (0-5 mg kg-'). Single spindle afferents from medial
beyond 8 Hz, where gain increased continuously toward gastrocnemious (MG) or tibialis anterior (TA) were recorded
1-1-1-3 at 20 Hz. Phase lag increased with frequency, from a from dorsal rootlets, usually six at a time. MG and TA
few degrees at 2 Hz to about 45 deg at 20 Hz. Above 2 Hz, spindles were tested with ramp and hold stretches repeating
VOR gain was not significantly different for the three every 6s, and the unit types established by conduction
142P Journal of Physiology (1994) 479.P
velocity and by testing with succinylcholine (Taylor et al. second group were paired with females and allowed to court,
1992). Stimulation along tracks through the deep cerebellar nest-build and begin incubation. 2DG was administered to
nuclei and peduncles was via a unipolar stainless steel this group on the first day of incubation at the time when
electrode at 30 deg to the vertical with tip rostral (100 ptA, the male bird was on the nest. Thirty minutes after
0 2 ins, 200 Hz). Another electrode tracked through the red injection, birds were killed by cervical dislocation, the
nucleus (RN) to monitor responses to cerebellar stimulation brains quickly removed, frozen in isopentane cooled to -45
and also to act as a stimulating electrode. Iron marks were to -55 °C with dry ice and serial transverse 20 ,m thick
made and tracks reconstructed from sagittal sections. cryostat sections cut. Alternate sections were either
Analysis of spindle stretch responses showed that the mounted on slides for staining with Cresyl Violet acetate or
most prominent effects were on static fusimotor activity. placed on coverslips on Kodak SB-5 X-ray film for 4 days at
This w-as increased to TA and decreased to AIG when -15 °C each with a series of ['4C]methylmethacrylate
stimulating in the brachium conjunctivum (BC) in sites standards precalibrated for 20 #um tissue sections. After
which evoked monosynaptic responses in the RN. developing, specific neural areas of chosen sections were
Stimulation in the RN in turn caused similar fusimotor identified by the superimposition of a Cresyl Violet stained
effects. There was also an enhancement of dynamic sister section upon the appropriate autoradiogram. For each
fusimotor outflow when stimulating in the rostro-medial neural area, 2DG uptake was quantitatively measured in
BC, effects which are also reproduced by RN stimulation. three to five separate sections for each brain and both right
Stimulation more caudally along a track through the and left areas were estimated for each section using a
posterior interpositus nucleus inhibited static output to Quantimet-image analysis system (Cambridge
both TA and MIG. This could be due to excitation of Instruments).
Purkinje axons passing to vestibular nuclei. Stimulation Statistically significant differences in glucose utilization
within the dorsal division of the lateral vestibular nucleus between non-breeding and incubating males were observed
caused reciprocal static excitation to TA and inhibition to in only four discrete regions. Cell bodies in the tuberal
MG. This was reversed in the ventral division. (basal) hypothalamus, nucleus ovoidalis and nucleus
We conclude that output pathways exist from the preopticus medialis exhibited significant (Student's t test;
cerebellum which can selectively excite static and dynamic P< 0 05) increases in 2DG uptake in the brains of
activity to flexors and to extensors. The problem remains to incubating doves. In contrast, the paleostriatum
determine under what conditions they are used naturally. primitivum decreased in activity (P< 0 05) at the onset of
Supported by Action Research and St Thomas's Hospital incubation.
Research Endowments. Brain sections for FOS immunocytochemistry were
prepared from male doves in four physiological states (i) first
REFERENCES day of incubation, (ii) second day of incubation, (iii) 12-14th
Goorassini, Al., Prochazka, A. & Taylor, J.L. (1993). J. Neurophysiol. day of incubation and (iv) non-breeding isolated controls
70,1853-1862. (n = 6, all groups). Immunoreactivity (FOS-ir) was detected
Jansen, J.K.S. & Alatthews, P.B.C. (1962). J. Physiol. 161, 357-378. using an antibody raised to a 22 amino acid synthetic
Taylor, A., Rodgers, J.F., Fowle, A.J. & Durbaba, R. (1992).
peptide corresponding to the c-terminal of chicken FOS.
