Abstract We have analyzed the
behavior of neurons of the lateral reticular
nucleus (LRN) during fictive respiration and
locomotion and found that some LRN neurons have
both central respiratory and locomotor
rhythms.
Experiments were conducted on decerebrate,
decerebellate, immobilized, and artificially
ventilated cats, with the spinal cord transected
at the lower thoracic cord. Fictive respiration
and fictive forelimb locomotion were ascertained
by monitoring activities from the phrenic nerve
and forelimb extensor and flexor nerves,
respectively.
Fictive locomotion was evoked by electrical
stimulation of the mesencephalic locomotor
region (MLR) or sometimes occurred
spontaneously. During fictive locomotion many
LRN neurons fired in certain phases of the
locomotion cycle; i.e., with respect to the
nerve discharge of the ipsilateral forelimb they
fired in either the extensor, flexor,
extensor-flexor, or flexor-extensor phase.
Firing of some LRN neurons was modulated
synchronously with central respiratory rhythm.
Neurons with inspiratory activity and those with
expiratory activity were both found. More than
half of these respiration-related LRN neurons
had locomotor rhythm as well.
The majority of the three types of LRN
neurons, i.e., neurons with only locomotor
rhythm, those with only respiratory rhythm, and
those with both respiratory and locomotor
rhythms, were antidromically activated by
electrical stimulation of the ipsilateral
inferior cerebellar peduncle. Electrical
stimulation of the upper cervical cord showed
that these LRN neurons, not only
locomotion-related but also respiration-related
neurons, received short latency inputs from the
spinal cord.
The LRN neurons studied were distributed
widely in the LRN, relatively densely in the
caudal two-thirds of the nucleus. No particular
differences were detected between the three
types of LRN neurons with respect to their
location in the nucleus. These results indicate
that the information about central respiratory
and locomotor rhythms that is necessary for
cerebellar control of the coordination between
respiration and locomotion converges, at least
partly, at the level of the LRN.
Introduction : The lateral reticular
nucleus (LRN). a major source of mossy fibers to
the cerebellum, reccives inputs from a large
nurnber of supraspinal motor centers and
intervenes in the spino-reticulo-cerehellar
pathway (SRCP), which conveys information
concerning the activity of spinal motor centers.
It is a large nucleus lying in the ventrolateral
part of the lower medulla and adjacent to the
ventral respiratory group (VRG), which plays a
major role in respiraton control.
During the course of our stddy on
respiratory neurons of the VRG we often
recorded, largely by chance, from LRN neurons:
their situation in the LRN was noticed after the
experiments by histological examination. These
LRN neurons often showed respiratory or
locomotor rhythm and sometimes the two rhythms.
Since there are few reports on their firing
properties during respiratory and locomotor
movements, we aimed to study the behavior of LRN
neurons in more detail during these movements.
In particular, there are no published reports
that observed single LRN neurons, having both
respiratory and locomotor rhythms.
At present, several key nuclei that
participate in respiratory control have been
identified in the medulla: e.g., the dorsal
respiratory groups (DRG), the VRG, and the
Botzinger complex. Before identification of
these nuclei, early studies surveyed the medulla
and pons searching for neurons that fired in
synchrony with the respiratory rhythm. Some of
such studies have found neurons with respiratory
activity in the LRN. However, not much attention
has been paid to these findings, presumably
bccause the role of the cerebellum in
respiration is vague and the respiration-related
LRN neurons, if any, are almost certainly not
involved in respiratory rhythm generation. In
consequence, at present, the firing properties
and distribution of the respiration-related LRN
neurons are not well known.
On the other hand, more studies have been
done on LRN neurons in relation to limb
movements. In particular, the behavior of LRN
neurons during scratching movements in the
hindlimb has been studied extensively. The LRN
neurons showed distinct modulation of their
discharge during scratching, most of them firing
in the extensor (Ex) phase of the scratching
cycle. Although the LRN neUrons are known to
fire in synchrony with forelimb locomotion as
well. the information about their behavior
during locomotion is limited compared with that
during scratching.
This study concerned with the LRN neurons
with respiratory and/or locomotor rhythms,
describes: ( 1 ) their firing, patterns during
respiration and locomotion. (2) their
distribution, (3) their projections to the
cerebellar peduncle and (4) inputs from the
spinal cord. [...]
