afferents to the abducens nucleus in the monkey and cat

THE JOURNAL OF COMPARATIVE NEUROLOGY 245~379-400 (1986)

Afferents to the Abducens Nucleus in the Monkey and Cat

T. LANGER, C.R.S. KANEKO, C.A. SCUDDER, m A . F . FUCHS Department of Physiology and Biophysics, and Regional Primate Research Center,

University of Washington, Seattle, Washington 98195

ABSTRACT The abducens nucleus is a central coordinating element in the genera-

tion of conjugate horizontal eye movements. As such, it should receive and combine information relevant , to visual fixation, saccadic eye movements, and smooth eye movements evoked by vestibular and visual stimuli. To reveal possible sources of these signals, we retrogradely labeled the afferents to the abducens nucleus by electrophoretically injecting horseradish peroxidase into an abducens nucleus in four monkeys and two cats. The histologic material was processed by the tetramethyl benzidine (TMB) method of Mesulam.

(1) In both species the largest source of afferents to the abducens nucleus was bilateral projections from the ventrolateral vestibular nucleus and the rostral pole of the medial vestibular nucleus. Scattered neurons were also labeled in the middle and caudal levels of the medial vestibular nucleus.

(2) Large numbers of neurons were labeled in the ventra1,margin of the nucleus prepositus hypoglossi in the cat and in the common margin of the nucleus prepositus and the medial vestibular nucleus in the monkey, a region we call the marginal zone.

(3) Substantial numbers of retrogradely labeled neurons were found in the dorsomedial pontine reticular formation both caudal and rostral to the abducens nuclei.

(4) In the monkey, large numbers of labeled neurons were present in the contralateral medial rectus subdivision of the oculomotor complex, while smaller numbers occurred in the ipsilateral medial rectus subdivision and elsewhere in the oculomotor complex. In the cat, large numbers of retrogradely labeled cells were present in a small periaqueductal gray nucleus immediately dorsal to the caudal pole of the oculomotor complex, and a few labeled neurons were also dispersed through the caudal part of the oculomotor complex.

(5) Occasional labeled neurons were present in the contralateral superior colliculus in both species. The size and distribution of the labeled neurons within the intermediate gray differed dramatically in the two species. In the cat, the retrogradely labeled neurons were very large and occurred predominantly in the central region of the colliculus, while in the monkey, they were small to intermediate in size and were distributed more uniformly within the middle gray.

(6) Among the afferent populations present in the monkey, but not in the cat, was a group of scattered neurons in the ipsilateral rostral interstitial nucleus of the medial longitudinal fasciculus and a denser, bilateral population in the interstitial nucleus of Cajal. A few labeled neurons were also present in the nucleus reticularis tegmenti pontis in the monkey, but not in the cat. Small numbers of retrogradely labeled neurons were present in the monkey's y-group and superior vestibular nucleus.

Accepted October 30,1985.

0 1986 ALAN R. LISS. INC.

380 T. LANGER ET AL.

It appears that while the monkey and the cat have similar sets of abducens afferents, there are definite differences in the details of the innervation of the abducens nucleus in the two species.

Key words: oculomotor system, Vestibular nuclei, reticular formation, horseradish peroxidase, cytoarchitecture

The abducens nucleus is clearly important in the generation of conjugate horizontal eye movements. The lateral rectus motoneurons, which abduct the ipsilateral eye, constitute a major population of the abducens nucleus. A second large population of abducens neurons, the internuclear neurons, has been shown to innervate the medial rectus motor nucleus in the contralateral oculomotor complex, but not the lateral rectus muscle (Graybiel and Hartweig, '74; Spencer and Sterling, '77; Highstein and Baker, '78; Steiger and Buttner-Ennever, '79; Carpenter and Batton, '80). Since the discharge patterns of the abducens internuclear neurons are qualitatively similar to those of abducens motoneurons (Fuchs and Luschei, '70; Schiller, '70; Delgado- Garcia et al., '83; Fuchs et al., '84), an increase in the level of activity in the abducens nucleus causes the contraction of both the ipsilateral lateral rectus muscle and the contralateral medial rectus muscle, thereby producing a conjugate horizontal shiR of gaze. Many abducens neurons project to other oculomotor structures. About one-third of the internuclear neurons have axon collaterals that extend caudal to the abducens nucleus (Highstein et al., '82; Strassman et al., '82), where they may innervate the nucleus prepositus hypoglossi and parts of the vestibular complex (Baker and McCrea, '79). In addition, a number of investigators have recently demonstrated a population of abducens neurons that project to the flocculus of the cerebellum (Alley et al., '75; Graybiel, '77a; Kotchabhakdi and Walberg, '77; Sat0 et al., '83; Blanks et al., '83; Blanks and Torigoe, '83; Langer et al., '85). On the basis of these observations, it is clear that the abducens nucleus has a much larger role in the generation of eye movements than simply contracting the lateral rectus muscle and that there may be several types of abducens neurons with different sets of connections.

Since the abducens nucleus participates in the generation of all conjugate horizontal eye movements, it must receive afferents conveying information relevant to visual fixation, saccadic eye movements, and smooth eye movements evoked by visual and vestibular stimuli. A few of these afferents have been studied both anatomically and physiologically. It is known that burst neurons reside in the dorsomedial pontine reticular formation immediately rostral and caudal to the abducens nucleus. In the cat, the caudal burst neurons have been shown to provide a powerful, widespread inhibition to the contralateral abducens nucleus during ipsilaterally directed saccades, and the rostral burst neurons are thought to excite the ipsilateral abducens neurons during ipsilaterally directed saccades. Neurons with similar discharge patterns have been found in the monkey's rostral pontine reticular formation (Luschei and Fuchs, '72; Keller, '74; Henn and Cohen, '761, but it is not known whether these neurons project to the ipsilateral abducens nucleus. There is evidence that the caudal pontine burst neurons innervate the contralateral abducens nucleus in the monkey (Scudder et al., '82). There are neurons in the vestibular complex that receive a direct vestibular nerve input and project to one or the other abducens nucleus

(McCrea et al., '80; Ishizuka et al., '80); these neurons are believed to be the interneurons of the vestibulo-ocular reflex in the cat. Similar neurons have been seen in the alert, behaving monkey (Scudder and Fuchs, '81).

In the cat brainstem, sources of abducens afferents have been demonstrated in the vicinity of each of these functional populations following the injection of horseradish peroxidase (HRP) into the abducens nucleus (Maciewicz et al., '75; Gacek, '79; Baker and Spencer, '81; Stanton and Greene, '81). Additional afferent sources have been found in the nucleus prepositus hypoglossi, the oculomotor corn- plex and the superior colliculus. There has been no comparable anatomical study in the monkey brainstem.

To confirm the existence of these afferent sources in the monkey brainstem, we injected HRP into the abducens nucleus and used a more sensitive chromogen than was generally used in the studies of the cat's abducens afferents. By using the tetramethyl benzidine (TMB) protocol (Mesulam, '78), we were able to label many more cells than had been reported in previous cat studies and some populations that had not been described at all. We also processed several cat brains by the same protocols to obtain brainstem series for comparison with our monkey series.

MATERIALS AND METHODS The normal material, used for the cytoarchitectonic anal-

ysis of the monkey's abducens nucleus, was derived frlDm three rhesus macaque brains perfused with saline and several liters of 10% formalin. The brains were held for a year or more in 10% formalin before they were slowly dehy- drated through a graded series of alcohols, two changes of ether-alcohol, and a graded series of low-viscosity celloidin in ether-alcohol (LaBossiere, '76). They were infiltrated with 12% celloidin under a vacuum and cured in chloroform vapors in the refrigerator. Alternating 40-pm and 80.pm sections were stained with cresylecht violet, lux01 fast blue, neutral red, or an unsuppressed Fink-Heimer silver stain (Fink and Heimer, '67) and mounted with Permount and a coverslip. The advantage of this procedure, particularly the long time in formalin, is that it suppresses the staining of glia, but not of neurons.

All other material was processed to demonstrate retrogradely transported HRP. The material with injections of HRP into the flocculus has been described elsewhere (Lan- ger et al., '85). The internuclear neurons were labeled by a large hydraulic injection of 30% Sigma type V1 HRP (in pH 8.6 Tris bufTer) in the oculomotor complex and the rostral mesencephalon contralateral to the labeled abducens nucleus. The lateral rectus motoneurons were labeled by plac- ing a large injection of HRP (30% Sigma type VI in pH 8.6 Tris buffer) in the belly and proximal end of the lateral rectus muscle, which had been exposed by a dissection of the lateral wall of the orbit.

One abducens nucleus in each of four monkeys (two Ma- caca mulatta and two M. fmcicularis) was injected with HRP (30% Sigma type VI in pH 8.6 Tris buffer). Two differ-

AFFERENTS TO THE ABDUCENS NUCLEUS 381

ent approaches were used to reach the abducens nucleus. In MoHAb-1 and MoHAb-2, a posterior fossa craniotomy and posterior vermalectomy exposed the floor of the fourth ventricle, allowing direct visualization of the genu of the facial nerve. A stimulating electrode was placed in the lateral rectus muscle to evoke an antidromic field potential in the ipsilateral abducens nucleus. The region of the abducens nucleus was explored systematically with a tung- sten microelectrode to optimize the amplitude of the evoked potential. A small electrophoretic injection of HRP (30% Sigma type VI in pH 8.6 Tris buffer) was placed at the site of the maximum potential. In MoHAb-5 the micropipette was passed through the anterior vermis, in the stereotaxic plane, rather than through the fourth ventricle. Since MoHAb-4 had been used in a chronic electrophysiologic study of the abducens nucleus, the location and extent of the nucleus were known in some detail. In that case the micropipette was passed through the recording microelectrode cannula and into the abducens nucleus. The two cat brains also received electrophoretic injections via the posterior cranial fossa.

Each experimental animal, except the monkey used in the chronic electrophysiological study, was anesthetized with Nembutal during the surgery, the exploration of the abducens nucleus, and the injection of HRP. The wound margins were anesthetized by topical application of benzocaine hydrochloride, and after the injection the wound was closed and more benzocaine was applied. The animal was sedated with Valium and/or Nembutal during the survival period. About 24 hours later, the animal was anesthetized with Nembutal and perfused transcardially with 250 ml of saline followed by 2 liters of 1% paraformaldehyde-2% glu- taraldehyde in 0.1 M phosphate buffer. The brain was quickly exposed, blocked, removed into 30% sucrose in 0.1 M phosphate buffer, and held overnight in a refrigerator. The next day the brainstem was sectioned in 80-pm sections on a freezing microtome, promptly treated with TMB (Me- sulam, '78), mounted, and dried in a warming oven. After a thorough drying, every other section was stained with neutral red and all the sections were coverslipped with Per- mount. Every section was examined with a Leitz Dialux 20 microscope and every other section was charted as in the illustrations. The quantitative methods used to assess the distribution of cell size were the same as have been described elsewhere (Langer, '85). They are based on measur- ing the area of the cell body image in the microscope.

