Anatomical substrates of oculomotor control.

It is convenient to describe oculomotor neuroanatomy in terms of five to six different eye movement types, each with relatively independent neural circuitry: saccades, vestibulo-ocular reflex, optokinetic response, smooth pursuit, vergence and, most recently added to the list, gaze-holding. Current research indicates that many structures participate in several eye movement types, such as the nucleus reticularis tegmenti pontis, frontal eye fields and pretectum. However, the circuits appear to run in parallel rather than being integrated.

Identification of motoneurons supplying multiply- or singly-innervated extraocular muscle fibers in the rat.

In mammals, the extraocular muscle fibers can be categorized in singly-innervated and multiply-innervated muscle fibers. In the monkey oculomotor, trochlear and abducens nucleus the motoneurons of multiply-innervated muscle fibers lie separated from those innervating singly-innervated muscle fibers and show different histochemical properties. In order to discover, if this organization is a general feature of the oculomotor system, we investigated the location of singly-innervated muscle fiber and multiply-innervated muscle fiber motoneurons in the rat using combined tract-tracing and immunohistochemical techniques. The singly-innervated muscle fiber and multiply-innervated muscle fiber motoneurons of the medial and lateral rectus muscle were identified by retrograde tracer injections into the muscle belly or the distal myotendinous junction. The belly injections labeled the medial rectus muscle subgroup of the oculomotor nucleus or the greatest part of abducens nucleus, including some cells outside the medial border of abducens nucleus. In contrast, the distal injections labeled only a subset of the medial rectus muscle motoneurons and exclusively cells outside the medial border of abducens nucleus. The tracer detection was combined with immunolabeling using antibodies for perineuronal nets (chondroitin sulfate proteoglycan) and non-phosphorylated neurofilaments. In monkeys both antibodies permit a distinction between singly-innervated muscle fiber and multiply-innervated muscle fiber motoneurons. The experiments revealed that neurons labeled from a distal injection lack both markers and are assumed to represent multiply-innervated muscle fiber motoneurons, whereas those labeled from a belly injection are chondroitin sulfate proteoglycan- and non-phosphorylated neurofilament-immunopositive and assumed to represent singly-innervated muscle fiber motoneurons. The overall identification of multiply-innervated muscle fiber and singly-innervated muscle fiber motoneurons within the rat oculomotor nucleus, trochlear nucleus, and abducens nucleus revealed that the smaller multiply-innervated muscle fiber motoneurons tend to lie separate from the larger diameter singly-innervated muscle fiber motoneurons. Our data provide evidence that rat extraocular muscles are innervated by two sets of motoneurons that differ in their molecular, morphological, and anatomical properties.

Medial rectus subgroups of the oculomotor nucleus ant their abducens internuclear input in the monkey.

Physiological experiments show that the abducens internuclear pathway is involved in the activation of only the medial rectus (MR) eye muscle. Previous anatomical experiments have shown that this pathway terminates in multiple foci within the oculomotor nucleus (OMN) of the monkey, and not only over the classical motoneuron subgroup. In this study the location of MR motoneurons in the monkey OMN is reinvestigated, and compared with the detailed pattern of terminations of the abducens internuclear pathway. The motoneurons were labelled by injections of retrograde tracer substances, HRP and [125I] wheat germ agglutinin (WGA), into extraocular muscles. Labelled MR motoneurons were found in three main divisions, called subgroup A, B, and C. Subgroup A corresponds mainly to the classical ventral MR subgroup. Subgroup B lies dorsal and caudal in OMN, occupying an area classically reserved for inferior rectus (IR). However, the representation of IR is shown to be further rostral in the dorsal OMN. Subgroup C is on the dorsomedial border of OMN. Its cells are significantly smaller than those of Group A and B. In addition C could be labelled independently of the other subgroups by small injections into the outer (orbital) layer of MR muscle. This indicates a functional difference between the subgroups. It is suggested that subgroup C may be important for the tonic component of MR activity, possibly convergence. The location of abducens internuclear terminals, labelled by the injection of tritiated amino acids into the abducens nucleus, corresponds exactly to the position of MR motoneurons. These experiments provide a new picture of the internal OMN organization, and support the physiological findings that the abducens internuclear pathway activates only MR motoneurons.