Discussion
This study has extended the previous
investigations of the activition of LRN neurons
during fictive respiration and during fictive
locomotion. We examined the firing properties of
LRN neurons during respiration and locomotion in
more detail and in addition we studied the
firing of LRN neurons when respiration and
locomotion occurred simultanously. In
particular, this is the first report finding
single LRN neurons that have both respiratory
and locomotor rhythms. These findings will
provide a new insight into the function of LRN
neurons and the cerebellar control of
respiration and locomotion.
Distribution and sampling, of the LRN
neurons
The LRN is divided into three parts: a
small-celled parvocellular, a large-celled
magnocellular and a rostral subtrigeminal part.
The parvocellular and magnocellular parts occupy
the ventrolateral and dorsolateral regions of
the nucleus, respectively and comprise the major
portion of the LRN. The rostral and lateral part
of the major portion gradually becomes the
subtrigeminal part which is located immediately
ventral to the spinal trigeminal tract. The
rostral and medial border of the major portion
of the LRN is unclear, with a gradual decrease
in cell density. Moreover. the border between
the parvo- and magnocellular parts can only be
vaguely delimited.
The present Study intended to sample LRN
neurons as widely as possible from the major
portion of the LRN. The present LRN neurons were
located in both the parvocellular and
magnocellular parts: the neurons found in the
close vicinity of the ventrolateral border of
the medulla belong to the parvocellular part.
However, no neurons were sampled from the
lateral part of the nucleus. This could be due
to sampling bias, because (1) the sampling from
the lateral edge of the brainstem is technically
difficult and (2) the lateral part corresponds
largely to the parvocellular part consisting of
small neurons. It is also possible that the
lateral part of the LRN receives afferents from
the spinal cord below L3 and such inputs were
removed in the present study.
With the reservation mentioned above, the LRN
neurons sampled were distributed throughout the
rostrocaudal extent of the major portion. This
is consistent with the previous results: LRN
neurons modulated with forelinib locomotion were
recorded throughout the LRN. In addition, the
present study shows that LRN neurons related to
forelimb locomotion are distributed more densely
in the caudal region of the major portion of the
nucleus, largely caudal to the obex. This is in
contrast to the distribution of LRN neurons that
respond to fictive scratching in the hindlimb:
such LRN neurons are located preferentially in
the rostral area of the LRN, rostral to the
obex. Therefore. it is possible that the LRN
neurons related to hindlimb movements are
located more rostrally and those related to
forelimb movements more caudally.
One hundred and five of the 122 LRN neurons
tested were antidromically activated from the
inferior cerebellar peduncle. Antidromic
activation of some other LRN neurons was
uncertain because of the follo\wing when
spontaneous locomotion was present we sampled
locomotion-related neurons first and then the
cerebellar peduncle was stimulated to examine
their antidromic responses. However, the
stimulation of the cerebellar peduncle almost
always evoked synchronous discharges of a large
number of LRN neurons. Since LRN neurons in
general had spike activity of large amplitude
easily exceeding 1 mV the spikes of the isolated
neuron were often obscured. In such cases, it
was difficult to confirm antidromic activation
even with a collision test. Nevertheless, the
present results suggest that almost all the LRN
neurons with respiratory and/or locomotor
rhythms project to thee cerebellum considering,
that the stimulating electrode could activate
not all axons that passed the cerebellar
peduncle. On thee other hand, when spontaneous
locomotion was absent, we first idenfied LRN
neurons by stimulating the cerebellar pedoncle
and then locomotion was evoked by MLR
stimulalion. This often activated previously
quiet neurons and the spikes of the isolated
neurons were contaminated by large spikes of
newly recruited neurons. Then. the initially
isolated neurons could not be characterized
during locomotion and had to be discarded. With
these experiences we tended to discard, from the
begining neurons with small spikes: thus the
present sampling was biased toward recording
from neurons with larger spikes.
Behavior of LRN neurons during
locomotion
Arshavsky and colleagues ( 1986a) have made
extensive study of neurons in the spinal cord
and brainstem that are related to limb
movements. In particular, they have analyzed the
neural mechanisms of fictive scratching that
involves only one hindlimb. They have clarified
many features of the input-output organization
of the cerebellum as weil as the organization of
thee spinal motor centers during scratching and
locomotion. They studied LRN neurons because the
LRN is the relay nucleus of the SRCR one of the
major afferent systems ofthe cerebellum. LRN
neurons show distinct modulation of their
discharge in synchrony with scratching and
loconiotion rhythms similarly to neurons oft he
other spinocerebellar tracts.