RESULTS Heterogeneity of the monkey abducens nucleus

In the monkey, the abducens nucleus is composed of several differentially distributed populations of neurons with markedly different projections (Fig. 1). In addition to the well-known motoneurons (Fig. 1, series B) and the internuclear neurons (Fig. 1, series A), we will consider the neurons in the abducens nucleus that labeled following injections of HRP in the flocculus (Fig. 1, series C), which we refer to as floccular afferents &anger et al., '85). Each of these populations can be associated with cytoarchitectural subdivisions of the abducens nucleus (Fig. 2).

CgtOarchitecture of the abducens nucleus. In normal histological material (Fig. Z), the abducens nucleus lies within the dorsomedial pontine tegmentum, in close asso- ciation with the genu of the facial nerve. Most caudally (Fig. ZA), the abducens nucleus extends dorsal to the ascending limb of the facial nerve as it collects into the com-

pact genu. At this level the abducens nucleus (Ab) is clearly demarcated by its large, chromatophilic neurons scattered in a comparatively open array over the dorsal surface of the facial nerve. It is readily differentiated from several other groups of neurons in the vicinity. Medially, there is a small, round mass of intermediate-sized chromatophilic neurons in the caudal supragenu nucleus (asterisks), which may be comparable to the supragenual area in the cat (Brodal, '52), and it has also been called the supragenicu- late nucleus (Olszewski and Baxter, '54). The medial Nbm) and ventrolateral vestibular nuclei extend along the lateral margin of the abducens nucleus.

The abducens nucleus attains its maximal cross-sectional area between the ascending and descending limbs of the facial nerve (Fig. ZA,C). Abducens neurons intrude dorsally into the space lateral and caudal to the genu (Fig. 2B) until they almost reach the ventricular surface. The caudal supragenu nucleus gradually diminishes in size until it is only a few scattered neurons dorsal to the rostral part of the genu (Fig. 2B). The lateral margin of the abducens nucleus remains proximate to the medial and ventrolateral vestibular nuclei and the medial part of the ventral margin becomes contiguous with the supragigantocellular reticular formation (NRS), a region rostral and dorsal to the nucleus reticularis gigantocellularis (also see Fig. 8). We were particularly interested in the NRS because, as will be shown below, it probably contains the inhibitory burst neurons (IBNs) as well as several other populations of neurons.

Distribution of motoneumns and internuclear neu- mm. In the experimental, HRP-labeled series, the motoneurons were most densely packed in the part of the abducens nucleus between the ascending and descending limbs of the facial nerve (Fig. 1, series B) and the internuclear neurons were concentrated ventrally in the most caudal sections, but shifted rapidly to occupy the lateral part of the nucleus over most of its rostrocaudal extent (Fig. 1, series A). This obvious restriction of certain neuronal groups to certain parts of the abducens nucleus is apparently not present in the cat (Steiger and Buttner-Ennever, '78; High- stein et al., '82). Spencer and Porter ('81) observed very similar patterns in their studies of the monkey's abducens nucleus. Keeping in mind that we are comparing HRP and Nissl material, it is interesting that in the Nissl-stained sections, the parts of the abducens nucleus that contain internuclear neurons have a greater density of cells than those parts that contain only motoneurons (Fig. 2). The parts containing only motoneurons have a characteristically lower density and the neuronal cell bodies appear uniform in size and morphology. Though the neurons in the latter region appear larger on average, a quantitative comparison of the sizes of the neurons in these two regions revealed only slight differences in average cell sue and in the distribution of cell size. Quantitative comparison of the populations of retrogradely labeled neurons in the two illustrated series (Fig. 1) yielded almost identical distributions of cell size.

For a sample of 947 neurons labeled retrogradely from an injection of lectin-HRP in the lateral rectus muscle, the mean and standard deviation of the cross-sectional area of the HRP-filled cell bodies (Langer, '85) were 479.58 k 179.00 pm2. The same measurements of a sample of 421 neuronal cell bodies that were labeled following a large injection of HRP that involved the contralateral oculomotor complex were 467.45 & 171.01 pm2. Given the possible variation because the samples were drawn from different brains, we


series A series B

Fig. 1. Distributions of three types of abducens neurons within the monkey’s abducens nucleus. Series A. Labeled neurons in the abducens nucleus following injection of HFtP into the contralateral oculomotor nuclei and rostral mesencephalon. Note that the cells occupy only certain parts of the abducens nucleus, a s was previously shown by Graybiel (‘77a). Series B. Labeled neurons in another monkey’s abducens nucleus following injection of HRP into the ipsilateral lateral rectus muscle. The motoneurons are almost entirely confined within the conventional borders of the abducens

series C :

flocculus afferents

nucleus, but a t more rostral levels (sections at top of figure) parts of the abducens nucleus are devoid of motoneurons and internuclear neurons. Series C. Labeled neurons in the abducens nuclei and adjacent parts of the raphe and the supragenu nuclei (asterisks) following injection of HRP into the right flocculus. Note the continuity between the labeled raphe neurons and the labeled population in the dorsomedial and rostral abducens nucleus. Note also the complementarity between this distribution and that of the motoneurons in series B.

AF’FERENTS TO THE ABDUCENS NUCLEUS 383

were not able to distinguish these two populations on the basis of their cell body sues.

Flocculur afferents. In caudal sections through the abducens nucleus (Fig. 1, series C, bottom sections), the neurons labeled retrogradely from the flocculus were primarily in the caudal supragenu nucleus, with only a few labeled neurons scattered in small clusters in the raphe between the abducens nuclei. About midway through the abducens nucleus, where the genu begins to pass over its dorsal surface (Fig. 1, series C , third and fourth sections from the bottom), there were sizable clusters of neurons in the raphe between the abducens nuclei. These were larger and more chromatophilic than those seen more caudally and ventrally. These islands of cells began to intrude between the medial longitudinal fasciculus (mlf) and the genu and then infiltrated between the fascicles of the immediately subjacent paramedian tracts, forming cellular bridges to an in- creasingly prominent collection of labeled neurons in the medial and dorsal margins of the abducens nuclei. In rostral sections through the abducens nucleus, there was a striking absence of the motoneurons in the immediate vicinity of the paramedian tracts, the mlf, and the genu of the facial nerve (Fig. 1, series B, top three sections), presumably because they were displaced by neurons that projected to the flocculus. This displacement was even more apparent in the rostral pole of the abducens nucleus, just anterior to the genu of the facial nerve (Fig. 1, series B and C, topmost

Ab c o D das Den Fac G ICP INC 10 IPN ML. mlf NM NPC NPO NRG NRP NRPV NRS NRTP OMC PN PPh py’ RB R M RN sc SCN SCP SN so Sol T Tmo Tt Vbi Vbl Vbm VbS

Abbreuiations

abducens nucleus cochlear nucleus nucleus of Darkschewitsch dorsal acoustic stria dentate nucleus facial nucleus genu of the facial nerve inferior cerebellar peduncle interstitial nucleus of Cajal inferior olive interpeduncular nucleus medial lemniscus medial longitudinal fasciculus marginal zone nucleus reticularis pontis caudalis nucleus reticularis pontis oralis nucleus reticularis gigantocellularis nucleus reticularis paramedianus nucleus raphe pontis, pars ventralis nucleus reticularis supragigantocellularis nucleus reticularis tegmenti pontis oculomotor complex pontine nuclei prepositus hypoglossi pyramidal tract retroflex bundle rostral interstitial nucleus of the mlf red nucleus superior colliculus superior central nucleus superior cerebellar peduncle substantia nigra superior olive tractus solitarius trochlear nucleus motor nucleus of the trigeminal nerve spinal tract of the trigeminal nerve inferior vestibular nucleus lateral vestibular nucleus medial vestibular nucleus superior vestibular nucleus

sections), where almost all of the neurons were labeled retrogradely from the flocculus and very few from the lateral rectus muscle or the oculomotor complex. We have referred to this region elsewhere (Langer et al., ’85) as the rostral cap of the abducens or the preabducens interfasci- cular nucleus. It continues rostrally, among the fascicles of the mlf. Measurements of the distribution of cell size in the population of neurons labeled retrogradely from the flocculus indicated that they were on average significantly larger (609 & 354 ,urn2; N = 374) than the motoneurons and internuclear neurons, apparently due to a small population of very large neurons labeled from the flocculus.

The monkey’s abducens nucleus is subtly different in the region that contains floccular derents (Fig. 2C,D). The neuronal cell bodies are more angular because of the Nissl substance extending further into their dendrites, but they are about as chromatophilic and on average about the same size as the other neurons in the nucleus. There is, however, a small population of very large neurons in this region that is apparently not present in the other parts of the abducens nucleus. The neurons within the parts of the abducens nucleus containing floccular afferents are similar to those among the paramedian tracts and in the contiguous raphe nuclei. They differ markedly from the cells in the nearby rostral supragenu nucleus, which also projects very heavily to the flocculus.

Thus, the cytoarchitectonic heterogeneity of the monkey’s abducens nucleus is apparently due to the differential distribution of at least three populations of neurons. The neurons that project to the flocculus occupy a region of the abducens nucleus complementary to the region occupied by the motoneurons and internuclear neurons. These floccular afferents are continuous with similar populations among the fascicles of the paramedian tracts and in the adjacent raphe. The internuclear neurons occupy only a portion of the region occupied by motoneurons. Cytoarchitectonically, the regions with internuclear neurons in series A (Fig. 1) are the regions that are more densely cellular (Fig. 2). The regions occupied by only motoneurons are more open, with larger neurons that stain more intensely and have more discrete Nissl bodies. A small number of motoneurons appear to lie below the conventional boundary of the abducens nucleus, within the subjacent reticular formation (Fig. 1, series B).

Afferents to the abducens nucleus in the monkey The following descriptions of the sources of afferents to

the abducens nucleus in the monkey are based primarily on the four cases whose injection sites are shown in Figure 3. For each case the section shown is through the maximal cross section of the injection site and the extent of the injection was judged as being between two estimates, indicated by the two densities of screen in Figure 3. The area covered by the denser screen is an estimate based largely on the distribution of dense, diffusely distributed label, and is probably the approximate extent of the effective injection site. The area covered by the lighter screen is a generous estimate based on the distribution of flair about the dense core of label. As can be seen in the photomicrograph of the injection site in MoHAb (Fig. 4), it is often difficult to put precise boundaries on injection sites in TMB-treated material.

