Revisão de anatomia e correlações clínicas do tronco encefálico, parte i: bulbo / Review of brain anatomy and clinical correlations, part i: medulla

O Tronco encefálico (TE) é uma estrutura singular do sistema nervoso central, pois nele passam tratos sensoriais ascendentes da medula espinal, tratos sensoriais da cabeça e do pescoço, os tratos descendentes motores originados no prosencéfalo (divisão mais rostral do encéfalo), e as vias ligadas aos centros de movimento dos olhos. Contém ainda os núcleos dos nervos cranianos e está envolvido na regulação do nível de consciência através de projeções ao prosencéfalo oriundas da formação reticular. Todas essas estruturas coexistem em um espaço muito exíguo, o que faz com que o TE seja um local muito sensível às alterações patológicas, sendo que os pacientes apresentam muitos sinais neurológicos mesmo com lesões muito pequenas nesse local. Compreender a anatomia interna do TE é essencial para o diagnóstico neurológico e a prática da medicina clínica. Outros profissionais da saúde também se beneficiam desse conhecimento para melhor manejo dos seus pacientes neurológicos. Essa revisão apresenta detalhes da anatomia macroscópica e microscópica do bulbo, bem como seus correlatos clínicos frente às lesões mais comuns dessa divisão particular do TE, conhecidas como síndromes bulbares.

The brainstem is a unique structure in the central nervous system, since it gives way to ascending sensory tracts from the spinal cord, sensory tracts from the head and neck, motor descending tracts originating from the forebrain, and the pathways connected to the eye movement centers. It also contains the cranial nerve nuclei and is involved in the regulation of consciousness levels through projections to the forebrain originating in the reticular formation. All these structures coexist in a very small space, which makes the brainstem very sensitive to pathological changes, with patients presenting several neurological symptoms even with very small brainstem lesions. Understanding the internal anatomy of the brainstem is essential for neurological diagnosis and the practice of clinical medicine. Other health professionals also benefit from this knowledge to better manage their neurological patients. This review presents detailed information on the macroscopic and microscopic anatomy of the medulla, as well as its clinical correlates in the face of the most common lesions of this particular division of the brainstem, known as medullary syndromes.

Parallel pathways from motor and somatosensory cortex for controlling whisker movements in mice.

Mice can gather tactile sensory information by actively moving their whiskers to palpate objects in their immediate surroundings. Whisker sensory perception therefore requires integration of sensory and motor information, which occurs prominently in the neocortex. The signalling pathways from the neocortex for controlling whisker movements are currently poorly understood in mice. Here, we delineate two pathways, one originating from primary whisker somatosensory cortex (wS1) and the other from whisker motor cortex (wM1), that control qualitatively distinct movements of contralateral whiskers. Optogenetic stimulation of wS1 drove retraction of contralateral whiskers while stimulation of wM1 drove rhythmic whisker protraction. To map brainstem pathways connecting these cortical areas to whisker motor neurons, we used a combination of anterograde tracing using adenoassociated virus injected into neocortex and retrograde tracing using monosynaptic rabies virus injected into whisker muscles. Our data are consistent with wS1 driving whisker retraction by exciting glutamatergic premotor neurons in the rostral spinal trigeminal interpolaris nucleus, which in turn activate the motor neurons innervating the extrinsic retractor muscle nasolabialis. The rhythmic whisker protraction evoked by wM1 stimulation might be driven by excitation of excitatory and inhibitory premotor neurons in the brainstem reticular formation innervating both intrinsic and extrinsic muscles. Our data therefore begin to unravel the neuronal circuits linking the neocortex to whisker motor neurons.

Dental afferents project onto gustatory neurons in the nucleus of the solitary tract.

The aim of this study was to investigate the inferior alveolar nerve (IAN) and chorda tympani (CT) projections onto gustatory neurons of the nucleus of the solitary tract (NST) in the rat by immunochemical and electrophysiological techniques. IAN afferents were retrogradely labeled. NST neurons were labeled either by retrograde tracer injection into the parabrachial nucleus (PBN) or by c-Fos mapping after CT activation. NST neurons responding to tastant stimulation were recorded in vivo before and after electrical stimulation of the IAN. Results from the immunolabeling approach showed IAN boutons "en passant" apposed to retrogradely labeled neurons from PBN and to CT-activated neurons in the NST. Recordings of single NST neurons showed that the electrical stimulation of the IAN significantly decreased CT gustatory responses. Analysis of these data provides an anatomical and physiological basis to support trigeminal dental and gustatory interactions within the brainstem.