Premotor neurons for vertical eye movements in the rostral mesencephalon of monkey and human: histologic identification by parvalbumin immunostaining.

In the monkey, premotor neurons for vertical gaze are located in the mesencephalic reticular formation: the rostral interstitial nucleus of the medial longitudinal fascicle (riMLF) contains medium-lead burst neurons, and the interstitial nucleus of Cajal (iC) acts as integrator for the eye-velocity signals to eye-position signals. Both nuclei lie adjacent to each other and are similar in appearance at the transition zone in Nissl-stained sections, which makes a delineation of the functionally different nuclei difficult in human. For a neuropathologic analysis of degenerative changes in saccadic disorders of patients, the histologic identification of the riMLF and the iC is important. The aim of this study is to identify both nuclei in human by using parvalbumin as a histologic marker. First, in monkeys the premotor neurons in riMLF and iC were identified by trans-synaptic labelling after injections of tetanus toxin fragment C into vertical-pulling eye muscles. Premotor neurons were found in the riMLF mainly ipsilateral to the corresponding eye muscle motoneurons and on both sides within the iC, but here the labelled cell populations differed: the contralateral side contained more medium-sized cells compared with the mainly small-sized cell population on the ipsilateral side. Double labelling showed that almost all premotor neurons in the iC and all premotor neurons in the riMLF were parvalbumin-immunoreactive. The immunocytochemical staining of human brainstem sections revealed the riMLF as a cluster of medium-sized, elongated parvalbumin-positive cells, with a similar appearance and at a similar location as that in monkey: a wing-shaped nucleus dorsomedial to the red nucleus, rostral to the traversing tractus retroflexus, dorsally bordered by the thalamo-subthalamic paramedian artery. The adjacent iC could be distinguished easily by its more densely packed, round parvalbumin-immunoreactive neurons. The exact identification of premotor neurons of the vertical system in the normal human brain provides a reference basis for the neuropathologic analysis of vertical gaze disorders at a cellular level.

Histological identification of premotor neurons for horizontal saccades in monkey and man by parvalbumin immunostaining.

The premotor excitatory and inhibitory burst neurons are essential for horizontal saccades. In the monkey, excitatory burst neurons lie in the ipsilateral paramedian pontine reticular formation, and the inhibitory burst neurons lie more caudally in the contralateral nucleus paragigantocellularis dorsalis. For a neuropathological analysis of degenerative changes in saccadic disorders of patients, the histological identification of the burst neuron areas in man is important. Here, we show that this is possible with parvalbumin immunostaining as a histological marker. First, in monkeys, the premotor burst neurons were backlabeled by injections of wheat germ agglutinin-horseradish peroxidase or cholera toxin subunit B into the abducens nucleus or tetanus toxin fragment C into the lateral rectus muscle and shown by double labeling to contain parvalbumin. Then, human brainstem sections were immunoreacted for parvalbumin, and, by comparing the resulting staining pattern to that in the monkey, the homologous burst neuron areas were defined in man. In the monkey, excitatory burst neurons were confirmed to the nucleus reticularis pontis caudalis and did not extend farther rostrally into the nucleus reticularis pontis oralis. All retrogradely labeled cells in both burst neuron areas were parvalbumin positive, and approximately 70% of the parvalbumin-positive cells were retrogradely labeled. Both burst neuron areas were highlighted by their parvalbumin staining pattern and could be outlined in man as well. The putative excitatory burst neuron area in man is in the medial part of the nucleus reticularis pontis caudalis (extending 2.5 mm mediolaterally), immediately rostral (250 microns) to the omnipause neurons and extending 2.2 mm rostrally, and the putative inhibitory burst neuron area lies in the medial part of the paragigantocellular nucleus caudal to the abducens nucleus, extending 1.8 mm caudally. The location of the burst neuron areas, including the burst neurons themselves, via parvalbumin immunostaining will help in the analysis of clinical cases with slow saccades.

Raphe nucleus of the pons containing omnipause neurons of the oculomotor system in the monkey, and its homologue in man.

