Central projections of the saccular and utricular nerves in macaques

Superior vestibular nucleus.

The SuVe is involved as a relay nucleus in the vertical vestibulo-ocular reflex (VOR). Secondary VOR neurons project by means of the medial longitudinal fasciculus to the ipsilateral oculomotor and trochlear nuclei (Highstein, 1973; Highstein and Reisine, 1979; Graf et al., 1983; McCrea et al., 1987a). The nucleus receives extensively overlapping inputs from all three semicircular canals (Naito et al., 1995; Dickman and Fang, 1996; Newlands and Perachio, 2003).

In gerbils, saccular afferents had only sparse SuVe terminations (Kevetter and Perachio, 1986). In cats, saccular projections to the caudal SuVe are reported from two intracellularly filled units (Imagawa et al., 1998). Saccular projections to the SuVe are lateral in pigeons (Dickman and Fang, 1996), lateral in guinea pigs based on application of HRP crystals to the saccular nerve (Gstoettner et al., 1992) or injection into the saccular neuroepithelium (Didier et al., 1987), and ventrolateral in squirrel monkeys (Naito et al., 1995). In general, these findings are confirmed here in macaques but dorsal terminals are also noted in the rostral half of the nucleus.

Reports of utricular projection patterns in other species have varied. The utricle had terminations primarily in the lateral and rostral parvicellular portions of the nucleus in gerbils (Newlands et al., 2002). Siegborn and Grant (1983) did not find that utricular afferents projected to the SuVe in cats after HRP application to the cut nerve stump. Yet in squirrel monkeys, the utricular afferents were found to project to the dorsolateral SuVe (Naito et al., 1995), whereas single unit fills of four neurons in cats demonstrated projection to the caudal and ventral portion of the nucleus only (Imagawa et al., 1995). The current study demonstrates a utricular projection that is primarily to the ventral and lateral parts of the nucleus.

All of these reports demonstrate that both otolith macula project to peripheral regions of the SuVe, not to the central, larger-celled area thought to be the origin of inhibitory second-order vestibulo-ocular fibers (Highstein and Reisine, 1979; Mitsacos et al., 1983a). These peripheral regions have been shown to project to the cerebellum and bilateral vestibular nuclei (Gacek, 1978; Pompeiano et al., 1978; Mitsacos et al., 1983b).

Lateral vestibular nucleus.

The LVe is characterized by Deiters' cells, which project to and receive inputs from the spinal cord (Wilson et al., 1967; Boyle and Pompeiano, 1980; Boyle et al., 1992; Xiong and Matsushita, 2001) and, thus, is primarily believed involved in vestibulospinal mechanisms. Neurons in the lateral and central LVe give rise to the lateral vestibulospinal tract and participate in postural control (Wilson et al., 1967; Boyle et al., 1992). Neurons in the lateral vestibulospinal tract project to the cervical and lumbar spinal cord. The lumbar-projecting neurons bypass the cervical cord and do not contribute to head posture and stabilization (Boyle, 2000). These neurons are located more dorsally and rostrally in the LVe of cats, where utricular and saccular afferent terminals are not seen, consistent with monosynaptic inputs being more prevalent to cervical projection than to lumbar-projecting LVe cells. Lateral vestibulospinal tract neurons projecting to the cervical cord arise from both the LVe and MVeMC, and many of these show physiologic evidence of monosynaptic afferent input (Boyle et al., 1992; Uchino et al., 1994; Sato et al., 1996). Although the LVe has been considered void of afferent input, more recent studies using BDA (Kevetter et al., personal communication; gerbil) and carbocyanine fluorescent dyes (Maklad and Fritzsch, 2002; mouse) have reported afferent terminals in the LVe. Similarly, in this study using BDA, such terminals are reported, although this connection is weak. Monosynaptic latencies were seen for production of excitatory postsynaptic potentials (EPSPs) in many spinal-projecting LVe neurons in squirrel monkeys (Boyle et al., 1992). These EPSPs were small and did not result in discharge, suggesting that afferent input to spinal-projecting LVe neurons are distal on the dendritic tree and may be outside the anatomic boundaries of the LVe (Highstein et al., 1987; Boyle et al., 1992).