Role of reactive oxygen species in brainstem in neural mechanisms of hypertension.

The involvement of reactive oxygen species such as superoxide is implicated in the pathogenesis of hypertension. The brain contains a high concentration of polyunsaturated fatty acids in its cell membranes. These fatty acids are targets of oxygen-derived free radicals. Thiobarbituric acid-reactive substances (TBARS), an indirect marker of oxidative stress, are increased in the brainstem of stroke-prone spontaneously hypertensive rats (SHRSP) compared with those of Wistar-Kyoto rats (WKY). In addition, the intensity of electron spin resonance signals taken from the rostral ventrolateral medulla (RVLM), a cardiovascular center, decreases more rapidly in SHRSP than in WKY. To confirm the role of reactive oxygen species in the RVLM or the nucleus tractus solitarius (NTS) in SHRSP, we transfected adenovirus vectors encoding the manganese superoxide dismutase (MnSOD) gene (AdMnSOD) or Cu/Zn-SOD gene (AdCu/ZnSOD) bilaterally into the RVLM or the NTS. After the gene transfer, blood pressure and heart rate of SHRSP, monitored by radio-telemetry system, were significantly decreased compared with non-treated SHRSP, but not WKY. Urinary norepinephrine excretion was significantly decreased in AdMnSOD- or AdCu/ZnSOD-transfected SHRSP, but not in WKY. Furthermore, we found that activation of NAD(P)H oxidase via Rac1 is a source of reactive oxygen species generation in the brain of hypertensive rats. Taken together, these results suggest that the increased oxidative stress in the RVLM and the NTS contribute to the central nervous system mechanisms underlying hypertension in SHRSP. We also found that atorvastatin has actions of reducing oxidative stress in the brain associated with sympatho-inhibitory effects.

Connectivity of an effective hypothalamic surgical target for cluster headache.

The purpose of this study was to look at the connectivity of the posterior inferior hypothalamus in a patient implanted with a deep brain stimulating electrode using probabilistic tractography in conjunction with postoperative MRI scans. In a patient with chronic cluster headache we implanted a deep brain stimulating electrode into the ipsilateral postero-medial hypothalamus to successfully control his pain. To explore the connectivity, we used the surgical target from the postoperative MRI scan as a seed for probabilistic tractography, which was then linked to diffusion weighted imaging data acquired in a group of healthy control subjects. We found highly consistent connections with the reticular nucleus and cerebellum. In some subjects, connections were also seen with the parietal cortices, and the inferior medial frontal gyrus. Our results illustrate important anatomical connections that may explain the functional changes associated with cluster headaches and elucidate possible mechanisms responsible for triggering attacks.

Chemical stimulation of visceral afferents activates medullary neurones projecting to the central amygdala and periaqueductal grey.

Cholecystokinin (CCK) stimulates gastrointestinal vagal afferent neurones that signal visceral sensations. We wished to determine whether neurones of the nucleus of the solitary tract (NTS) or ventrolateral medulla (VLM) convey visceral afferent information to the central nucleus of the amygdala (CeA) or periaqueductal grey region (PAG), structures that play a key role in adaptive autonomic responses triggered by stress or fear. Male Sprague-Dawley rats received a unilateral microinjection of the tracer cholera toxin subunit B (CTB, 1%) into the CeA or PAG followed, 7 days later, by an injection of CCK (100 microg/kg, i.p.) or saline. Brains were processed for detection of Fos protein (Fos-IR) and CTB. CCK induced increased expression of Fos-IR in the NTS and the VLM, relative to control. When CTB was injected into the CeA, CTB-immunoreactive (CTB-IR) neurones were more numerous in the rostral NTS ipsilateral to the injection site, whereas they were homogeneously distributed throughout the VLM. Double-labelled neurones (Fos-IR+CTB-IR) were most numerous in the ipsilateral NTS and caudal VLM. The NTS contained the higher percentage of CTB-IR neurones activated by CCK. When CTB was injected into the PAG, CTB-IR neurones were more numerous in the ipsilateral NTS whereas they were distributed relatively evenly bilaterally in the rostral VLM. Double-labelled neurones were not differentially distributed along the rostrocaudal axis of the NTS but were more numerous in this structure when compared with the VLM. NTS and VLM neurones may convey visceral afferent information to the CeA and the PAG.