We considered the possibility that HRP extended into the structures surrounding the abducens nucleus, but generally were able to rule out the possibility on the basis of one

AFFERENTS TO THE ABDUCENS NUCLEUS

MoHAb'

MoHAb 5

Fig. 3. Injection sites in the four monkeys used in this study. The area covered with dense screen is our estimate of the minimal extent of the effective injection site. The area covered by light screen is our maximal estimate.

or another observation. For example, consider the injection site for MoHAb-4 that is illustrated in Figure 4. The reactivity in and above the genu of the facial nerve is almost entirely due to reactive glial cells. We frequently obtain the same reaction about the microelectrode tracks, when there is no HRP involved. There were no labeled neurons in the facial motor nucleus. There was filling of glial and axonal processes in the lateral part of the abducens nucleus and in the subjacent reticular formation, but this label does not appear to be sufficient to label neurons retrogradely. In the illustrated section, microscopic examination of the dorsomedial pontine reticular formation immediately subjacent to the noninjected abducens nucleus revealed many retrogradely labeled neurons, presumably IBNs (see below). If the injection extended into the region ventromedial to the injected abducens nucleus, then one would expect to find retrograde labeling of known atTerents to the IBNs. Our anatomical studies in the cat (Langer and Kaneko, '83) and monkey (unpublished observations) and physiological stud-

Fig. 2. Cytoarchitecture of the abducens nucleus in a series of transverse sections of a normal monkey brainstem. The sections are in caudal (A) to rostra1 (D) order. These sections are representative of the abducens nucleus in the monkey, and substantiate the cytoarchitectonic variation between the different regions of the monkey's abducens nucleus. Compare with distributions in Figure 1. A detailed description is given in the text. Cali- bration bar = 1 mrn.

385

ies in the cat (Curthoys et al., '81) have shown that the region containing IBNs is densely innervated by omnipause neurons (OPNs), which lie in the raphe between the rootlets of the abducens nerve. In the monkey, we have both physiological and anatomical evidence (unpublished observations; Biittner-Ennever and Pause, '85) that the cell bodies of the great majority of OPNs are in the distinctive, well- differentiated, bilaminate raphe nucleus (NRPv in Fig. 2) between the rootlets of the abducens nerve. The nucleus that contains the OPNs is the ventral part of a collection of raphe nuclei that appear to correspond to nucleus raphe pontis in the cat (Taber et al., '60). When the dense core of label in an abducens nucleus injection site extends into the part of the dorsomedial reticular formation, which presumably contains IBNs, there are labeled neurons in this ventral subdivision of the nucleus raphe pontis (NRPv); therefore, the presence of labeled neurons in the NRPv indicates that the effective injection site may include IBNs. Only in MoHAb-2 was there a very faint labeling of many of the cells in NRPV, but no other neuronal populations labeled in MoHAb-2 that were not labeled in other cases.

In none of the cases was there perceptible involvement of the vestibular nuclei or the supragenu nuclei. The fascicles in the medial margin of the abducens nucleus provided an effective barrier to the medial spread of label, though in MoHAb-2 and MoHAb-4 there was some infiltration of label into the neuropil between the fascicles, where one might

386 T. LANGER ET AIL.

Fig. 4. Photomicrograph of the injection site in MoHAb-4, the brain from which the illustrated series was drawn. The four white arrows point to the presumed injection site. While reactive material was present throughout the nuclear cross section and in the medial part of the subjacent reticular formation, most of it was clearly due to labeled axons. Dense reactivity not contained in axons was restricted to the medial margin of the nucleus (arrows) adjacent to the paramedian tracts. This was also judged to be the locus of the effective injection site based on the populations of retrogradely labeled neurons seen in this and other brains. Calibration bar = 1 mm.

expect to involve floccular af€erents. We have placed injections of HRP into the raphe between the abducens nuclei (work in progress) and obtained patterns of labeled neurons very different from those present in the cases described in this paper; therefore, there was probably no effective involvement of the raphe nuclei in the present cases. With the possible exception of MoHAb-2, in which some IBNs may have been within the effective injection site, there did not appear to be an involvement of the subjacent reticular formation. Each of the injections appeared to involve only a small fraction of the abducens nucleus.

The injections in MoHAb-1 and MoHAb-2 were made via a caudal approach, through the fourth ventricle and obliquely into the brainstem. The micropipette passed through the lateral margin of the nucleus prepositus hypoglossi to enter the abducens nucleus just lateral to the ascending limb of the facial nerve. Placement of the HRP at the point of maximal antidromic field potential, following electrical shocks to the lateral rectus muscle, resulted in an injection centered in the caudal part of the nucleus. There was apparently no involvement of any other oculomotor region, except possibly the medial supragigantocellular reticular formation in MoHAb-2.

MoHAb-4 had been used previously for chronic electrophysiological studies of the abducens nucleus and surrounding regions, and the location of abducens neurons was known from recording spike-trigger-averaged abducens motoneurons (&udder and Fuchs, '81). Our approach was tilted 20" lateral to the sagittal plane, thus avoiding involvement of the nucleus prepositus hypoglossi. The injection in MoHAb-5 was made via a vertical approach. The comparatively small number of labeled cells suggests that the effective injection site was very small.

The injections in MoHAb-1 and MoHAbB were centrally placed within the caudal part of the nucleus, while the injections in MoHAb-2 and MoHAb-4 were in the medial margin of the nucleus, adjacent to the paramedian tracts. These differences, along with the intrinsic heterogeneity of the abducens nucleus described above, may account for some of the variability in the labeled populations in the different cases as seen below.

Interstitial nucleus of Cajal and rostra1 interstitial nucleus. The rostralmost labeled neurons were found in a moderate projection from the interstitial nucleus of Cajal and the contiguous rostra1 interstitial nucleus of the mlf (RDT) (Fig. 5); their presence was surprising since they had not been reported previously and we did not find such a projection in the cat series (see Afferents to the Abducens Nucleus in the Cat section of the Results). It was prominent in only two cases, MoHAb-2 and MoHAb-4. A few labeled cells were seen in MoHAb-1, but none in MoHAb-5. The medial placement of the injections in the two animals with these projections suggest that these interstitial nuclei project to the floccular afferent neurons.

We have considered the possibility that these neurons were labeled by uptake of HRP by fibers passing through the abducens nucleus, but feel that this is unlikely because, when the interstitial populations were labeled, there were about as many neurons labeled as in the oculomotor corn- plex (Fig. 5). In MoHAb-4 there was no direct involvement of the vestibular nuclei nor, apparently, of the region that contains IBNs (see above). The labeled interstitial neurons were about the same sizes as their unlabeled neighbors. They tended to be clustered in certain parts of the interstitial nuclei and most were ipsilateral to the injected abducens nucleus, though many aIso occurred in the contralateral interstitial nucleus of Cajal (Fig. 5D-F).

Oculornotor complex. Certain parts of the oculomotor complex were densely labeled in all the brains. The majority of the labeled neurons occurred in regions that contained a fine, granular, anterograde label, but were not uniformly distributed in those regions. Most of the retrogradely labeled neurons were intercalated among the fascicles of the mlf (Fig. 9A) and in the lateral margins of the anterogradely labeled area; the great majority were contralateral to the injection, but a small number occurred in mirror symmetric positions in the ipsilateral nucleus. The distribution of these neurons is consistent with most, of them lying within the medial rectus subdivision of the oculomotor complex (Buttner-Ennever, '81). A small number of labeled neurons were seen elsewhere in the oculomotor complex and in the trochlear nuclei.

Superior colliculus. The contralateral superior coll icu- lus was labeled in all cases; the label always occurred in the intermediate gray, and the labeled neurons were intermediate in size (Fig. 6). There was, on average, about one labeled neuron per section. There was no definite preference for any rostrocaudal region of the superior colliculus to be labeled. Occasional neurons were labeled in contiguous parts of the pretectum.

Nucleus reticularis tegmenti pontis. There was a small collection of retrogradely labeled neurons scattered in the dorsal margin of the ipsilateral nucleus reticularis tegmenti pontis, primarily medially (Fig. 6). A few neurons were labeled more ventrally, within the body of the nucleus, or in the medial part of the contralateral nucleus. The labeled neurons were generally large, multipolar cells, approximately the same sizes and shapes as the unlabeled


B x 4

" J

E x4

x 3 F x 4

Mesencephalon

Fig. 5. Distribution of retrogradely labeled neurons in the mesencephalon, including the rostral interstitial nucleus of the mlf @IN), the interstitial nucleus of Cajal (INC), and the oculomotor complex (OMC) of MoHAb- 4. Labeled neurons were scattered through a large part of the ipsilateral RIN (A-C); only a few cells were labeled contralateral to the injection (B). The labeled populations in the OMC and the INC were bilateral. However, the labeled neurons were denser ipsilaterally in the INC (D-F) and contra-

neurons in the vicinity. There was a smooth transition between this population in the nucleus reticularis tegmenti pontis and the scattered labeled neurons in the overlying part of the nucleus reticularis pontis oralis (Fig. 6 0 .

Nuclei reticularis pontis oralis et caudalis. One of the larger pools of retrogradely labeled neurons in the brainstem rostral to the abducens nucleus was a dispersed population of neurons in the dorsomedial nucleus reticularis pontis caudalis (Fig. 6E) and the ventromedial part of the nucleus reticularis pontis oralis (Fig. 6C,D), where it extends over the dorsum of the nucleus reticularis tegmenti

G x 3

x 4

I x 4

laterally in the OMC (E-I), particularly in the lateral margin of the medial rectus motor nucleus and among the adjacent fascicles. Small numbers of retrogradely labeled neurons were present in other parts of the OMC. In this and following figures the arabic numerals beside the sections indicate the number of histological sections that were combined in the illustrated section. This and all subsequent charts of the monkey brainstem are for MoHAb-4.

pontis. It was not apparent, on the basis of the neuronal morphoIogy, that these were separate groups of neurons (Fig. 6). The labeled neurons were sparsely distributed and the transitions so gradual that it was difficult to draw boundaries, even though there was a shift from the ventromedial quadrant to the dorsomedial quadrant of the tegmentum as one moved caudally. The regions that were labeled were characterized by their small- to intermediate- sized multipolar neurons, and differed from the laterally adjacent rostra1 pontine reticular formation, which contains some very large neurons. The neurons retrogradely


Fig. 6. Distribution of retrogradely labeled neurons in the nucleus reticularis tegmenti pontis (A-C), nucleus reticularis pontis oralis (C, D), and nucleus reticularis pontis caudalis (El of MoHAb-4. All the cells in the superior colliculus are plotted together in section D. A population of small neurons was labeled in the dorsal gray in this series. but not in other series.

labeled from the abducens nucleus were medium and small multipolar cells.