J. Physiol. 456, 629-644.
Birds were perfused with Zamboni's fixative under terminal
anaesthesia and serial vibratome sections prepared for
immunocytochemistry using the ABC method. FOS-ir was
observed in cell nuclei in several areas of the forebrain,
Identification of brain areas activated at the onset particularly the hyperstriatum ventrale, hyperstriatum
of incubation behaviour in male ring doves accessorium and nucleus septalis lateralis. Numbers of
R.WzA. Lea, G. Georgiou, Q. Li and P.J. Sharp* FOS-ir cell nuclei in these areas did not differ significantly
between incubating and non-breeding doves. FOS-ir was
Department of Applied Biology, University of Central significantly greater (P< 0 001) in the tuberal (basal)
Lancashire, Preston PR1 2HE and *Roslin Institute hypothalamus of male doves during the first 2 days of
(Edinburgh), liidlothian EH25 9PS incubation compared to non-breeding and late incubation
This study sought to identify areas of the avian brain males. In addition, male doves at all stages of incubation
activated at the onset of parental behaviour in the male ring exhibited significantly more FOS in the nucleus preopticus
dove (Streptopelia risoria). In this monomorphic species, both anterior.
sexes participate in the 2 week incubation of two eggs. Two It is thus concluded that the expression of incubation
markers of neural activity were employed; 2-deoxyglucose behaviour in ring doves is associated with changes in
(2DG) autoradiography and immuno-cytochemistry for the activity and gene transcription in cells of these
presence of the nuclear protein product, FOS. hypothalamic nuclei.
["4C]-labelled 2DG dissolved in saline solution was given
intravenously via a brachial vein at the dose of 125 ,aCi (kg
body weight)ml to two groups of male doves (n =5, 1)0th
groups). One group control remnainedl in isolation whilst the
Journal of Physiology (1994) 479.P 143P
Very slow electrophysiological signals in man: Gardner-Medwin, A.R., Swithenby, S.J., Fiaschi, K., Elbert, T. &
advantages of study by magnetic measurement Kowalik, Z. (1993). Advances in Biomagnetism: Proc. 9th Int. Conf.
on Biomagnetism (Vienna), pp. 9-10.
Department of Physiology, University College London,
London WCJE 6BT
Magnetic measurement with SQUID technology Differing immune functions, including Natural
(Superconducting Quantum Interference Devices) is a Killer cell activity, after left vs. right magnetic
physiological tool that is usually considered best suited to stimulation of human T-P-O cortex
the study of signals in the 1 Hz to 1 kHz domain, for V.E. Amassian, K. Henry*, H. Durkint, S. Chicet,
example magnetocardiography, magnetoencephalography,
and sensory evoked fields. Slowly changing signals (say over J.B. Cracco*, M. Somasundaram*, N. Hassan*,
periods of 20 min: ca 1 mHz) have been regarded as poorly R.Q. Cracco*, P.J. Maccabee* and L. Eberle
suited to the technology, since it is difficult to screen out Departments of Physiology, *Neurology and tPathology,
interfering magnetic fields in this frequency domain. SUNY Health Science Center, Brooklyn, NY 11203, USA
Recently developed techniques for correlation of magnetic A study of differential immune responses to magnetic coil
signals with continuous small movements induced in the (MC) stimulation of left vs. right temporo-parieto-occipital
subject (Gardner-Medwin et al. 1993) have improved the (T-P-O) cortex (Amassian et al. 1994) has been extended to
signal-to-noise ratio so much for quasi-DC magnetic Natural Killer (NK) cells, identified by the CD 16 marker,
recording, that it may be possible to observe new slow and neutrophils. The same five of us, three males (M) and
phenomena, in a range of physiological and pathological two females (F), all right-handed, were subjects. As
situations, that are undetectable with voltage recording. previously, an ovoid MC, 6-0 x 5-5 cm o.d. (Cadwell
laboratories), was sited over T-P-O cortex and 100 stimuli
given per session at 100 % (M) or 90 % (F) of maximum
intensity. Subjects had to identify vocally briefly presented
visual stimuli. Venous blood was drawn: (1) immediately
before, (2) immediately after stimulation lasting 1A-2' h of
one hemisphere, (3) 4-5 h later and (4) at 1-3 day intervals.