Patterns of fos expression in the rostral medulla and caudal pons evoked by noxious craniovascular stimulation and periaqueductal gray stimulation in the cat.

Functional imaging studies and clinical evidence suggest that structures in the brainstem contribute to migraine pathophysiology with a strong association between the brainstem areas, such as periaqueductal gray (PAG), and the headache phase of migraine. Stimulation of the superior sagittal sinus (SSS) in humans evokes head pain. Second-order neurons in the trigeminal nucleus that are activated by SSS stimulation can be inhibited by PAG stimulation. The present study was undertaken to identify pontine and medullary structures that respond to noxious stimulation of the superior sagittal sinus or to ventrolateral PAG stimulation. The distribution of neurons expressing the protein product (fos) of the c-fos immediate early gene were examined in the rostral medulla and caudal pons of the cat after (i) sham, (ii) stimulation of the superior sagittal sinus, (iii) stimulation of the superior sagittal sinus with PAG stimulation, or (iv) stimulation of the PAG alone. The structures examined for fos were the trigeminal nucleus, infratrigeminal nucleus, reticular nuclei, nucleus raphe magnus, pontine blink premotor area, and superior salivatory nucleus. Compared with all other interventions, fos expression was significantly greater in the trigeminal nucleus and superior salivatory nucleus after SSS stimulation. After PAG with SSS stimulation, on the side ipsilateral to the site of PAG stimulation, fos was significantly greater in the nucleus raphe magnus. These structures are likely to be involved in the neurobiology of migraine.

Activation by cutaneous or visceral noxious stimulation of spinal neurons projecting to the medullary dorsal reticular nucleus in the rat: a c-fos study.

The involvement of spinal neurons in the transmission of cutaneous and visceral nociceptive input to the medullary dorsal reticular nucleus was studied. Rats were injected with cholera toxin subunit B in the left dorsal reticular nucleus and subjected 4 days later to noxious mechanical, thermal or chemical stimulation of the proximal internal aspect of the left thigh, or to chemical stimulation of the urinary bladder. Sections of spinal segments T13-L3 were processed immunocytochemically for cholera toxin subunit B and Fos protein. The percentage of double-labelled cells in the population of Fos-positive cells was higher in lamina I (1-4%) than in deeper laminae (0-0.7%) following all stimuli. The percentage of double-labelled cells in the population of retrogradely labelled cells was 30-53% in lamina I and 0-5% in laminae III-X. Visceral stimulation activated more retrogradely labelled lamina I cells than any kind of cutaneous stimulation. Pyramidal cells were activated in higher numbers than multipolar and flattened cells after thermal cutaneous or visceral stimulation, and in lower numbers than multipolar cells after mechanical stimulation. These results suggest that, in the experimental conditions used, spinal cord cells conveying noxious input to the dorsal reticular nucleus are concentrated in lamina I. They further indicate that the spinal-dorsal reticular nucleus pathway plays a major role in the transmission of nociceptive visceral input, and point to the preferential involvement of pyramidal cells in cutaneous thermal and visceral processing.

Different roles of alpha 2-adrenoceptors of the medulla versus the spinal cord in modulation of mustard oil-induced central hyperalgesia in rats.

We attempted to determine the roles of spinal versus medullary alpha 2-adrenoceptors in modulation of central hyperalgesia in rats. Central hyperalgesia was produced by applying mustard oil (50%) to the skin of the ankle of one hindpaw. The threshold for eliciting a hindlimb flexion reflex was determined by applying a series of calibrated monofilaments to the glabrous skin of the hindpaw contralaterally (= control) or ipsilaterally to the mustard oil-treated ankle (= outside the area of primary hyperalgesia). Medetomidine (an alpha 2-adrenoceptor agonist; 1 micrograms), atipamezole (an alpha 2-adrenoceptor antagonist; 2.5 micrograms) or saline was microinjected into the lateral reticular nucleus of the medulla, the nucleus raphe magnus, or intrathecally to the lumbar spinal cord 12 min before the mustard oil treatment. Following saline injections, mustard oil produced a significant decrease of the hindlimb withdrawal threshold in the mustard oil-treated limb but not in the contralateral limb. Atipamezole in the lateral reticular nucleus produced a complete reversal of the hyperalgesia but no effect on the threshold of the intact limb. However, atipamezole in the raphe magnus nucleus or in the lumbar spinal cord did not produce a significant attenuation of the hyperalgesia. Medetomidine in the spinal cord, but not in the lateral reticular nucleus, reversed the hyperalgesia. At this dose range (up to 3 micrograms), medetomidine in the spinal cord of nonhyperalgesic control rats did not produce any significant change in the withdrawal response of hindlimbs or in the tail-flick latency. The results indicate that neurogenic inflammation induces significant plastic changes in the function of alpha 2-adrenergic pain regulatory mechanisms. In rats with mustard oil-induced central hyperalgesia, an alpha 2-adrenoceptor antagonist produces an antihyperalgesic effect due to an action on the caudal ventrolateral medulla, whereas an alpha 2-adrenoceptor agonist produces an enhanced antinociceptive effect due to a direct action on the spinal cord.