During hindlimb scratching, most LRN neurons
fire in the vicinity of the Ex phase of the
scratching cycle. This firing of LRN neurons
contrasts with that of ventral spinocerebellar
tract (VSCT) neurons. whicli fire largely in the
FI phase. Thus. it is stugested thal the two
pathways to the cerchellm, the SRCP and the
VSCT, convey messages about procceses occurring
in different phases of the scratching cycle.
Although the behavior of LRN neurons during
forcelimb locomotion has been studied by using
preparations similar to the present study, it
has not been reported in which phase of the
locomotion cycle LRN neurons fire. The present
results show that there are many LRN neurons
that fire in the FI phase as well in the FI-Ex
and Ex-FI phases, although LRN neurons .. in fre
in the Ex phase are the most abundant. A direct
comparison between the present results in
forlimb locomotion and the previous results in
hindlimb scratching is difficult, because: ( 1 )
scratching involves only one limb and the cycle
consists of short EX phase and long FI phase in
contrast to the locomotion cycle; and (2) the
locomotion was sudied in the forelimb and
hindlimb movements. It is possiblethat the LRN
neurons linked to hindlimb movements did not
respond to forelimb locomotion and thus have not
been studied in the present study.
The present as well as previous results
obtained in immobilized animals show that the
rhythmic activity of LRN neurons is of central
but not peripheral origin. During hindlimn
scratching this central drive originates in the
motor center of the lumber cord and influences
of supraspinal motor centers on LRN neurons are
of minor importance. In the case of forelimb
locomotion, at least one source of rhythmic
input from the spinal cord has been identified
Arshavsky showed thal virtually all (26 of 27)
of the C3-4 propriospinal neurons (C3-4 PNs)
that project to the LRN had locomotor rythm. The
central drive to C3+4 PNs originatesin the
cervical cord, and influences of supra spinal
motors centers are of minor importance. However,
it has not been reported in which phase of the
locomotion cycle the C3+C4 PNs fire. Other
neurons or tracts that convey information about
forelimb locomotion to the LRN are not known and
remain to be studied.
Behavior of LRN neurons during
respiration
Gesell et al. (1936) were the first to record
from single respiratory neurons in the brain
making extensive tracking by fine needle
electrodes. They recorded from respiratory
neurons in the areas around the solitary tract
and the nucleus ambiguus in the dog which at
present are known as the DRG and the VRG,
respectively. In this pioneering study, they
recorded from both inspiratoy and expiratory
neurons in the ventrolateral medulla and
localized these neurons in the LRN. Vibert et
al. ( 1976). who mapped respiratory neurons in
the pons and medulla in the cat, also reported
the existence of respiratory neurons in the LRN,
although they did not describe the type of
respiratory neurons recorded from.
The respiralory neurons recorded in the
present study are presumably from the same
population of respiratory neurons reported in
the previous studies. We have added more
information about their firing patterns and
locations in the LRN. The respiratory rhythm in
these LRN neurons is of central origin, as is
the locomotor rhythm. The respiration related
LRN neurons fired in the absence of rhythmical
sensory inputs, which was shown by stopping the
ventilator in the immobilized animal. This
characteristic is similar to that of respiratory
neurons of the medullary respiratory center such
as the DRG VRG. and Botzinger complex. It
remains to be elucidated whether the respiratory
inputs are conveyed to the LRN directly from the
medullary respiratory center, or indirectly via
the spinal cord, or from both. There are
suggesyions that some DRG neurons and
swallowing-related respiratory neurons in the
medulla have axon collaterars in the LRN.
Antidromic mapping by electrical stimulation of
decrementing inspiratory neurons and
decrementing expiratory neurons of the Botzinger
complexalso suggests their projections to the
LRN in a few cases.
Inputs from the spinal cord
The LRN receives information from a number of
ascending tracts from the spinal cord as well as
from supraspinal motor centers; We stimulated
the ventral part of the C3/4 spinal cord to
examine whetheir or not the LRN neurons with
respiratory and/or locomotor rhythms received
inputs from the spinal cord. The stimulation
could activate three well-studied tracts that
ascend the lateralfuniculus: there are the
bilateral ventral flexor reflex tract (bVFRT),
the ipsilateral fore-limb tract (iFT) and
ascending branches or somata of the C3-4 PNs.