Most of these rostral pontine afYerents were located rostral to the rootlets of the abducens nerve, but a few labeled neurons were aIso seen among the rootlets (Fig. 9E) and a few occupied the parvicellular reticular formation at the same rostrocaudal levels (Figs. 6E, 7A). Almost all of the retrogradely labeled neurons in the rostral pontine reticular formation were ipsilateral to the injected abducens nucleus; however, there were a few labeled neurons in a restricted region of the contralateral reticular formation a short distance rostral to the rootlets of the abducens nerve Fig. 6D,E).

Dorsal grug of the pons. In the illustrated case (MoHAb- 4) and in MoHAb-1 there were a few small neurons labeled in a restricted region of the dorsal gray, between the mlf and the ventricular surface (Fig. 6E). In the illustration, the superimposition of several sections gives this region the appearance of a substantial projection, but in the histological sections it was clear that these neurons constituted a very small fraction of the neurons in the nucleus, which is composed of small, densely packed neurons. This region contained no neurons in MoHAb-2, and was destroyed by the micropipette in MoHAb-5. Owing to its capriciousness, we cannot definitely conclude that this is a source of abdu- cells derents.


Caudal Pons and Rostra1 Medulla

Fig. 7. Distribution of labeled neurons in the pontine (A-C) and medullary (D, E) reticular formation, the vestibular complex (I-F), the marginal nucleus (D), and the nucleus prepositus hypoglossi (D, El of MoHAb4. The asterisk indicates the supragigantocellular reticular formation population. The combined population of the ventrolateral and rostra1 medial vestibular nuclei is shown in C. See the text for details.


Fig. 8. Photomicrograph of a Nissl-stained transverse section through the marginal zone (M) and the supragigantocellular reticular formation (NRS). Note the marginal zone lies between the medial vestibular nucleus (Vbm) and the nucleus prepositus hypoglossi (Pph). The supragigantocellular reticular formation lies dorsal to the gigantocellular reticular formation (NRG) and ventral to the nucleus prepositus hypoglossi. The medial, more densely cellular, part of the NRS contains IBNs. Calibration bar = 0.5 mm.

Supragigantocellular reticular formation. Most of the retrogradely labeled neurons in the reticular formation caudal to the abducens nucleus were in the dorsomedial part of the caudal pontine reticular formation and rostral medullary reticular formation, contralateral to the injection (Fig. 7B-E, asterisks). The part of the reticular formation occupied by these retrogradely labeled neurons is dorsal to the gigantocellular region and extends from the rootlets of the abducens nerve to the nucleus reticularis paramedianus. Because this region is cytoarchitectonically similar throughout, and to provide a shorthand for the following description, we call it the supragigantocellular reticular formation. The part of this region within the medulla cor- responds to the nucleus reticularis paragigantocellularis dorsalis (Olszewski and Baxter, '54) and the rostral part, in the caudal pons, is a part of the caudal nucleus reticularis pontis caudalis (Olszewski and Baxter, '54; Brodal, '57; Taber, '61).

Retrogradely labeled cells were almost entirely confined to the medial part of the supragigantocellular region (Fig. 7B-E) and were predominantly intermediate-sized multipolar neurons (Fig. 9F). Following a large injection into the abducens nucleus, about half of the neurons in the rostral part of the labeled region contained HRP, but the density declined in progressively more caudal sections until only a few neurons were labeled in the sections through the rostral pole of the paramedian reticular nucleus. The supragigantocellular reticular formation tends to be more cellular and less fascicular medially, and it is this medial part that contains virtually all of the retrogradely labeled neurons. The most rostral of the labeled neurons, just caudal to the most caudal rootlets of the abducens nerve, were immediately ventromedial and medial to the caudal pole of the abducens nucleus (Fig. 7B). More caudally, the labeled neurons occurred in the medial supragigantocellular region, below the ascending limb of the facial nerve (Fig. 7C), and they continued to occupy a similar medial position below

the nucleus prepositus hypoglossi (Fig. 7D), extending back to abut upon the paramedian reticular nucleus. There was an occasional labeled neuron in the lateral part of the supragigantocellular region or among the fascicles of the paramedian tracts in the medial boundary of the supragigantocellular region.

Although the caudal part of the supragigantocellular projection was contiguous with a labeled population in the nucleus prepositus hypoglossi (Fig. 7D,E), the retrogradely labeled cells in the supragigantocellular reticular formation had about twice the average cross-sectional area of those in the nucleus prepositus hypoglossi. Very few labeled neurons were seen elsewhere in the reticular formation caudal to the abducens nucleus. A few labeled neurons were found in the ipsilateral supragigantocellular reticular formation (Fig. 7D), primarily caudal and ventral to the contralateral population. An occasional cell was labeled in the ipsilateral nucleus reticularis paramedianus (Fig. 7D,E).

Vestibular nuclei and nucleus prepositus hypoglossi. One of the densest groups of retrogradely labeled neurons was in the ventrolateral vestibular nucleus and adjacent parts of the medial vestibular nucleus, bilaterally (Fig. 7C). A second, comparably dense group of labeled neurons was contralateral to the injected abducens nucleus in a small nucleus in the common margin of the medial vestibular nucleus and the nucleus prepositus hypoglossi. We have called this nucleus the marginal zone (M in Fig. 7D) because it links the nucleus prepositus and the medial vestibular nucleus, but is cytoarchitectonically differentia- ble from both (Fig. 8). In other studies in progress, we have found that this same region differs from the adjacent portions of the medial vestibular nucleus and the nucleus prepositus in the connections that it has with other oculomotor structures, and preliminary physiological studies suggest that it contains units of a particular functional type. We will consider each of these populations in some detail and finish with a brief survey of other retrogradely labeled neurons in the vestibular complex.

The marginal zone and the nucleus prepositus hypoglossi. One of the most numerous and most densely packed collections of labeled neurons was the group in the marginal zone, which links the medial vestibular nucleus and the nucleus prepositus hypoglossi over much of their common boundary (Fig. 7D). Following an injection in the abducens nucleus, the great majority of the neurons in the contralateral marginal zone were retrogradely labeled. A small number were also labeled in the ipsilateral marginal zone. The neurons in the marginal zone are intermediate to large in size, and are fairly uniform in shape and staining properties (Fig. 9D).

The density of labeled neurons falls off very rapidly bloth medial and lateral to the marginal zone (Fig. 7D). Thiere were small, labeled neurons scattered throughout the rostrocaudal extent of the nucleus prepositus hypoglossi, but they were somewhat more frequent in its ventral margin, where it is traversed by a mediolaterally coursing band of axons. Lateral to the marginal zone, in the medial vestibular nucleus, there were substantial numbers of retrogradely labeled neurons scattered through the nucleus, but they constituted a very small fraction of the neurons present. They were also morphologically different from the cells in the marginal zone (see below).

Ventrolateral vestibular nucleus and rostral medial vestibular nucleus. There was a prominent population of retrogradely labeled neurons in a region that surrounds the

Fig. 9. Photomicrographs of retrogradely labeled neurons in several populations. A. Labeled neurons among the fascicles of the mlf and in the lateral margin of the oculomotor complex. The arrows point to some of the labeled neurons. B. Neurons labeled in the rostra1 pole of the ipsilateral medial vestibular nucleus. The arrows indicate some of the less densely labeled neurons. C. Labeled neurons in the contralateral ventrolateral vestibular nucleus at the level of the abducens nucleus. The extent of the nucleus is indicated by the triangles. D. Labeled neurons in the contralat-

era1 marginal nucleus. Some of the less densely labeled neurons are indicated by the arrows. E. Labeled neurons in the reticular formation ventral to the injected abducens nucleus. The retrogradely labeled neurons are indicated by arrows; r6 indicates the rootlets of the abducens nerve, which were filled with label. F. Labeled neurons in the contralateral supramagnocellular reticular formation. Some of the less densely labeled neurons are indicated by ari-ows. Calibration in C = 300 p m for A. B. and D. and 760 pni for C. l?, and F


rostroventral aspect of the lateral vestibular nucleus (Fig. 7C). Many of these neurons occupied a part of the vestibular complex that has been called the ventrolateral vestibular nucleus (Maciewicz et al., '77), and the remainder were in the adjacent ventrolateral part of the rostral pole of the medial vestibular nucleus. There were some morphological differences between these two collections of labeled neurons, but they seemed to form a confluent collection of neurons.

The labeled neurons in the ventrolateral vestibular nucleus were in the dorsal part of the nucleus, adjacent to the very large multipolar neurons in the lateral vestibular nucleus (Deiter's nucleus). The labeled neurons were intermediate- to large-sized multipolar neurons, typical of the nucleus (Fig. 9C). Following a large injection in the abducens nucleus, most of the neurons in this region of the ventrolateral vestibular nucleus were labeled-possibly all of the neurons if both ipsilateral and contralateral groups are combined.

The projection from the ventrolateral vestibular nucleus to the abducens nuclei was bilateral and varied from case to case, with the contralateral population tending to be larger. The ipsilateral population was frequently difficult to assess since the faintly labeled cells in the ipsilateral ventrolateral vestibular nucleus were often obscured by the halo of the nearby injection site.

Many of the other labeled neurons in this region occupied the part of the medial vestibular nucleus that is contiguous with the ventrolateral vestibular nucleus (Fig. 7C). There were also some small foci of labeled neurons even more medial in the rostral pole of the medial vestibular nucleus. The labeled neurons in the medial vestibular nucleus (Fig. 9B) were smaller than those in the ventrolateral vestibular nucleus, as might be expected from the smaller average size of the neurons in the involved parts of the medial vestibular nucleus. These medial vestibular populations were distributed more or less symmetrically on the two sides.

Intermediate levels of the medial vestibular nucleus. For a considerable distance caudal to the ventrolat- era1 vestibular nucleus, within the medial vestibular nucleus, there were labeled neurons scattered through much of the nuclear cross section (Fig. 7D-F). Most of these neurons were at about the same rostrocaudal levels as the marginal zone, but there were occasional retrogradely labeled neurons even in the most caudal sections through the vestibular complex. There was considerable variety in the size and shape of these neurons; some were large and chromatophilic, while others were small. Two additional small groups of labeled neurons were seen bilaterally in the y- group (Fig. 7D,E) and in the ipsilateral superior vestibular nucleus (Fig. 7B,C). They did not appear to be remarkably different from the unlabeled neurons in their vicinity. An occasional labeled neuron was seen in the inferior vestibular nucleus. None were seen in the lateral vestibular nucleus.

Afferents to the abducens nucleus in the cat The following descriptions of the m e r e n t s to the abdu-

cens nucleus in the cat are based primarily on two cases, CHAb-1 and CHAb-3. In both of these cases the micropipette passed through the fourth ventricle and obliquely into the brainstem. Because of the slightly different anatomy of the cat brainstem there was much less likelihood of involving the nucleus pr-epositus hypoglossi than in the monkey

brain. In both cases the injection was placed in the center of the abducens nucleus (Fig. 10). Comparison of the extent of the reactivity in the injected abducens nucleus with the distribution of retrogradely labeled neurons in the contralateral supragigantocellular region suggests that the injection did not involve the supragigantocellular reticular formation. There was no label deposited in the nucleus prepositus hypoglossi. The illustrated sections were drawn from CHAb-3.