Omnipause neurons take part in the generation of saccadic eye movements. They lie around the midline in the caudal pontine reticular formation, in an area usually ascribed to the nucleus raphe pontis (rp). In this study of the monkey (Macaca fascicularis and M. mulatta), we describe four series of experiments aimed at establishing that omnipause neurons lie within a distinctive cytoarchitectonic entity, which we call the nucleus raphe interpositus (rip): (1) cytoarchitectural study, (2) recording-lesion experiments to establish in which cell group omnipause neurons lie, (3) cytochrome oxidase distribution in the omnipause region and neighboring structures, and (4) neuroanatomical tracing experiments to demonstrate afferents to the omnipause region. In the detailed cytoarchitectural study of the midline structures in the caudal pons and rostral medulla, a distinctive group of neurons (rip) adjoining the ventrocaudal border of rp and dorsal to the nucleus raphe magnus (rm) is described. The striking features of rip are the uniformly arranged, narrow row of the cells either side of the midline, and the extensive horizontally oriented dendritic trees of its neurons. The abducens rootlets (NVI) pass through the reticular formation at the same rostrocaudal level as rip and form a reliable landmark for its location. Cytochrome-oxidase-stained sections demonstrated additional differences between rip and adjacent cell groups: in rip the neurons and their extensive dendrites stained strongly, but not the surrounding neuropile, whereas in rp both the neurons and the neuropile stained darkly, so that individual neurons were difficult to see. Unlike rp, rip coincides with the location of omnipause neurons, and lesions marking the sites of individual omnipause units lay within its boundaries. Tritiated leucine was injected into superior colliculus (sc), which is known to have monosynaptic connections with omnipause neurons. Labelled axons and patterns of silver grains taken to indicate the presence of terminal branching were found in and around rip, but no significant labelling was seen in rp or rm. It is concluded that the omnipause neurons lie within the rip in the monkey. These functional and morphological differences between rip and the adjacent raphe nuclei are used to justify its characterization as an independent cell group in the monkey. In order to relate these findings to man, cytochrome oxidase experiments were carried out on the human brainstem, and the pattern of staining at the level of the abducens rootlets was correlated with the cytoarchitecture.(ABSTRACT TRUNCATED AT 400 WORDS)

Efferent pathways of the nucleus of the optic tract in monkey and their role in eye movements.

To clarify the role of the pretectal nucleus of the optic tract (NOT) in ocular following, we traced NOT efferents with tritiated leucine in the monkey and identified the cell groups they targeted. Strong local projections from the NOT were demonstrated to the superior colliculus and the dorsal terminal nucleus bilaterally and to the contralateral NOT. The contralateral oculomotor complex, including motoneurons (C-group) and subdivisions of the Edinger-Westphal complex, including motoneurons (C-group) and subdivisions of the Edinger-Westphal complex, also received inputs. NOT efferents terminated in all accessory optic nuclei (AON) ipsilaterally; contralateral AON projections arose from the pretectal olivary nucleus embedded in the NOT. Descending pathways contacted precerebellar nuclei: the dorsolateral and dorsomedial pontine nuclei, the nucleus reticularis tegmenti pontis, and the inferior olive. Direct projections from NOT to the ipsilateral nucleus prepositus hypoglossi (ppH) appeared to be weak, but retrograde tracer injections into rostral ppH verified this projection; furthermore, the injections demonstrated that AON efferents also enter this area. Efferents from the NOT also targeted ascending reticular networks from the pedunculopontine tegmental nucleus and the locus coeruleus. Rostrally, NOT projections included the magnocellular layers of the lateral geniculate nucleus (lgn); the pregeniculate, peripeduncular, and thalamic reticular nuclei; and the pulvinar, the zona incerta, the mesencephalic reticular formation, the intralaminar thalamic nuclei, and the hypothalamus. The NOT could generate optokinetic nystagmus through projections to the AON, the ppH, and the precerebellar nuclei. However, NOT also projects to structures controlling saccades, ocular pursuit, the near response, lgn motion sensitivity, visual attention, vigilance, and gain modification of the vestibulo-ocular reflex. Any hypothesis on the function of NOT must take into account its connectivity to all of these visuomotor structures.

Neuroanatomical identification of mesencephalic premotor neurons coordinating eyelid with upgaze in the monkey and man.