Beating the competition: the reliability hypothesis for Mauthner axon size.

The Mauthner cell has an axon that is among the largest in diameter of any vertebrate neuron. It is commonly thought that the large size is needed for short latency escape responses involving a major contraction of the trunk musculature. Previous work, however, has shown that there is nothing unique about the strength of the Mauthner initiated response, compared to responses initiated by other smaller cells, and it is debatable that there is any important improvement in response latency due to Mauthner axon size. In this paper we advance an alternative explanation: although the Mauthner cell has a powerful excitatory influence on motoneurons, the large size of the Mauthner axon is most important in rapidly spreading an inhibitory signal that turns off other competing motor commands. Such competing commands are likely to arise in the presence of ongoing swimming behavior or ambiguous stimuli that could activate a fast turn either toward or away from the stimulus. These stimuli include apparent food items, or lures, presented by predators (such as anglerfish) and escape eliciting sounds which, in the presence of background noise, may have 180 degrees directional ambiguity. Thus, large size of the axon contributes most to the reliable expression of the escape behavior. We base this reliability hypothesis on a retrospective analysis of previous neurophysiological data and new anatomical measurements of the diameters of the large spinal cord axons from which we calculated conduction velocities. Our calculations show that the Mauthner-derived inhibition is fast enough that it allows an escape response to occur even when a conflicting motor command enters the spinal cord at the same time as the Mauthner axon impulse. The rapid spread of inhibitory influence, along with excitation, may be a general feature of motor system cells with large axonal diameters.

Distributions and colocalization of neuropeptide Y and somatostatin in the goldfish brain.

The distributions of single- and double-labelled neuropeptide Y- (NPY-) and somatostatin-immunoreactive (SOM-IR) perikarya and processes were determined in the goldfish brain using immunoperoxidase and immunofluorescence techniques, respectively. In double-labelled material, it was evident that although these two peptides showed markedly similar distributions, they were colocalized in very few instances. A high degree of colocalization of NPY and SOM was noted in the neurons of the ventrolateral telencephalon (VI), the entopenduncular nucleus (NE) and, to a lesser extent, in the dorsocentral nucleus of the telencephalon (Dc). In Vl and NE, neurons showing NPY-IR displayed SOM-IR and vice versa. The only other instance of colocalization was that noted in the brainstem, where SOM and NPY were colocalized in the large cell bodies of the medial column of the vagal motor complex. Single-labelled SOM- and NPY-IR neurons shared a very similar distribution in various nuclei in the diencephalon and in the optic tectum. Colocalization was also noted within fibers throughout many nuclei of the telencephalon and within fibers innervating the swim bladder, one of the peripheral organs to which neurons of the medial column of the vagal motor complex project. Processes in the torus semicircularis and vagal lobe showed single-labelled immunoreactivity for both SOM and NPY in distinct laminar patterns. Large single-labelled SOM-IR terminals appeared to form pericellular baskets in the eminentia granularis of the cerebellum. Single-labelled NPY- or SOM-IR fibers were also found in the secondary gustatory nucleus and tract, the facial lobe, descending trigeminal tract, reticular formation and spinal cord. As in mammalian species, select groups of neurons in teleosts colocalize the neuropeptides SOM and NPY.

Parabrachial nuclear complex. A comparative study of its cytoarchitectonics in birds and some mammals including man.