The mesencephalon. The labeled neurons in the mesencephalon were confined almost entirely to the superior colliculus, the oculomotor complex, and the central gray immediately dorsal tc, the oculomotor complex. The cells dorsal to the oculomoto; complex were the most numerous.

Superior colliculus. The neurons labeled in the superior colliculus (Fig. 11A,B) were large and very large chromatophilic multipolar cells confined almost entirely to the intermediate gray layer of the colliculus contralateral to the injection. An occasional cell occupied the deep gray layer. Unlike those in the monkey, the labeled neurons occurred more frequently in the central part of the collicu. lus and were particularly dense rostrally, but were absent from the rostral and caudal poles. It is of interest that the most densely labeled region was the part concerned with. central vision and the horizontal meridian, but the small size, loose organization, and variability of this population from case to case suggest that it is a comparatively minor input to the abducens.

Oculomotor complex. Labeled neurons were distributed. bilaterally in parts of the oculomotor complex, with approximately the same density on the two sides (Fig. 11A,B). The labeled neurons within the oculomotor complex were generally smaller than the large chromatophilic neurons that define the nucleus and that are presumably the m o toneurons. There were no retrogradely labeled neurons in the rostral third of the oculomotor complex. In the middle third (Fig. llA), small numbers of labeled neurons were dispersed among the various subnuclei. In the caudal part of the complex (Fig. 11B) the labeled neurons were primar-- ily in the overlying central gray, but there were also many among the oculomotor neurons.

Since the great majority of the retrogradely labeled neurons do not lie within the cytoarchitectonically defined oculomotor complex, as was previously found by Maciewicz et al. ('751, we hesitate to call them oculomotor internu-. clear neurons, although they may serve in the same capac- ity as the retrogradely labeled neurons with the medial. rectus nucleus (Biittner-Ennever, '81; see also Results, Af- ferents to the Abducens Nucleus in the Monkey).

In the central gray dorsal to the caudal pole of the oculo.. motor complex there is a densely populated, bilateral band of retrogradely labeled neurons that extends from the mid- line to the lateral boundary of the central gray. There were notably fewer labeled neurons ipsilateral to the injection. This densely labeled central gray region is apparent in Nissl-stained sections as a poorly defined nucleus within the central gray, which melds with the dorsolateral margin of the posterior oculomotor complex. A large fraction of its cells project to the contralateral abducens nucleus and many project to the ipsilateral abducens nucleus. The l a beled cells were generally small- to intermediate-sized,, multipolar neurons, with a tendency to be oriented parallel to the dorsal boundary of the oculomotor complex. As the transition between the oculomotor complex and this nucleus is not abrupt, it is not always possible to assign


Fig. 10. Injection sites for the two cats used in this study. The denser screen indicates our estimate of the effective injection site and the lighter screen, the extent of the flair. Dots show locations of individual labeled neurons in the supramagnocellular and medial vestibular populations.

labeled cells unambiguously to one nucleus or the other. A few retrogradely labeled neurons were dispersed among the fascicles of the mlf and, occasionally, in the reticular formation ventrolateral to the mlf.

The pons. Labeled neurons were scattered through the medial part of the nucleus reticularis pontis oralis and in the dorsomedial quadrant of the part of the nucleus reticularis pontis caudalis that lies rostra1 to the abducens nuclei (Fig. 11C,D). Most of the neurons labeled in the pontis oralis were intermediate-sized cells confined to its medial region, dorsal to the nucleus reticularis tegmenti pontis. The cells retrogradely labeled in the pontis caudalis were confined to its dorsomedial region, adjacent to the paramedian tracts, but generally separated from the ascending tract of Deiters. At all levels the retrogradely labeled neurons were a small fraction of the neurons present in their vicinity. Unlike the monkeys, the cats had no labeled cells in the nucleus reticularis tegmenti pontis.

Pontomedullary groups. As in the monkey, the great majority of the labeled neurons in the cat’s brainstem were in the vestibular complex, the nucleus prepositus hypo-

glossi, or in the supragigantocellular reticular formation. The general pattern was similar in the two species.

Supragigantocellular reticular formation. As in the monkey, there was a region of the reticular formation dorsal to the gigantocellular region that differed most ob- viously in lacking the very large cells characteristic of the gigantocellular region. Within the medial supragigantocellular region (Fig. 11E-H), labeled neurons were most frequent in the contralateral reticular formation medial and ventral to the caudal pole of the abducens nucleus. There may even be some intermingling of these neurons with neurons of the abducens nucleus. In more caudal sections the labeled neurons were found in the medial reticular formation below the ascending limb of the facial nerve. Caudally, this population of intermediate to large multipolar neurons was contiguous with the population of smaller labeled cells in the ventral margin of the nucleus prepositus hypoglossi (Fig. 11F-HI.

Nucleus prepositus hypoglossi. There does not appear to be a marginal zone in the cat brainstem, but there was a dense labeled population of small neurons in the ventral


Fig. 11. Distribution of retrogradely labeled neurons in CHAb-3. The screen over the abducens nucleus indicates the injection site. Each dot represents one retrogradely labeled neuron. The numbers under the section letters indicate the number of histological sections that were compressed into the illustrated section.


margin of the contralateral nucleus prepositus hypoglossi (Fig. 11F-4. These labeled neurons were largely confined to the fascicular ventral margin of the nucleus prepositus hypoglossi. The overlying, more cellular, part of the nucleus prepositus had many fewer labeled neurons. Small numbers of labeled neurons occurred in the medial margin, intermingled with the fascicles of the adjacent mlf. In the caudal part of the nucleus there was a more uniform distribution of labeled neurons throughout the nucleus bilaterally. Most caudally, some labeled cells were scattered in the rostral pole of the nucleus intercalatus. No labeled neurons were seen in the nucleus of Roller.

Vestibular complex. The greatest numbers of labeled neurons were in the vestibular complex, principally in the ventrolateral vestibular nucleus and the rostral pole of the medial vestibular nucleus (Fig. 11E,F). A small number of labeled cells occurred in the superior vestibular nucleus of CHAb-3 (Fig. 11D), but only an occasional neuron was labeled there in CHAb-1.

The dense, bilateral populations of labeled neurons in the ventrolateral vestibular nucleus and contiguous parts of the rostral pole of the medial vestibular nucleus tended to be more numerous ipsilateral to the injection. The most rostral labeled neurons were seen in the extreme rostral pole of the medial vestibular nucleus adjacent to the abducens nucleus (Fig. 11E,F). Slightly more caudally they were joined with those in the ventrolateral vestibular nucleus to form a ban? just dorsal to the ventral margin of the vestibular complex (Fig. 11F). The labeled cells in the medial vestibular nucleus were generally small, while those in the ventrolateral vestibular nucleus were generally intermediate-sized cells.

Just caudal to the ventrolateral vestibular nucleus, the labeled neurons dispersed to fill most of the cross section of the medial vestibular nucleus ipsilateral to the injection (Fig. 11G,H) and some labeled cells were found even in adjacent parts of the inferior vestibular nucleus (Fig. llH,U. Contralateral to the injected abducens nucleus, there were comparatively few labeled neurons. Further caudally in the medial vestibular nucleus, caudal to the cerebellomedullary junction, there were about equal numbers of labeled cells on the two sides. The frequency of labeled neurons was much less than was seen rostrally, and there was a gradual decline in frequency in progressively more caudal sections until only occasional cells were labeled in the most caudal sections through the medial vestibular nucleus. Almost all of the cells labeled in the medial vestibular nucleus were small.

DISCUSSION The afferents to the abducens nucleus originate in a lim-

ited number of distinct brainstem regions. The general regions involved in the cat and the monkey are similar, but frequently the morphology and detailed distributions of the involved neurons were rather different. For many of these regions, specific functions have been attributed to the neurons on the basis of their anatomical connections andor physiology. We will consider each retrogradely labeled region by reviewing and comparing results in the cat (Ma- ciewicz et al., '75, '77; Gacek, '79; Baker and Spencer, '81; Stanton and Greene, '81; this paper) with our findings in the monkey. Where possible we will consider the potential functional role of each afferent population.

Rostra1 interstitial nucleus of the rnlf and interstitial nucleus of Cajal

The most rostral abducens afferents in the monkey brainstem were from the rostral interstitial nucleus of Cajal. We were surprised by the labeling of these neurons, since they were not labeled in the cat either in our series or in those previously reported. Consequently, we considered the possibility that they might be due to the spread of label into the vestibular nuclei (Steiger and Buttner-Ennever, '79) or the nucleus prepositus hypoglossi (McCrea et al., '79), or that they might be due to the uptake of HRP by damaged axons in the mlf or other paramedian tracts. It is improba- ble that the neurons were labeled owing to involvement of the vestibular nuclei since the most effective injection sites for labeling these interstitial nuclei were in the medial part of the abducens nucleus and neither injection appeared to involve the vestibular nuclei. While one of the injections that labeled these nuclei was made by passing the micropipette through the rostral pole of the nucleus prepositus hypoglossi, the most effective injection was made by using a rostral approach that completely avoided the nucleus prepositus. Although we cannot say with certainty, we do not believe that the labeled neurons were back-filled from cut axons because (1) their distribution within the nuclei, the granularity of their label, and the distribution of label within the individual neurons were similar to the patterns of labeling of the neurons in the other regions; (2) there was minimal damage to the axons of the paramedian tracts since only the micropipette tip passed among them and there were no labeled cells in the facial nucleus despite more damage in the genu caused by the shank of the micropipette; and (3) the most effective injection was placed more laterally and dorsally, lateral to the dense fascicles of the paramedian tracts. Consequently, we are led to consider that the labeled neurons in the rostral interstitial nucleus of the mlf and the interstitial nucleus of Cajal may well be valid afferents to the abducens nucleus.

If they are truly projecting to the abducens nucleus, why don't these cells always take up label? The answer to this question seems to lie in the placement of the injections relative to the various neuronal populations within the abducens nucleus. The medially placed injections that were effective in labeling the interstitial neurons maximally involved the flocculus afferents and minimally involved the internuclear neurons. The motoneurons were involved in all these cases. It is possible that the rostral interstitial nuclei are projecting to a subset of the abducens neurons that probably includes the floccular afferents.