Except during blinks, movements of the upper eyelid are tightly coupled to vertical eye movements. The premotor source for the coordination of lid and eye movements is unknown. The present paper provides the anatomical identification of a new premotor cell group in the rostral mesencephalon of the monkey and human, which lies in close proximity to the premotor center for vertical saccades and is thought to participate in lid-eye coordination. After injections of a retrograde transsynaptic tracer (tetanus toxin fragment C or BII(b)) into the levator palpebrae (LP), the superior rectus (SR), or the inferior oblique (IO) muscle of macaque monkeys, a small circumscribed group of premotor neurons was labeled in the central gray of the rostral mesencephalon, but not after superior oblique or inferior rectus muscle injections. This group lies immediately rostral to the interstitial nucleus of Cajal and medial to the rostral interstitial nucleus of the medial longitudinal fasciculus, each of which contain premotor neurons for vertical saccades, and was termed the M-group. Injections of tritiated leucine into the M-group led to afferent labeling primarily over LP motoneurons. In addition, label was present over the SR- and IO-motoneuron subgroups in the oculomotor nucleus and frontalis muscle motoneurons in the facial nucleus. This projection pattern of the M-group suggests a role in the coordination of the upper eyelid and eyes during upgaze. Double-labeling experiments in macaque monkeys revealed that the M-group is strongly parvalbumin immunoreactive and contains high levels of cytochrome oxidase activity. With these two histochemical markers, the homologue of the M-group was identified in the human brain as well.

Projections from the superior colliculus motor map to omnipause neurons in monkey.

Descending projections from the superior colliculus (SC) motor map to the saccadic omnipause neurons (OPNs) were examined in monkeys by using anterograde transport of tritiated leucine. The SC was divided into three zones: the rostral pole of the motor map, a small horizontal saccade zone in central SC, and a large horizontal saccade zone in caudal SC. Tracer injections into the intermediate layers of the three zones led to different patterns of silver grain deposits in and around nucleus raphe interpositus (RIP), which contains the OPNs: 1) From the rostral pole of the motor map, coarse axon branches of the crossed predorsal bundle spread medially into the RIP, branched, and terminated predominantly unilaterally over cells on the same side. 2) From the small horizontal saccade zone, the axon branches were of a finer caliber and terminated diffusely in the RIP, mainly on the same side. 3) From the large horizontal saccade zone, no terminal labeling was found within the RIP. 4) From the rostral pole of the motor map and small horizontal saccade zone, fiber branches from the ipsilateral descending pathway terminated diffusely over RIP. 5) In addition, terminal labeling in reticulospinal areas of the pons and medulla increased in parallel with the size of the saccade according to the SC motor map. The results suggest that there are multiple projections directly onto OPNs from the rostral SC but not from the caudal SC associated with large gaze shifts. The efferents from the rostral pole of the motor map may subserve the suppression of saccades during visual fixation, and those from the small horizontal saccade zone could inhibit anatagonist premotor circuits.

Pretectal projections to the oculomotor complex of the monkey and their role in eye movements.

The nucleus of the optic tract (NOT) is associated with the generation of optokinetic nystagmus (OKN), whereas the olivary pretectal nucleus (ol), which lies embedded in the primate NOT, is believed to be essential for the pupillary light reflex. In this anatomical study of the pretectum, projections from NOT and ol to structures around the oculomotor nucleus were traced in the monkey, to determine which cell groups they innervated. 1. 3[H]-leucine injections were placed into NOT and ol, and labelled terminals were observed just outside the classical oculomotor nucleus (nIII), in the "C-group' and midline cell clusters, both of which contain small motoneurons of the extraocular eye muscles. In addition, there were strong projections to the lateral visceral cell column of the Edinger-Westphal complex (lvc), but not to the Edinger-Westphal nucleus (EW) itself. All of these projections were mainly contralateral. 2. NOT efferents terminated over the ipsilateral medial accessory nucleus of Bechterew (nB), but not over the adjacent nucleus Darkschewitsch. 3. Injections of a retrograde tracer into the oculomotor complex showed that the pretectal afferents described above originated mainly from the dorsomedial part of NOT and from ol. 4. The use of a transsynaptic retrograde tracer, tetanus toxin fragment (BIIb), established the monosynaptic nature of the connection between dorsomedial NOT (contralaterally) and ol (bilaterally), to the small extraocular motoneurons outside classical nIII. The "C-group' motoneurons may play a role in vergence, and lvc in pupillary constriction and depth of focus. Our results imply that NOT and ol participate in the control of some aspects of the near-response, which may be important in the generation of some components of OKN in primates.