The cytoarchitectonic structure of the ncl. parabrachialis of 20 bird species, 17 mammals and man was studied in sections stained with cresyl violet. This comparative study showed the six cytoarchitectonic subnuclei of the ncl. parabrachialis (subncl. medialis, dorsomedialis, dorsalis, lateralis, dorsolateralis and ventralis) to be unevenly developed in the vertebrates under study. The degree of development of the subnuclei was tabulated to bring out the developmental trends of the ncl. parabrachialis in different groups of vertebrates. Only three of the six subnuclei were found enlarged in higher primates and in man--the subncl. dorsolateralis, lateralis and medialis, the latter two (subncl. lateralis and medialis) being also larger in insectivores. The most primitive structure is that of the subncl. ventralis with features very much like those of a small-cell reticular formation. Cytoarchitectonically speaking, the parabrachial subnuclei appear to be of polyphyletic origin: the medial and dorsomedial subnuclei being close to the substantia grisea centralis, the ventral subnucleus to the small-cell lateral reticular formation, and the lateral and dorsolateral subnuclei to the ncl. pedunculo-pontinus.

Double-labelled nigral compacta and reticulata cells from injections in the reticular formation and in the striatum. An experimental study using retrograde double labelling with HRP and iron-dextran in the rat.

14 rats were studied for the nigro-reticular projection. All animals had a relatively large quantity of dextran-fer injected in the middle of the striatum (to label the nigro-striatic cells and to facilitate distinction between the compacta and reticulata zones). 5 days later, HRP was injected in the reticular formation of the ponto-mesencephalic junction (in the ncl. pedunculo-pontinus, pontis oralis et caudalis, and in the caudal parts of the ncl. cuneiformis). Retrogradely labelled nigro-striatic, nigro-reticular and double-labelled nigrostriato-reticular cells were found. In the zona compacta substantiae nigrae, about 1/3 of the total number of labelled cells were identified as nigro-striatic, 1/3 as nigro-reticular, and 1/3 as collateralizing to the striatum and the reticular formation. In the zona reticulata substantiae nigrae, 2/3 of the total number of labelled cells constituted the nigro-reticular and 1/6 the nigro-striatic projections, while 1/6 collateralized to the striatum and the reticular formation. Retrograde HRP transport in the non-collateralized nigro-reticular and in the collateralized nigrostriato-reticular projections originated mainly from the ncl. pedunculo-pontinus and ncl. pontis oralis. Retrograde transport from the ncl. pontis caudalis was much less prominent. The numerous connections projecting from the zona compacta to the reticular formation point to the essentially dopaminergic nature of this projection. Collaterals of the nigro-striatic projection directed towards the reticular formation have not yet been described in the literature available to the author.

Gudden's tegmental nuclei and their connections to the hypothalamus and the reticular formation. I. An experimental study using retrograde labelling with HRP or iron-dextran in the rat.

31 rats were studied for the retrograde transport of HRP or dextran-fer from the corpus mamillare, anterior hypothalamus or reticular formation to the cells of the ventral (VG) and dorsal (DG) Gudden's tegmental nuclei. Judging by the position of the projection cells the VG was found divisible into the peripheral and central parts, a finding out of line with the conventional cytoarchitectonic division. The two parts differ in part of their projections. While both project jointly to the medial mamillary body, the peripheral part has a separate projection to the whole lateral hypothalamus whereas the central part projects also to the RF (ncl. pedunculopontinus and ncl. pontis oralis). The group of cells localized close to the cranial field of the VG, cytoarchitectonically referred to as the pars compacta ncl. centralis superioris, has the same projections as the external parts of the VG and can, therefore, be seen as a component of the VG peripheral part like the cells over the caudal field of the VG (pars intrafascucularis localized already among FLM fibres). The central part is made up of cells localized in the middle portions of the bulk of the VG cells. The dorsal Gudden's tegmental nucleus projects its ventral part to the lateral mamillary body, to the medial and lateral hypothalamus and to the RF (ncl. pontis oralis). The dorsal part of the DG projects solely to the RF (ncl. pontis oralis).

Gudden's tegmental nuclei and their connections to the hypothalamus and the reticular formation. II. An experimental study using retrograde double labelling with HRP and iron-dextran in the rat.

12 rats in double experiments were studied for the simultaneous retrograde transport of HRP and Dextran-fer from injections in the mamillary body and in the reticular formation or in selected diencephalic structures into the cells of Gudden's tegmental nuclei. The results of this particular study corroborated the findings of retrograde transport in single experiments (see Part I) showing all the cells of ventral Gudden's nucleus as projecting into the mamillary body, the cells of its peripheral parts as projecting also via a collateral to the lateral hypothalamus, and the cells of its central parts as projecting, in addition, via a collateral to the reticular formation (ncl. pedunculo-pontinus and ncl. pontis oralis). Cells of the ventral part of the ncl. dorsalis tegmenti Guddeni project into the lateral mamillary body and some of then also through a collateral into the whole of the hypothalamus. Part of its cells project independently into the RF (ncl. pontis oralis). The dorsal part of the ncl. dorsalis tegmenti Guddeni projects bilaterally to the ncl. pontis oralis.