It is puzzling that these structures associated with vertical eye movements should project into the abducens nucleus, which is associated with horizontal eye movements. We have comparatively little information about the physiology of the interstitial nucleus of Cajal in the monkey, but in the cat its activity is related to the vertical vestibular- ocular reflex (Anderson et al., '79; King et al., '80). Simi- larly, it is known that the rostral interstitial nucleus of the mlf participates in vertical eye movements (King and Fuchs, '79; Buttner et al., '77; Graybiel, '77). By contrast,~dI abducens neurons are horizontal burst-tonic neurons (see Fuchs et al., '85, for a review). There are two factors that should be considered. First, even when most prominent, this is a numerically minor projection. It would be expected to have a small effect, compared with the vestibular, marginal, and supragigantocellular afferents. Second, there has

3%

been no concerted effort to look for vertical eye movement sensitivity in abducens neurons. It is worth noting that there is also a small but consistent population of internuclear neurons in the oculomotor complex that lie outside the medial rectus subdivision and therefore may also trans- mit information about nonhorizontal eye movements.

The absence of labeled neurons in these interstitial nuclei in the cat brainstem is in agreement with the results of previous anatomical (Maciewicz et al., '75, '77; Gacek, '79; Baker and Spencer, '81; Stanton and Greene, '81) and physiological (Schwindt et al., '74) studies.

Oculomotor complex While both the cat and the monkey have afferent projec-

tions from the oculomotor complex or its immediate vicinity, the distribution of these neurons is very different in the two species. In the monkey, there is a good correspondence between the location of many of the labeled internuclear neurons, the location of the medial rectus motoneurons, and the terminals of the abducens internuclear neurons. This relationship, which has been documented extensively by Buttner-Ennever ('811, suggests a reciprocal interconnec- tivity between those two pools of horizontal-eye-movement- related neurons that might coordinate their actions. It is curious that the internuclear neurons in the medial rectus motor nucleus occupy only the lateral margin of the nucleus, as do the abducens internuclear neurons; the purpose of such a differential distribution is obscure. No such ar- rangement has been described in the cat's abducens nucleus (Steiger and Biittner-Ennever, '78; Highstein et al., '82). It is not apparent what role might be played by the other retrogradely labeled neurons in the oculomotor complex. Many of the other labeled neurons were in the medial rectus subdivision ipsilateral to the injected abducens nucleus, but a fair number were in other subdivisions that are concerned with vertical eye movements.

The existence of neurons in and about the oculomotor complex that project to the abducens nucleus is well estab- lished in the cat, largely on anatomical grounds (Maciewicz et al., '75; Maciewicz and Phipps, '83; this paper). In contrast to the situation in the monkey, comparatively few of these cells lie in the oculomotor complex proper. Those that do do not show a marked preference for any subdivision though they do avoid the rostral third of the complex, which contains the medial rectus subdivision (Naito et al., '74).

The great majority of the neurons labeled in the cat were in an inconspicuous nucleus in the central gray immediately dorsal to the caudal pole of the oculomotor complex. Within that nucleus, a large fraction of the neurons project to the contralateral abducens nucleus and a smaller fraction to the ipsilateral nucleus. Between the two projections virtually every neuron projects to the abducens nucleus; some may project bilaterally (Maciewicz and Phipps, '83). We know nothing of the physiology of these neurons.

Superior colliculus Cut. The retrogradely labeled neurons in the cat's supe-

rior colliculus are among the largest cells present in the intermediate gray layer, but only a small subset of these characteristically very large, chromatophilic neurons were labeled even following a fairly large injection in the abducens nucleus. The retrogradely labeled neurons were primarily in the rostral central part of the colliculus, the part associated with the central visual field (Feldon et al., '70; Stein et al., '75), and matching somesthetic and auditory fields. These results confirm previous results of HRP studies as well as the prior physiological (Grantyn and Grantyn,

T. LANGER ET AL.

'76) and anterograde anatomical (Edwards and Henkel, '76) studies.

Grantyn and Grantyn ('82) showed that very large cells in the cat superior colliculus, such as those labeled by the abducens injections, have exceedingly extensive axon ar- bors, generally involving large portions of the reticular formation. It is therefore likely that the abducens nucleus is only one of many places that these cells innervate.

Monkey. The characteristics of the retrogradely labeled cells in the monkey's superior colliculus are remarkably different from those in the cat. The neurons themselves are intermediate in size, have a variety of shapes, and frequently are not particularly chromatophilic. Since the large, chromatophilic, multipolar collicular neurons are not present in monkey, it is possible that monkey colliculus lacks this extensively arborized cell type. If so, it may be that monkey collicular neurons innervate more restricted regions and possibly only the abducens nucleus. Even with large injections in the contralateral abducens nucleus, there were comparatively few retrogradely labeled neurons in the superior colliculus. Consequently, it is probable that the superior colliculus projection carries a weak signal from the intermediate gray of the monkey's superior colliculus to the abducens nucleus, as it does in the cat (Grantyn and Grantyn, '76).

Nucleus reticularis tegmenti pontis Certain regions of the nucleus reticularis tegmenti pontis

(NRTP) project very lightly into the ipsilateral abducens nucleus in the monkey, but not in the cat. To our knowl- edge, this projection has not been reported previously. In- terestingly, the region of the NRTP labeled in this stud!{ also projects to the flocculus (Langer et al., '85). It is probable that the NRTP innervates more neurons within the abducens nucleus than those that project to the flocculus, since NRTP neurons were labeled following all the injections in the abducens nucleus.

The labeled neurons in the NRTP were located near the region where saccade-related neurons have been recorded. Those neurons discharge a burst of action potentials for saccades of a particular amplitude and direction, and have been found rostral to (Keller and Crandall, '83) as well as caudodorsal to (Hepp and Henn, '83) the area where labeled neurons are found. Although it seems likely that there 1s some overlap between the anatomically and physiologically defined regions, labeled neurons constitute only a small fraction of the neurons present and therefore a correspondence between the labeled and saccade-related neurons is tenuous. We can say that other cells in the region that discharge for optokinetic stimuli (Keller and Crandall, '83) appear to be located more rostral than our retrogradely labeled cells.

Nuclei reticularis pontis oralis et caudalis Early anterograde tracing experiments in the cat (Butt-

ner-Ennever and Henn, '76; Graybiel, '77b), using tritiated amino acids, showed that the oral and caudal pontine reticular formations innervate the abducens nucleus. However, contemporary experiments that used small injections of HRP confined to the abducens nucleus failed to label neurons in this region of the cat (Maciewicz et al., '77; Gacek, '79). We have been able to demonstrate these projections in both the cat and the monkey, probably because we used a much more sensitive chromophore.

Our results were expected on the basis of a large body of evidence from anatomical, physiological, and ablation experiments (see Fuchs et al., '85, for a review) that strongly


that they provided monosynaptic inhibition in the contralateral abducens motoneurons, inhibiting them during saccades toward the side of the recorded B N . Recently, Scudder and colleagues ('82) demonstrated a similar collection of neurons in the monkey brainstem.

Rostrally, about 50% of all units encountered in the supragigantocellular region discharge in relation to eye movements, but all responsive units are horizontal burst neurons responding to ipsilaterally directed saccades (Scud- der and Fuchs, unpublished observations). At more caudal levels the encounter rate for burst neurons drops until they are seldom encountered ventral to the nucleus prepositus hypoglossi. These observations agree well with our anatomical observations, since in the region immediately ventral to the caudal pole of the abducens nucleus about half of the cells are labeled following a large injection, but the percent- age drops rapidly as one passes caudal to the ascending limb of the facial nerve. Since the IBNs and abducens derents are distributed similarly, it is likely that the retrogradely labeled cells we observed in this region were IBNs.

As far as we know, all IBNs are located contralateral to the abducens nucleus that they innervate, but in the monkey there is a small collection of retrogradely labeled neurons in the ipsilateral supragigantocellular reticular formation as well. These neurons lie further caudal and ventral than the contralateral population, and they may very well be another type of neuron.

Rostra1 medial and ventrolateral vestibular nuclei The most numerous and extensive collection of labeled

neurons in both species was in the vestibular complex, principally in the ventrolateraI vestibular nucleus and the contiguous rostral pole of the medial vestibular nucleus. Consequently, one might expect that these cells would have a very strong influence on activity in the abducens nucleus. Scudder and Fuchs ('81) have described a compact collection of units within the rostral pole of the medial vestibular nucleus and in the ventrolateral vestibular nucleus of the monkey that discharge in relation to eye position (tonic eye position sensitivity) and head velocity (vestibular sensitivity) and that cease firing during saccades (pause). Because of these properties they have been called tonic-vestibular- pause units, or "VP cells (Pola and Robinson, '78). TVP cells receive a monosynaptic input from the ipsilateral vestibular nerve and synapse upon contralateral abducens motoneurons (Scudder and Fuchs, '81); therefore they are interneurons in the three-neuron arc of the vestibulo-ocular reflex. TVP cells are not intermixed with other types of units that discharge with eye movements or with units that are unrelated to eye movements. Reconstruction of recording tracks has shown that TVP cells lie within the region of the vestibular complex that is most densely labeled following HRP injections in the abducens nucleus.

Experiments that would demonstrate TVP cells have not been done in the cat, but McCrea et al. ('80) and Ishisuka et al. ('80) have demonstrated vestibular units that receive a direct vestibular nerve input, code eye position in their discharge, and project to the abducens nucleus. The distribution of these units is approximately the same as the distribution of retrogradely labeled neurons in our cat series. It is therefore likely that this region of the vestibular complex is the most important link in the direct pathway for the horizontal vestibulo-ocular reflex in cat and monkey.

supports the existence of immediately premotor burst neurons in this area that participate in the generation of horizontal saccades. The numbers of labeled neurons are far greater in the monkey than in the cat, just as the frequency of encountering burst neurons is greater in the monkey (Kaneko and Fuchs, unpublished results). Although the labeled neurons were scattered through a substantial ex- panse of reticular formation in both species, in the monkey they were confined almost entirely to the region dorsomedial or medial to the magnocellular regions of the oral and caudal pontine reticular formation. In the cat the labeled neurons mingled with the large chromatophilic neurons that characterize the region.

The small group of retrogradely labeled cells rostral to the rootlets of the contralateral abducens nerve may be the occasionally nuschei and Fuchs, '72) or even frequently (Henn and Cohen, '76) recorded burst neurons in this region that discharge maximally for contralaterally directed saccades. It is interesting that in the cat there are no burst neurons that discharge for contralaterally directed saccades (Kaneko et al., '811, and we found no retrogradely labeled neurons in the rostral contralateral pontine reticular formation.

The scattered labeled cells in the central reticular formation about the rootlets of the ipsilateral abducens nerve in the monkey's brainstem form a substantial collection of neurons. Unfortunately, they are scattered so diffusely and broadly that it is unlikely that we will ever understand their role unless they are identified by antidromically acti- vating them from the abducens nucleus. These neurons have not been observed in the cat.