Motoneurons of twitch and nontwitch extraocular muscle fibers in the abducens, trochlear, and oculomotor nuclei of monkeys.

Eye muscle fibers can be divided into two categories: nontwitch, multiply innervated muscle fibers (MIFs), and twitch, singly innervated muscle fibers (SIFs). We investigated the location of motoneurons supplying SIFs and MIFs in the six extraocular muscles of monkeys. Injections of retrograde tracers into eye muscles were placed either centrally, within the central SIF endplate zone; in an intermediate zone, outside the SIF endplate zone, targeting MIF endplates along the length of muscle fiber; or distally, into the myotendinous junction containing palisade endings. Central injections labeled large motoneurons within the abducens, trochlear or oculomotor nucleus, and smaller motoneurons lying mainly around the periphery of the motor nuclei. Intermediate injections labeled some large motoneurons within the motor nuclei but also labeled many peripheral motoneurons. Distal injections labeled small and medium-large peripheral neurons strongly and almost exclusively. The peripheral neurons labeled from the lateral rectus muscle surround the medial half of the abducens nucleus: from superior oblique, they form a cap over the dorsal trochlear nucleus; from inferior oblique and superior rectus, they are scattered bilaterally around the midline, between the oculomotor nucleus; from both medial and inferior rectus, they lie mainly in the C-group, on the dorsomedial border of oculomotor nucleus. In the medial rectus distal injections, a "C-group extension" extended up to the Edinger-Westphal nucleus and labeled dendrites within the supraoculomotor area. We conclude that large motoneurons within the motor nuclei innervate twitch fibers, whereas smaller motoneurons around the periphery innervate nontwitch, MIF fibers. The peripheral subgroups also contain medium-large neurons which may be associated with the palisade endings of global MIFs. The role of MIFs in eye movements is unclear, but the concept of a final common pathway must now be reconsidered.

Localizing value of torsional nystagmus in small midbrain lesions.

BACKGROUND: The topodiagnostic value and specificity of nystagmus in patients with mesencephalic lesions and its relation to tonic torsional deficits and vertical saccade deficits is controversial and anecdotal. METHODS: The authors examined 11 patients with vascular MRI-identified mesencephalic lesions and clinical evidence of vertical-torsional nystagmus on gaze straight ahead, focusing on the three-dimensional nystagmus components recorded with the three-dimensional search coil technique. RESULTS: Combined lesions of the rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF) and the interstitial nucleus of Cajal (iC) are much more frequent than riMLF and, in particular, iC lesions alone. Eight patients showed contralesional torsional nystagmus with a conjugate vertical component on gaze straight ahead and had anatomic (MRI) and clinical evidence (slowing of vertical saccades) for riMLF involvement. Tonic ocular torsion and the subjective visual vertical were shifted to the contralesional side (n = 7). Torsional nystagmus to the ipsilesional side was uncommon (n = 3) and found in patients with midbrain lesions involving the iC, all of whom also had decreased time constants of the slow phases of gaze-evoked nystagmus. CONCLUSIONS: Contrary to previous proposals, contralesional torsional nystagmus was the most frequent direction and is probably not compensatory for contralesional tonic ocular torsion. Small amplitude vertical saccades with normal velocities in association with ipsilesional torsional nystagmus may indicate isolated iC lesions. Torsional nystagmus following mesencephalic lesions may last for years and may help to distinguish rostral (riMLF) from caudal (iC) midbrain lesions.

Abnormalities of torsional fast phase eye movements in unilateral rostral midbrain disease.

In a patient with a unilateral rostral midbrain lesion, three-dimensional scleral search coil eye movement recordings demonstrated slowing of ipsidirectional torsional fast phase eye movements without any abnormalities of torsional slow phases. On high-resolution MRI, the brainstem lesion localized to the area of the efferent pathways from the rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF). This supports the experimental finding that unilateral inactivation of the riMLF results in a loss of ipsidirectional torsional fast phase eye movements and the hypothesis that there is lateralization of function throughout the torsional fast eye movement generating system.

Ptosis and supranuclear downgaze paralysis.