Central projections of the octavolateralis nerves of the clearnose skate, Raja eglanteria.

The central projections of first-order lateral line and octavus nerve afferents of the clearnose skate, Raja eglanteria, were determined by nerve degeneration and horseradish peroxidase techniques. The octavolateralis area of the medulla, which receives these afferents, is organized into dorsal, intermediate, and ventral longitudinal columns of cells and neuropil. Fibers that innervate the electroreceptive sense organs enter the dorsal longitudinal column via the dorsal root of the anterior lateral line nerve and terminate within the dorsal nucleus. Mechanoreceptive fibers from neuromasts of the head and trunk are carried by the ventral root of the anterior lateral line nerve and posterior lateral line nerve, respectively. Both nerves enter the intermediate longitudinal column and terminate throughout the rostrocaudal extent of the intermediate nucleus. Fibers of the ventral root of the anterior lateral line nerve are confined to the medial portion of the intermediate nucleus and posterior lateral line nerve fibers to the lateral portion. In addition, ascending mechanoreceptive fibers from both head and trunk neuromasts project to the vestibulolateral lobe of the cerebellum. Octavus nerve afferents enter the medulla and terminate primarily within the four octaval nuclei that comprise the ventral longitudinal column. Rostrocaudally, these nuclei are the anterior, magnocellular, descending, and posterior octaval nuclei. A few ascending axons continue beyond the anterior octaval nucleus and course to the vestibulolateral lobe of the cerebellum. Some descending axons emanate from the descending octaval nucleus and course to the reticular formation and intermediate nucleus. Therefore, electroreceptive lateral line, mechanoreceptive lateral line, and octavus nerve afferents project ipsilaterally and terminate predominantly within separate medullary nuclei. The significance of octavus nerve projections to the intermediate nucleus and overlap of mechanoreceptive and octavus afferents within the vestibulolateral lobe of the cerebellum cannot be determined until it is known which fibers of the inner ear sense organs project to these areas. Retrograde transport of horseradish peroxidase results in the labeling of large multipolar cells, both ipsilaterally and contralaterally, within a column of gray that lies dorsolateral to the reticular formation. These cells are interpreted as the cell of origin of the efferent components of the anterior and posterior lateral line nerves.

Efferent vestibular neurons: a study employing retrograde tracer methods in the pigeon (Columba livia).

Efferent neurons innervating the vestibular labyrinth and cochlea of the pigeon have been identified by means of a variety of retrograde tracers: [3H]-adenosine (Ad), horseradish peroxidase (HRP), Evan's Blue (EB) and Bisbenzimide (Bb). Discrete injections into individual cristae ampullares of the semicircular canals, into the macula utriculi, or into several of these end organs resulted in similar patterns of neuronal labelling. Efferent vestibular neurons were always found within a small portion of the nucleus reticularis pontis caudalis (RP), ventrolateral to the abducens nucleus on both sides. No systemic difference in the locations of labelled cells was found following injection into different sensory epithelia. Cell counts following injections into individual cristae did not differ significantly from those following injections into all three cristae. The injections into all cristae in both labyrinths yielded cell counts that were much lower than twice the number of cells labelled by injections into the three cristae on one side only. When HRP was injected into the right lateral canal crista and Ad into the right posterior canal crista, a high proportion of neurons was labelled with both compounds (61% of the HRP-labelled cells and 67% of the Ad-labelled cells). Injections of EB into all three cristae on the right side and Bb into all three cristae on the left side produced a smaller percentage of doubly labelled cells (10% of the EB-labelled cells and 6% of the Bb-labelled cells). It is concluded, therefore, that there is a considerable degree of collateralization within one labyrinth. Fewer collaterals of efferent neurons are directed to both labyrinths. Since each semicircular canal represents head rotation in one direction and one plane, it is unlikely that efferents which contact several different movement sensors can provide sensory motor control that is specific for directions and planes of head movements. Control injections of these tracers into the cochlea yielded labelled cells in a different reticular structure, the nucleus reticularis paragigantocellularis lateralis (Pgc), on both sides, as well as in the RP. It is proposed that the Pgc cells represent cochlear efferents, while the RP neurons are related to the macula lagenae, an otolithic organ of balance in the pigeon.