The short-latency responses in more than 70% of
LRN neurons from the ipsilateral C3/4 cord are
consistent with the resuslts that the LRN
neurons receive monosynaptic inputs from these
tracts.
The latencies shorter than 2 ms in fig 10
suggest monosynaptic activation from the bVFRT
and iFT. Since the conduction velocities of the
ascending branches of the C3-4 PNs are about
twice as slow as those of the bVFRT and iFT, the
latencies shorter than 4 ms in Fig. 10 may also
be within the monosynaptic range.
A large portion of the monosynaptic
connections from both the bVFRT and the iFT to
LRN neurons are inhibitory. Although the present
extracellular study hardly detected such
inhibitory connections, it is highly probable
that many of the present LRN neurons had
monosynaptic inhibitory inputs from the spinal
cord. The above argument holds only for the
ipsilateral stimulation. It is difficult to
compare the present contralateral stimulation
with that of the previously studies.
In the previously studies, in all cases. the
spinal cord contralateral to the LRN was
transected at C2-3. The contralateral
stimulation in such preparations activated
descending pathway,. in particular the
vestibulospinal tract which excites
monosynaptically the bVFRT neurons terminating
in the contralateral LRN. In contrast, the
present experimental situation was too
complicated, activating both descending and
ascending tracts. Nevertheless, the
short-latency responses from the contralateral
spinal cord suggest the existence of
monosynaptic ascending connections, which have
not been specified yet.
Respiration. locomotion, and
cerebellum
The cerebellum receives at least two sorts of
information from the spinal cord during limb
movements. One is the information from
peripheral sensory apparatus and it is used to
monitor on-going movements. The other is the
information about the activity of spinal motor
centers. During hindlimb movements, the former
is carried by the dorsal spinocerebellar tract
(DSCT) and the latter by the VSCT. During
forelimb movements, respective information is
carried by the cuneocerebellar tract (CCT) and
the rostral spinocerebellar tract (RSCT). In
addition to these direct spinocerebellar tracts,
the indirect spinocerebellar pathway through the
LRN, i.e.. the SRCP exists. The SRCP carries
information about the activity of spinal motor
centers rather than the activity of peripheral
sensory apparatus.
In this respect, the respiratory system is
organizedsimilarly to the locomotor or
scratching system. Hirai and colleagues found
two spinocerebellar tracts that convey rhythmic
respiratory information. Some DSCT neurons
located in Clark's column were shown to have
respiratory rhythm and ascend ipsilaterally.
These DSCT neurons receive peripheral inputs
that respond to movernents of the chest wall. A
group of neurons located in laminae VII and VIII
of the thoracic spinal cord were found to have
respiratory rhythm and to ascend in the
contralateral spirial cord. In contrast with the
DSCT neurons, the firing of these neurons is
modulated by central respiratory driv, some of
them receiving additional peripheral inputs. As
shown in the present study, the path through the
LRN also carries respiratory information to the
cerebellum. In addition to these inputs, the
pathway through the inferior olivary nucleus
functions in the respiratory system as well as
in the limb movement system.
Respiration-related neurons are recorded from
both cerebellar cortex and nuclei. Involvement
of the cerebellum in control of respiration is
evident from stimulation and lesion studies.
However. the role of the cerebellum in
respiration has not been established. It is
suggested that the cerebellum functions not only
in postural reflexes but also in respiratory,
cardiovascular, and other autonomic reflexes. In
particular, respiratory and locomotor inputs to
the cerebellum may function in homeostatic
adjustments of respiratory movements during
locomotion, and this may he revealed as
interactions between central rhythm generators
of respiration and locomotion.
The present study has shown that respiratory
and locomotor inputs to the cerebellum converge,
at least partly. on the LRN neurons.
Furthermore, LRN neurons are known to have
conjoint inputs, from macular and neck
receptors, and may function in control of neck,
limbs and extrinsic eye musculature. Since a
large nimber (more then 50%) of the LRN neurons
studied have convergent neck and macular inputs,
it is quite possible that other inputs, e.g.,
respiratory and locomotor inputs, also converge
on these LRN neurons. If this is this case,
single LRN neurons relay a large number of
inputs of various modalities to the
cerebellum.
Is yawning a
brainstem phenomenon ? Wimalaratana HS,
Capildeo R. A stroke patient who stretched his
hemiplegic arm during yawning
Lancet 1988; 1; 8580;
300