A patient developed the unusual combination of a supranuclear downward gaze paralysis and bilateral ptosis. It was caused by a single midbrain glioma. Other ocular motor functions were intact. The neuropathologic examination showed a tumor growing mainly around the third ventricle and the aqueduct. The findings agree with recent experimental evidence that a network of neural elements involved in eyelid control lies in the supraoculomotor area immediately dorsal to the oculomotor nucleus.

Pathophysiology of slow vertical saccades in progressive supranuclear palsy.

OBJECTIVES: To investigate the relative roles of burst neurons (which generate the saccadic command) and omnipause neurons (which gate the activity of burst neurons) in the pathogenesis of slow saccades in progressive supranuclear palsy (PSP). BACKGROUND: Experimental inactivation of mesencephalic burst neurons impairs vertical but not horizontal saccades. Experimental inactivation of omnipause neurons causes slowing of both horizontal and vertical saccades. Combining saccadic with vergence movements in healthy subjects induces small, high-frequency, conjugate oscillations, which indicate that omnipause neurons are inhibited. METHODS: The authors studied seven patients with PSP, six patients with other parkinsonian syndromes, and seven age-matched control subjects. They compared vertical saccades of similar sizes made with or without associated vergence movements. They compared the speed of vertical and horizontal saccades. RESULTS: Five patients with PSP and the six patients with other parkinsonian made vertical saccades in combination with horizontal vergence; all showed conjugate horizontal oscillations (29 to 41 Hz) during 27% to 93% of saccade-vergence trials. Vertical saccades made in conjunction with vergence movements were not speeded up or increased in size compared with saccades made between equidistant targets for the PSP or parkinsonian groups. Vertical saccades were slowed more than horizontal saccades in the PSP group (p < 0.005) but not in the parkinsonian group. CONCLUSIONS: Dysfunction of omnipause neurons ("gate dysfunction") is unlikely to be the primary cause of slow vertical saccades in progressive supranuclear palsy. Deficient generation of the motor command by midbrain burst neurons is the more likely cause.

Spatial distribution of field potential profiles in the cat cerebellar cortex evoked by peripheral and central inputs.

The present study was designed to characterize the spread of excitation within the frontal plane of the cat cerebellar cortex following different types of stimuli. In particular, experiments were performed to determine whether the spread of excitation evoked by mossy fibre inputs proceeds primarily along the parallel fibres ("beam-like" spread) or whether these inputs activate non-propagated foci ("patches") in the cerebellar cortex. Field potentials were recorded within a frontal plane as a medial to lateral array at different depths in parallel tracks. The recordings were made following electrical stimulation of different forelimb nerves and functionally related areas of the sensorimotor cortex as well as during passive paw movements. The resulting spatial grid of responses provides discrete spatio-temporal information reflecting the activation of specific cerebellar afferents and the neuronal interactions they evoke. The method employed demonstrates the spatial distribution of the temporal sequence of excitability changes throughout all the cerebellar cortical layers. In general, the characteristics of the responses in the intermediate cerebellar cortex depended on the source of the signals. Activity patterns evoked by peripheral nerve stimulation showed more clustered foci compared with those following electrical stimulation of functionally related areas of the sensorimotor cortex. The centrally evoked profiles were generally more homogeneous. The largest number of foci were observed following passive movements around the wrist joint. The spread of excitation in the vertical direction was evaluated by the spatial shift of the line of reversal of the N3/P2-potential (zero-isopotential line). Lines of reversal for peripherally-evoked activity patterns were approximately 90 microns closer to the molecular layer than those evoked by central stimulation in animals in which recordings have been performed in lobule Vc. The opposite was found for recordings in lobule Vb, where potential reversals following peripheral stimulation were located 40 microns deeper than those evoked following central stimulation. Cortical inputs resulted in a more proximal activation of lobule Vc Purkinje cell dendrites than in lobule Vb. This type of input processing thus seems to be lobule dependent. A beam-like spread of excitation could not be demonstrated. For both climbing fibre and mossy fibre afferent systems multiple foci were found in the frontal plane. The foci due to mossy fibre activation arose from the granular layer and expanded vertically to the molecular layer. For the climbing fibre system the foci were restricted to the molecular layer, where they merged to form a superficial band of activation. Although the data presented in this paper favour a focal distribution of activity, they do not exclude beam-like propagation along the parallel fibres, because of the difficulty of detecting this pattern in response to the stimuli. The "beam"- and "patch"-like hypotheses need not be mutually exclusive. Each could contribute to a specific stage of the temporal-spatial processing in the cerebellar cortex in a functional and task-specific manner.