The lymphocyte subsets: CD8+ (suppressor-cytotoxic),
41 CD4+ (helper-inducer) and CD16+ (NK) cells were counted
by flow microfluorimetry. Left and right sides were
stimulated at 1-8 week intervals. After left-sided
2 pT cm- stimulation, the maximum change from control value of
NK cell count was +88 (M), +29 (M), +131 (F), -41 (M),
50 mm -52 (F) % and after right-sided stimulation was -20 (M),
-25 (M), -68 (F), -36 (M), -25 (F), i.e., responses to
Fig. 1. Vectors showing horizontal gradients of vertical field stimulation of the two sides usually differed. In plots of NK
(parallel to the interaural axis) in successive movement cycles at vs. CD8+ or CD4+ cell counts, the correlation (p) for NK
the peak of the light-induced increase of the corneo-retinal and CD8+ cells was +0 747 + 0-259 (mean + S.D., n= 10),
current, 6-8 min after light onset following dark adaptation
(Gardner-Medwin et al. 1993). but was lower with CD4+ cells (+ 0 405 + 0 288). Unlike T
To illustrate the magnitudes involved, consider a simple cell subsets, neutrophil counts increased with left- and
situation with a 50 ,sA.mm horizontal current dipole source right-sided stimulation, which implies that a single humoral
20 mm under a flat surface of tissue with resistivity
factor could not account for all our findings. The
500 Qcm. The maximum potential difference detectable at reproducibility of the findings and effect of altering the
the surface would be 80 uV (or much less if the surface is the experimental conditions were tested on a subject. Visual
stimulation with vocalization (two sessions), or viewing and
scalp, with bone underneath). The maximum magnetic field listening to a teaching TV tape yielded similar
20 mm above the surface would be 3 pT. These signals are
well above noise levels at > 1 Hz, but would be easily lost in CD4+ changes within 6 h of MC stimulation but the increase
was reduced when sensory stimuli and motor responses were
drift and interference at 1 mHz. The maximum field absent (Fig. 1).
gradient would be 1P2 pT cm', which is readily detectable at
DC with movement correlation (Fig. 1).
Supported by the Wellcome Trust.
144P Journal of Physiology (1994) 479.P
(I) Normalized cell counts ipsilateral responses had substantiallyT longer latencies than
0 50 V/I
in the active contralateral homologous muscle (mean
, 40 X Visual input: verbal reaction task difference 8 6 + 1X0 ms (mean ± S.E.M., n = 16) for the three
30- : Visual input: verbal reaction task proximal muscles) and because in 4/5 subjects tested with
c' 20- E.....l
t visual or auditory input: no task
1No the small butterfly coil we could evoke ipsilateral responses
o TV - visual and auditory: no task in these muscles. In contrast, the ipsilateral response in IDI
to DC stimulation generally had high thresholds (> 1-4
a55 CD8+ CD4+ times the passive contralateral threshold) and the ipsi-
contralateral latency difference was smaller (4-1 ± 0-6 ms,
Fig. 1. Left-sidled transcranial magnetic T-P-O stimulation.
mean + S.E.M., 71 = 6). No ipsilateral responses were seen
with the butterfly coil.
REFERENCE Our results confirm the presence of ipsilateral responses
Amassian, V.E., Henry, K., Duikin, H., Chice, S., Cracco, J.B., to TMS in proximal upper limb muscles in man
Somasundaiam, AM., Hassan, N., Cracco, R.Q., Alaccabee, P.J. & (Wassermann et al. 1991). The prevalence of such responses
Ebeile, L. (1994). J. Physiol. 475.P, 22P. after damage to the motor system may result from the
enhancement of transmission in the responsible pathways;
these may involve slow corticospinal, cortico-reticular or
Activation of ipsilateral upper limb muscles by Supported by Action Research and the WVellcome Trust.
transcranial magnetic stimulation in man REFERENCE
Anna P. Basu, Ailie Turton and R.N. Lemon Carr, L.J., Hariison, L.AM., Evans, A.L. & Stephens, J.A. (1993). Brain
Department of Anatomy, Downing Street, Cambridge 116,1223-1247.
Wassermann, E.M., Fuhr, P., Cohen, L.G-. & Hallett, M. (1991).
CB2 3DY Nleurology 41, 1795-1799.
The motor cortex exerts a strong excitatory influence
over muscles of the contralateral upper limb, but weaker
activation of ipsilateral muscles may also exist. These
ipsilateral influences are interesting because of their possible Potentiation of muscle responses to transcranial
role in recovery) after stroke. Ipsilateral EMG responses to magnetic stimulations induced by stimulation of
transcranial magnetic stimulation (TMS) have been the median nerve
reported in patients with unilateral damage to the motor
pathways (e.g. Carr et al. 1993), but there is still a debate as A.R. Jamshidi Fard and E.M. Sedgwick
to whether such responses can also be obtained in normal Department of Clinical Neurological Sciences, University of
subjects. Southampton, Southampton General Hospital,
We have investigated this question in fifteen normal Southampton S09 4XY
volunteers, aged 19-59 years, with ethical committee
approval. Surface EMG recordings were made from Motor responses evoked by trascranial magnetic
pectoralis major, deltoid, biceps and first dorsal interosseous stimulation (TMS) or transcranial electrical stimulation
(IDI). We used a MIagstim 200 stimulator (maximum output (TCS) can be facilitated by a prior conditioning stimulus to
15 Tesla) and a double cone (DC) coil (110 mm diameter), the an afferent nerve. Two facilitation periods are described;
junction region of which was located at a point 6 cm lateral short (SI), when the nerve stimulus is given near 0 to 10 ms
to vertex. We also used a small (50 mm diameter) butterfly after cranial stimulation, long (LI), when nerve stimulation
coil, positioned at the same location. Wihile subjects made is given 25-60 ms before the cranial one (Troni et al. 1988;
bilateral contractions of all sampled muscles, we used the Deletis et al. 1992).