Supragigantocellular reticular formation Another population of pontine reticular afferents to the

abducens nucleus lies within the dorsomedial reticular formation immediately ventromedial and caudal to the caudal pole of the contralateral abducens nucleus, in the medial part of a region that we have called the supragigantocellular reticular formation. This region is composed of the nucleus paragigantocellularis dorsalis described by Olszewski and Baxter ('54) and a portion of their nucleus pontis caudalis in the dorsomedial part of the caudal pontine reticular formation, which they described separately but did not name. Our reasons for considering these two regions as a single unit are that they are cytoarchitectonically quite similar and there is a population of abducens afferents that appears to extend through the entire region. Other studies in our laboratory indicate that this region may have other connections that set it apart from the surrounding structures (unpublished studies). Of course, we have not demonstrated that all the labeled neurons in the contralateral supragigantocellular region are of the same functional type, but physiological studies in the cat (Hikosaka and Kawa- kami, '77; Hikosaka et al., '78) and the monkey (Scudder et al., '82) suggest that all the retrogradely labeled neurons may be inhibitory burst neurons (IBNs).

Maciewicz and colleagues ('77) were the first to demonstrate a population of retrogradely labeled neurons in this region of the cat brainstem following injections of HRP in the abducens nucleus. Concurrently, Graybiel ('77b) demonstrated the same connection using anterograde tracing methods. Others (Hikosaka et al., '78; Kaneko and Fuchs '81; Yoshida et al., '82) showed that units in the same region of the cat brainstem were horizontal burst neurons and

398

Intermediate levels of the medial vestibular nucleus Whereas the labeled cells were the majority of the neu-

rons in the ventrolateral vestibular nucleus, the labeled cells in the middle rostrocaudal levels of the medial vestibular nucleus were scattered and constituted a small fraction of the neurons present. The retrogradely labeled neurons were dispersed through most of the cross section of the ipsilateral nucleus at the more rostral levels. Curiously, there was a definite tendency for the ipsilateral and the contralateral afferents to occupy complementary regions of the medial vestibular nucleus, with the contralateral cells confined to the ventral margin of the nucleus.

Y-group A few retrogradely labeled neurons were present in the

y-group in the monkey, but not in the cat. As this was a consistent finding, the y-group probably provides a sparse afferent input to the abducens. The role of this input is obscure since the activity of y-group neurons reflects the combined head and eye movements in the vertical direction (Chubb and Fuchs, ’81). It is very unlikely that we will be able to see this signal in the discharge patterns of abducens neurons unless special conditions are imposed on the oculomotor system.

Subpopulations in the vestibular groups Even within some of the regions that we have considered,

there appeared to be small enclaves of labeled cells that were morphologically distinct. At this time, however, the physiological mapping of the vestibular complex has not been done in enough detail to justify our making such fine distinctions.

Marginal zone In the monkey, our injections have revealed an appar-

ently important new oculomotor nucleus. The marginal zone is densely labeled following HRP injections in the abducens and can be seen with ease in Nissl-stained material, though it appears to be a subdivision of the nucleus prepositus caudally and a subdivision of the medial vestibular nucleus rostrally. The great majority of its neurons project to the contralateral abducens nucleus, although a few project to the ipsilateral nucleus. In a monkey brain in which we placed a large injection in the rostral mesencephalic reticular formation, including virtually all of the vertical eye-movement-related regions, the marginal zone was conspicuous by the absence of retrogradely labeled cells (Langer and Kaneko, unpublished observations). The fact that this nucleus projects so intensely to the abducens nucleus and not at all to the vertical eye movement regions indicates that it transmits a signal related exclusively to horizontal eye movements. As yet there is no definite physiological evidence for this possibility, though exploratory recordings in this general region typically find horizontal burst-tonic units (Luschei and Fuchs, ’72; McFarland and Fuchs, unpublished observations). It appears that the cat does not have a marginal zone: such a nucleus cannot be identified cytoarchitectonically, nor is there a dense collection of labeled neurons in the border between the medial vestibular nucleus and the nucleus prepositus hypoglossi.

Nucleus prepositus hypoglossi Both cat and monkey have abducens afferents from the

nucleus prepositus hypoglossi. The retrogradely labeled cells were distributed throughout the rostrocaudal length

T. LANGER ET AL.

of the nucleus contralateral to the injection and, to a lesser extent, through the ipsilateral nucleus. In the monkey the labeled cells were denser bilaterally in the rostral part of the nucleus and there were very few labeled neurons in the caudal part of the ipsilateral nucleus. In the cat brainstem the labeled neurons in the rostral part of the nucleus were largely in the ventral margin of the nucleus, where it blends into the reticular formation. The central, more cellular, part was labeled less frequently despite the greater cell density. In the caudal part of the nucleus the distribution became more uniform and labeled neurons were found in all parts of the nuclear cross section. In the cat, the collection of retrogradely labeled neurons was particular1,y dense in the ventral, fascicular part of the rostral end of the nucleus prepositus hypoglossi, which may indicate that that region is the homologue of the marginal zone in th.e monkey.

In both cat and monkey, neurons with eye-position and eye-velocity sensitivity have been recorded in the prepositus (Lopez-Barneo et al., ’82) and the marginal zone (Lus- chei and Fuchs, ’72), respectively. The exact function fix this abducens input remains to be demonstrated.

CONCLUSIONS In both the cat and the monkey, the abducens nucleus

receives a variety of inputs. The major inputs are the same in the two species. Both receive a massive vestibular input, apparently from the interneurons of the vestibulo-ocul ar reflex. Both receive substantial innervation of the pontine reticular formation, presumably from excitatory and inhibitory burst neurons. Both receive what is probably a substantial horizontal burst-tonic input from the marginal zone and/or the nucleus prepositus hypoglossi. Both receive a substantial input from the oculomotor complex or its ’immediate vicinity, and both receive a small input from the superior colliculus.

On the other hand, the monkey abducens nucleus receives afYerents from additional structures, additional regions within the same structures, and usually more neurons from any particular structure. For example, in the monkey’s rostral pontine reticular formation there were many more retrogradely labeled neurons distributed over a larger volume of the reticular formation, and there were many more labeled cells in the monkey’s marginal zone and nucleus prepositus hypoglossi than in the cat’s nucleus prepositus. A number of the regions with small projections to the monkey’s abducens nucleus (rostral interstitial nucleus of the mlf, interstitial nucleus of Cajal, y-group, superior vestibular nucleus, and parts of the oculomotor complex) contain units related to vertical eye movements. Many of’ the larger populations of retrogradely labeled neurons in the monkey’s brainstem had small enclaves of strikingly drffer- ent morphological types, which may correspond to as-yet- unappreciated functionally different unit types. All of these observations, taken together, indicate that the system of afferents impinging upon the abducens nucleus is built on a common basic plan in both species, but there are many ways in which they have diverged.

ACKNOWLEDGMENTS We would like to thank Tin Tran and Karl Muhlbach for

photographic assistance, Katie Sacket for secretarial help, Eileen LaBossiere for histological assistance, and Kate Schmitt for editorial and word processing aplomb. We also thank our colleagues, M. Mustari, C. Schupert, J. Mc-


Farland, and J. Phillips, for critical review of earlier ver- sions of the manuscript, and Dr. R. Spencer for his helpful criticism of the final version.

This study was supported by grants EY03212, EY00745, and RR00166 from the National Institutes of Health, and BNS8218901 from the National Science Foundation.

LITERATURE CITED Alley, K., R. Baker, and J.I. Simpson (1975) Afferents to the vestibulocere-

bellum and the origin of the visual climbing fibers in the rabbit. Brain Res. 98582-589.

Anderson, J.H., W. Precht, and C. Pappas (1979) Changes in the vertical vestibulo-ocular reflex due to kainic acid lesions of the interstitial nucleus of Cajal. Neurosci. Lett. 14259-264.

Baker, R., and R. McCrea (1979) The parabducens nucleus. In H. Asanuma and V.J. Wilson (eds): Integration in the Nervous System. New York Igaku-Shoin, pp. 97-121.

Baker, R., and R.F. Spencer (1981) Synthesis of horizontal conjugate eye movement signals in the abducens nucleus. Jpn. J. EEG EMG 31:49- 59.

Blanks, R.H.I., and Y. Torigoe (1983) Two neuronal groups as associated with the medial longitudinal fasciculus (mln of cat and rat that conveys vestibular impulses to the cerebellar flocculus. Soc. Neurosci. Abstr. 9:608.

Blanks, R.H.I., W. Precht, and Y. Torigoe (1983) Afferent projections to the cerebellar flocculus in the pigmented rat demonstrated by retrograde transport of horseradish peroxidase. Exp. Brain Res. 52293-306.

Brodal, A. (1952) Experimental demonstration of cerebellar connections from the peri-hypoglossal nuclei (nucleus intercalatus, nucleus praepos- itus hypoglossi and nucleus of Roller) in the cat. J. Anat. 86:llO-129.

Brodal, A. (1957) The Reticular Formation of the Brain Stem. Anatomical Aspects and Functional Correlations. The Henderson Trust Lecture. Edinburgh: Oliver and Boyd.

Biittner-Ennever, J.A. (1981) Anatomy of medial rectus subgroups in the oculomotor nucleus of the monkey. In A.F. Fuchs and W. Becker (eds): Progress in Oculomotor Research (Developments in Neuroscience). New York: ElsevierNorth-Holland. 12.247-252.

Biittner, J.A., and V. Henn (1976) An autoradiographic study of the pathways from the pontine reticular formation involved in horizontal eye movements. Bram Res. 108t155-164.

Biittner-Ennever, J.A., and M. Pause (1985) Neuroanatomic identification of a raphe nucleus in the pons associated with omnipause neurons of the oculomotor system in the monkey. SOC. Neurosci. Abstr. 11:1042.

Biittner, U., J.A. Buttner-Ennever, and V. Henn (1977) Vertical eye movement related unit activity in the rostra1 mesencephalic reticular formation of the alert monkey. Brain Res. 130239-252.

Carpenter, M.B., and R.R. Batton I11 (1980) Abducens internuclear neurons and their role in conjugate horizontal gaze. J. Comp. Neurol. 1891191- 209.

Chubb, M.C., and A.F. Fuchs (1981) The role of the dentate nucleus and y- group in the generation of vertical smooth eye movements. Ann. N.Y. Acad. Sci. 374:446-454.

Curthoys, I.S., S. Nakao, and C.H. Markham (1981) Cat medial pontine reticular neurons related to vestibular nystagmus: Firing pattern, location, and projection. Brain Res. 22235-94.

Delgado-Garcia, J.M., F. Del Pozo, and R. Baker (1983) The properties of antidromically identified abducens motoneurons in the alert cat. SOC. Neurosci. Abstr. 9:1088.