A review of otolith pathways to brainstem and cerebellum.

Our knowledge of otolith pathways is developing rapidly, but is still far from complete. Primary afferents from the sacculus and utricle terminate mainly in the lateral, inferior and caudal superior vestibular nuclei, and the ventral cerebellum, in particular the nodulus. Otolith signals descend via reticulo- and vestibulospinal pathways in the spinal cord to influence neck motoneurons and ascending proprioceptive afferents. Utricular information can reach the extraocular eye muscles via mono-, di-, and multisynaptic pathways, but saccular afferents probably only by multisynaptic pathways. The otolith signals are relayed from the vestibular nuclei, medullary reticular formation, inferior olive, and lateral reticular nucleus to sagittal zones in the caudal cerebellar vermis (nodulus and uvula), and influence the deep cerebellar nuclei. The graviceptive information could be channeled by the cerebellar efferents back to the vestibular and inferior olive complex, or fed into ascending pathways that would innervate the mescencephalon, the thalamus, and cerebral cortex.

Pathways from cell groups of the paramedian tracts to the floccular region.

A group of cells lying along the midline of the mid-medulla, nucleus pararaphales, is shown to play a role in vertical eye movements. Its efferents project along the midline, then pass laterally to follow the ventral external arcuate fibers around the surface of the medulla into the restiform body. The fibers terminate in the flocculus and ventral paraflocculus. This nucleus is one of the "cell groups of the paramedian tracts," which, based on their connectivity, could provide a motor-feedback signal for eye-head position to the cerebellum. Lesions of these pathways could lead to gaze-evoked nystagmus.

Motor and sensory innervation of extraocular eye muscles.

Eye muscles are unusual in several ways; one is that they have up to three different layers-the inner global layer, the outer orbital layer, and in some species an external marginal layer has been described. In sheep this is called the "peripheral patch layer." Three different types of proprioceptors are found in eye muscles-muscle spindles, Golgi tendon organs, and palisade endings. A survey of the organization of their location leads us to the hypothesis that each receptor is confined to a separate layer of the eye muscle. The palisade endings are associated with the global layer, the muscle spindles lie predominantly in the orbital layer, and the Golgi tendon organs are found only in the peripheral patch layer. This well-organized scheme may help us to understand the proprioceptive system in eye muscles.

Modern concepts of brainstem anatomy: from extraocular motoneurons to proprioceptive pathways.

The extraocular muscles, unlike the skeletal muscles, contain non-twitch muscle fibers. Recent experiments have located the non-twitch motoneurons. They lie around the periphery of the oculomotor, trochlear and abducens nuclei, separate from the more usual twitch motoneurons that cluster within the boundaries of the classical motor nuclei. The premotor inputs to non-twitch neurons were traced by the injection of rabies virus into the distal tip of the lateral rectus muscle. Retrogradely labeled cells were found in areas associated with the neural integrator, vergence and smooth pursuit premotor areas, but not the saccadic premotor burst neurons or the direct vestibulo-ocular pathways. The rabies tracing emphasizes for the first time that the central mesencephalic reticular formation (cMRF) and the supraoculomotor area exert direct premotor control over the non-twitch motoneurons. Because the two sets of motoneurons do not receive the same afferents, they must have different functions; these are not yet clarified. These results are not compatible with the concept of a single final common pathway from motoneurons to eye muscles. Putative sensory receptors, palisade endings, are located at the tips of non-twitch muscle fibers reminiscent of an inverted muscle spindle, which would make the non-twitch motoneurons, gamma-motoneurons. We propose that twitch motoneurons are the major source of tension used for eye movements, whereas non-twitch motoneurons are more important for fine alignment of the eyes. Furthermore, the non-twitch motoneurons could be controlled through sensory feedback networks (including perhaps proprioceptive signals from the palisade endings) that are relayed through the superior colliculus and via cMRF to the non-twitch motoneurons. The clinical repercussions of these hypotheses are discussed.