DC coil at intensities of up to twice that of the threshold of These facilitation periods were examined in more detail
the response in the passive contralateral IDI. In most in ten normal consenting subjects. The study has ethical
subjects both hemispheres were investigated. committee approval. Focal cortical TMIS was applied by a
Nine subjects showed ipsilateral short-latency responses figure-of-S coil over the 'hot spot' for the hand muscles and
to DC stimulation. Of 108 averages (27 hemispheres x the strength adjusted to be just above twitch threshold for
4 muscles), 22 showed ipsilateral activation. Responses were the relaxed muscle. Conditioning electrical stimuli were
seen in deltoid (n = 1), pectoralis major (9), biceps (6) and applied to the median nerve at the wrist again at a strength
IDI (6). Ipsilateral responses were much smaller than in the just suprathreshold for a twitch in abductor pollicis brevis
contralateral muscle. muscle. The conditioning-test (C-T) interval was varied
Because of the size and position of the DC coil, 'ipsilateral' from -80 to +10 with respect to the magnetic stimulus and
responses could be due to activation of the opposite five magnetic stimuli were tried at each interval. The
hemisphere. unconditioned motor evoked potentials (MEPs) given is the
For the more proximal muscles this is unlikely since the mean + S.D. of twelve trials. The conditioned response was
Jo4rntal of Physiology (1994) 479.P 145P
considered significant if the mean amplitude of five trials (TA) were studied using low intensity stimulation (two
exceeded mean + 3 S.D. of the unconditioned responses. times perception threshold) of the sural nerve. The shocks
The results confirm the short facilitation period when the were given at sixteen phases of the step cycle in normal
C-T interval was -6 to +3 ms. Consideration of the timing subjects walking forward (FW) or backward (BW) on a
indicates that this must occur at spinal segmental level. The treadmill.
long period of facilitation lasted from 27-70 ms but in all In ST, facilitatory responses were present at end stance
subjects w-as divided into two periods (27-35 and 55-70 ms), and early swing during FW. In BW, however, these
separated by an interval of about 20 ms during which the responses are most prominent in the middle and at the end
test response fell to control levels. The maximum of swing. Suppressive responses occurred in the second half
facilitation observed was 734 % (620 5 + 126 9, mean + S.D.) of swing in FW, while during BW they were seen at end
for the short period and 386 % (251 + 97 9, mean + S.D.) stance. Hence, for both FW and BW a phase-dependent
during the long periods. reflex reversal occurred in ST during swing, but the sign of
The long late facilitations may be cortical as the earliest the reversal was different for the two conditions (from
facilitation began at 27 ms but the afferent volley reaches facilitation to suppression in FW but from suppression to
the sensory cortex at 20 ms. It is unlikely that this facilitation during BW). Moreover, the reversal point
facilitation occurs at spinal level due to summation of the occurred earlier in the swing phase during BW as compared
afferent volley with a volley in the small cortico-spinal (CS) to FW. In the other muscles, there was a 48-100 %
fibres. The near threshold magnetic stimuli are unlikely to reduction in the amplitude of the facilitatory responses
excite the small CS fibres and there are no reports of during late swing in FW and during early swing in BW.
magnetically induced volleys in these fibres. It is concluded that touchdown is not likely to be a main
The long interval facilitation consisting of two element involved in phase-dependent reflex reversals in
temporally separate processes implies separate cortical humans. A better explanation for the reversal is offered by
mechanisms creating a bimodal excitability cycle at the assuming that the FW reversal depends on a central motor
level of motor cortex. program. During BW this program runs in reverse,
resulting in a shift of the reversal point.
Supported by Mlucorn2 (BRA6615) and Nato 910574.
Deletis, V., Schild, J.H., Beric, A. & Dimitrijevic, AI.R. (1992).