Edwards, S.B., and C.K. Henkel(1978) Superior colliculus connections with the extraocular motor nuclei in the cat. J. Comp. Neurol. 179t451-468.

Feldon, S., P. Feldon, and L. Kruger (1970) Topography of the retinal projection upon the superior colliculus of the cat. Vision Res. 10:135- 143.

Fink, R.P., and L. Heimer (1967) Two methods for selective silver impreg- nation of degenerating axons and their synaptic endings in the central nervous system. Brain. 4r369-374.

Fuchs, A.F., and E.S. Luschei (1970) Firing patterns of abducens neurons of alert monkeys in relationship to horizontal eye movements. J. Neuro- physiol. 33:382-392.

Fuchs, A.F., C.R.S. Kaneko, and C.A. Scudder (1985) Brainstem control of saccadic eye movements. Ann. Rev. Neurosci. 8t307-337.

Fuchs, A.F., C.A. Scudder, and C.R.S. Kaneko (1984) Discharge characteristics of internuclear neurons in the monkey. SOC. Neurosci. Abstr. 10:391

Gacek, R.R. (1979) Location of abducens afferent neurons in the cat. Exp. Neurol. 64:342-353.

Grantyn, A.A., and R. Grantyn (1976) Synaptic actions of tectofugal pathways on abducens motoneurons in the cat. Brain Res. 105:269-285.

Grantyn, A., and R. Grantyn (1982) Axonal patterns and sites of termination of cat superior colliculus neurons projecting in the tecto-bulbo- spinal tract. Exp. Brain Res. 462434256,

Graybiel, A.M. (1977a) Organization of oculomotor pathways in the cat and rhesus monkey. Dev. Neurosci. 139-88.

Graybiel, A.M. (197713) Direct and indirect preoculomotor pathways of the brainstem: Autoradiographic study of the pontine reticular formation in the cat. J. Comp. Neurol. 175:37-78.

Graybiel, A.M., and E.A. Hartwieg (1974) Some afferent connections of the oculomotor complex in the cat: An experimental study with tracer tech- niques. Brain Res. 81:543-551.

Henn, V., and B. Cohen (1976) Coding of information about rapid eye movements in the pontine reticular formation of alert monkeys. Brain Res. 108:307-325.

Hepp, K., and V. Henn (1983) Spatio-temporal recoding of rapid eye movement signals in the monkey paramedian pontine reticular formation. Exp. Brain Res. 52105-120.

Highstein, S.M., and R. Baker (1978) Excitatory termination of abducens internuclear neurons in medial rectus motoneurons: Relationship to syndrome of internuclear ophthalmoplegia. J. Neurophysiol. 41,1647- 1661.

Highstein, S.M., A. Karabelas, R. Baker, and R.A. McCrea (1982) Compari- son of the morphology of physiologically identified abducens motor and internuclear neurons in the cat: A light microscopic study employing the intracellular injection of horseradish peroxidase. J. Comp. Neurol. 208:369-381.

Hikosaka, O., and T. Kawakami (1977) Inhibitory reticular neurons related to the quick phase of vestibular nystagmus: Their location and projection. Exp. Brain Res. 27:377-396.

Hikosaka, O., Y. Igusa, S. Nakao, and H. Shimazu (1978) Direct inhbitory synaptic linkage of pontomedullary reticular burst neurons with abducens motor neurons in the cat. Exp. Brain Res. 33:337-352.

Ishizuka, N., H. Mannen, S. Sasaki, and H. Shimazu (1980) Axonal branches and terminations in the cat abducens nucleus of secondary vestibular neurons in the horizontal canal system. Neurosci. Lett. 16143-148.

Kaneko, C.R.S., C. Evinger, and A.F. Fuchs (1981) The role of cat pontine burst neurons in the generation of saccadic eye movements. J. Neuro- physiol. 46r387-408.

Kaneko, C.R.S., and A.F. Fuchs (1981) Inhibitory burst neurons in alert trained cats: Comparison with excitatory burst neurons and functional implications. In A.F. Fuchs and W. Becker (eds): Progress in Oculomotor Research (Developments in Neuroscience) 12:63-71. New York: Elsevieri North-Holland.

Keller, E.L. (1974) Participation of medial pontine reticular formation in eye movement generation in monkey. J. Neurophysiol. 37t316-332.

Keller, E.L., and W.F. Crandall (1983) Neuronal responses to optokinetic stimuli in pontine nuclei of behaving monkey. J. Neurophysiol. 49:169- 187.

King, W.M., and A.F. Fuchs (1979) Reticular control of vertical saccadic eye movements by mesencephalic burst neurons. J. Neurophysiol. 42861- 876.

King, W.M., W. Precht, and N. Dieringer (1980) Synaptic organization of frontal eye field and vestibular afferents to interstitial nucleus of Cajal in the cat. J. Neurophysiol. 43;912-928.

Kotchabhakdi, N., and F. Walberg (1977) Cerebellar afferents from neurons in motor nuclei of cranial nuclei demonstrated by retrograde axonal transport of horseradish peroxidase. Brain Res. 137:158-163.

LaBossiere, E. (1976) Histological Processing for the Neural Sciences. Springfield Charles C. Thomas.

Langer, T. (1985) Basal interstitial nucleus of the cerebellum: A cerebellar nucleus related to the flocculus. J. Comp. Neurol. 235:38-47.

Langer, T.P., and C.R.S. Kaneko (1983) Efferent projections of the cat OCU-

lomotor reticular omnipause neuron region: An autoradiographic study. J. Comp. Neurol. 217,188-306.

Langer, T., A.F. Fuchs, C.A. Scudder, and M.C. Chubb (1985) Afferents to the flocculus of the cerebellum in the rhesus macaque as revealed by retrograde transport of horseradish peroxidase. J. Comp. Neurol. 235:l- 25.

Lopez-Barneo, J., C. Darlot, A. Berthoz, and R. Baker (1982) Neuronal activity in prepositus nucleus correlated with eye movement in the alert cat. J. Neurophysiol. 47:329-352.

400 T. LANGER ET AI,.

Luschei, E.S., and A.F. Fuchs (1972) Activity of brain stem neurons during eye movements of alert monkeys. J. Neurophysiol. 35445-461.

Maciewicz, R., and B.S. Phipps (1983) The oculomotor internuclear pathway: A double retrograde labeling study. Brain Res. 262:1-8.

Maciewicz, R.J., C.R.S. Kaneko, S.M. Highstein, and R. Baker (1975) Mor- phophysiological identification of interneurons in the oculomotor nucleus that project to the abducens nucleus in the cat. Brain Res. 96:60-- 65.

Maciewicz, R.J., K. Eagen, C.R.S. Kaneko, and S.M. Highstein (1977) Ves- tibular and medullary brain stem afferents to the abducens nucleus in the cat. Brain Res. 123:229-240.

McCrea, R.A., R. Baker, and J. Delgado-Garcia (1979) Afferent and efferent organization of the prepositus hypoglossi nucleus. Progr. Brain Res. 503353-665.

McCrea, R.A., K. Yoshida, A. Berthoz, and R. Baker (1980) Eye movement related activity and morphology of second order vestibular neurons terminating in the cat abducens nucleus. Exp. Brain Res. 40:468-473.

Mesulam, M.-M. (1978) Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reaction-product with superior sensitivity for visualizing neural afferents and efferents. J. His- tochem. Cytochem. 26:106-117.

Naito, H., K. Tanimura, N. Taga, and Y. Hosoya (1974) Microelectrode study on the subnuclei of the oculomotor nucleus in the cat. Brain Res. 81:215- 231.

Olszewski, J., and D. Baxter (1954) Cytoarchitecture of the Human Brain Stem. New York Karger.

Pola, J., and D.A. Robinson (1978) Oculomotor signals in medial longitudinal fasciculus of the monkey. J. Neuropbysiol. 41:245-259.

Sato, Y., T. Kawasaki, and K. Ikarashi (1983) Afferent projections from the brainstem to the three floccular zones in cats. 11. Mossy fiber projections. Brain Res. 2723748.

Schiller, P.H. (1970) The discharge characteristics of single units in the oculomotor and abducens nuclei of the unanesthetized monkey. Exp. Brain Res. 10:347-362.

Schwindt, P.C., W. Precht, and A. Richter (1974) Monosynaptic excitatory and inhibitory pathways from medial midbrain nuclei to trochlear mo-

toneurons. Exp. Brain Res. 20:223-238. Scudder, C.A., and A.F. Fuchs (1981) Projections from the vestibular to

abducens nuclei as revealed by spike triggered averaging. SOC. Neurosci. Abstr. 7:777.

Scudder, C.A., T. Langer, and A.F. Fuchs (1982) Probable inhibitory burst neurons in the monkey. SOC. Neurosci. Abstr. 8:157.

Spencer, R.F., and J.D. Porter (1981) Innervation and structure of extraocular muscles in the monkey in comparison to those of the cat. J. Comp. Neurol. 198:649-665.

Spencer, R.F., and P. Sterling (1977) An electron microscopic study of moto- neurones and interneurones in the cat abducens nucleus identified by retrograde intraaxonal transport of horseradish peroxidase. J. Comp. Neurol. 17665-86.

Stanton, G.B., and R.W. Greene (1981) Brain stem afferents to the periabdu- cens reticular formation (PARF) in the cat. Exp. Brain Res. 44r419-426.

Steiger, H.-J., and J.A. Buttner-Ennever (1978) Relationship between motoneurons and internuclear neurons in the abducens nucleus: A double retrograde tracer study in the cat. Brain Res. 148:181-188.

Steiger, H.-J., and J.A. Buttner-Ennever (1979) Oculomotor nucleus afferents in the monkey demonstrated with horseradish peroxidase. Brain Res. 16O:l-15.

Stein, B.E., B. Magalhaes-Castro, and L. Kruger (1975) Superior colliculus: Visuotopic-somatotopic overlap. Science 189:224-226.

Strassman, A., R.A. McCrea, and S.M. Highstein (1982) Structure function correlation of abducens motor and internuclear neurons in relation to eye movements: An intracellular HRP study in the alert squirrel monkey. SOC. Neurosci. Abstr. 8:156.

Taber, E. (1961) The cytoarchitecture of the brain stem of the cat. I. Brain stem nuclei of cat. J. Comp. Neurol. 116:27-70.

Taber, E., A. Brodal, and F. Walberg (1960) The raphe nuclei of the brainstem in the cat. I. Normal topography and cytoarchitecture and general discussion. J. Comp. Neurol. 114:161-187.

Yoshida, K., R. McCrea, A. Berthoz, and P.P. Vidal (1982) Morphological and physiological characteristics of inhibitory burst neurons controlling horizontal rapid eye movements in the alert cat. J. Neurophysiol. 48:761-- 784.

afferents to the abducens nucleus in the monkey and cat

Documents