Electroe neeph. Clin. Neurophysiol. 85, 302-310. REFERENCES
Ti oni, W., Cantello, R., DeMlattei, Al. & Bergamini, L. (1988). In Non-
Invasive Stimulation of Braini and Spinal Cord, ed. Rossini, P. & Duysens, J., Trippel, Al., Horstmann, G.A. & Dietz, VT. (1990). Exp.
Marsden, C.D., pp. 73-83. Alan R. Liss, Inc. Brain Res. 82, 351-358.
Yang, J.F. & Stein, R.B. (1990). J. Nleurophysiol. 63 (5), 1109-11170.
Phase-dependent modulation and reversal of sural
nerve induced reflexes during human backward Effects of plantar nerve stimulation on the
walking transmission of late flexion reflexes in the
decerebrate spinal cat
L. Murrer, A.A.M. Tax, V. Dietz* and J.E.J. Duysens
B.A. Conway*, D.T. Scott, J.S. Riddell and
Department of Mledical Physics & Biophysics, K. U. AI.R. Hadian
Nijmegen, The NVetherlands and *Department of Clinical
Neurology & Neurophysiology, University of Freiburg, *Bioengineering Unit, University of Strathclyde and
Freiburg i. Br., Germany Institute of Physiology, University of Glasgow, Glasgow
In man, the amplitude of reflex responses elicited by low In the acute spinal cat administration of L-DOPA results
intensity stimulation of cutaneous afferents from the foot in the depression of short latency flexion reflexes and the
depends on the phase of stimulation within the step cycle appearance of long latency, long-lasting flexion reflexes.
(Duysens et al. 1990; Yang & Stein, 1990). This 'phase- The organization of these late flexion reflexes is believed to
dependent modulation' is expressed in its most extreme reflect the organization of a spinal generator for locomotion
form in reflex reversals. In some leg muscles a reversal from (Lundberg, 1979). In previous studies on fictive locomotion
facilitatory to suppressive responses occurs at end swing in acute spinal cats, Conway et al. (1987) demonstrated that
during forward walking (FW). To test the idea that this stimulation of extensor lb afferents produces an extensor
reversal occurs in anticipation of touchdown, experiments resetting of the locomotor rhythm. This effect is
were done during backward walking (BW). If touchdown is accompanied by inhibition of the interneuronal network
an important factor, one would expect a similar reversal to responsible for the generation of ipsilateral late flexion
occur at end swing during BW as well. reflexes. Here the effects of low threshold plantar nerve
The phase-dependent modulation of medium-latency (P2) stimulation on the transmission of ipsilateral late flexion
responses in various leg muscles such as semitendinosus (ST), reflexes are reported.
biceps femoris (BF), rectus feinoris (RF) and tibialis anterior Cats were deeplyT anaesthetized (02/N20 and halothane)
146P Journal of Physiology (1994) 479.P
and decerebrated. Spinalization (T12) and lumbar
laminectomy were subsequently performed. Both hindlimbs
were denervated and selected nerves dissected for
stimulation or recording purposes. These included the
lateral, deep and medial branches of the plantar nerve. On
completion of surgery anaesthesia was discontinued and the
cats paralysed and ventilated (see Conway et al. 1987).
Following administration of nialamide (50 mg kg', i.v.) and
L-DOPA (70-100 mg kg', i.v.) late flexion reflexes, recorded
as electroneurograms, were evoked by trains of stimuli
(5-50 times threshold, T) applied to hindlimb nerves.
Continuous low threshold stimulation (20-50 Hz, 1-5-2 T)
of deep or lateral plantar nerves resulted in complete
suppression of late flexion reflex transmission. Short bursts
of stimuli delivered to these nerves during late flexor
reflexes abruptly terminated flexor activity. In contrast,
stimulation of the medial plantar nerve (< 2 T) tended to
potentiate flexor reflexes. These results demonstrate that
low threshold afferents from the plantar nerves interact
with the spinal circuits responsible for the generation of late
flexion reflexes. Together with observation from
spontaneously walking premammillary cats (Duysens, 1977)
the results suggest that sensory input from the plantar
aspect of the foot is important in the reflex control of spinal
stepping. Further work is required to determine the relative
contribution of low threshold muscle and cutaneous
afferents to the above observations.
Supported by the MRC.
Conway, B.A., Hultborn, H. & Kiehn, 0. (1987). Exp. Brain Res. 68,
Duysens, J. (1977). J. Neurophysiol. 40, 737-751.
Lundberg, A. (1979). Prog. Brain Res. 50,11-28.