Medial vestibular nucleus, parvicellular part.

This nucleus has medium to small neurons and is believed to be involved in projection to oculomotor nuclei, the medial vestibulospinal tract, and the vestibular commissural system (Gacek, 1969; Newlands et al., 1989; Minor et al., 1990; Boyle et al., 1992; Xiong and Matsushita, 2001). This nucleus receives input from all of the semicircular canals (Kevetter and Perachio, 1985, 1986; Sato et al., 1989; Siegborn et al., 1991; Gstoettner et al., 1992; Naito et al., 1995; Dickman and Fang, 1996). In general, the rostral portion of the nucleus is believed more involved in ocular motor reflexes and the caudal portion with spinal projections, although there is great overlap in these systems and many neurons project in both systems (Graf and Ezure, 1986; Highstein et al., 1987; McCrea et al., 1987a, b; Boyle et al., 1992; Büttner-Ennever, 1992; Scudder and Fuchs, 1992).

In contrast to canal inputs, otolith inputs do not ubiquitously occupy the MVePC. In this study, we found saccular projections were light, except for the lateral dorsal corner of the nucleus where it abuts the SuVe and LVe, which is consistent with the preponderance of previous reports. Saccular projections to the MVePC were not seen in gerbils (Kevetter and Perachio, 1986) and were reported as weak in cats based on single unit injections (Imagawa et al., 1998). Didier et al. (1987) found saccular projections to the rostral extent of the MVePC in guinea pigs. Naito et al. (1995), in squirrel monkeys, reported light saccular nerve projections throughout the MVePC. Potentially, in this latter study, there may have been spread of tracer to adjacent end organs or these findings may represent a species difference.

Utricular nerve afferents in gerbils project to the entire MVePC, except for the most central portion of the nucleus (Newlands et al., 2002). The utricular nerve was noted to project to the MVePC in cats (Siegborn and Grant, 1983, specifics not given) and squirrel monkeys (Naito et al., 1995, central part of the nucleus rostrally becoming more lateral as one moves caudally). After injection into single utricular units, Imagawa et al. (1995) described the distribution of terminals as similar to canals but sparser. In pigeons, the utricular and saccular projection was described as lateral (Dickman and Fang, 1996). The current study demonstrates limitation of utricular nerve terminals to the lateral border of the nucleus where it abuts the SpVe and the MVeMC and the rostral pole of the nucleus where it abuts the SuVe. In these locations, the afferents might be involved in ocular motor or spinal reflexes (Boyle et al., 1992). This region contains convergent central neurons that are sensitive to linear and angular acceleration but do not have eye movement sensitivity (Dickman and Angelaki, 2002) and, thus, are potentially involved in the medial vestibulospinal tract. Similarly, neurons believed involved in the VOR with both eye movement and vestibular sensitivity such as position vestibular pause or eye head velocity neurons are located in this region (Scudder and Fuchs, 1992). Many of these cell types have both angular and linear sensitivity and are potentially involved in control of both the linear and rotational VOR (Angelaki et al., 2001).

Medial vestibular nucleus, magnocellular part.

Older studies reporting afferent input to the ventral and ventrorostral LVe likely refer to the MVeMC. Studies tracing the individual canal (Siegborn and Grant, 1983; Kevetter and Perachio, 1986; Siegborn et al., 1991; Gstoettner et al., 1992; Naito et al., 1995; Kevetter et al., personal communication) and otolith nerves (Siegborn and Grant, 1983; Kevetter and Perachio, 1986; Didier et al., 1987; Naito et al., 1995; Newlands et al., 2002) have found that all of the end organs project significantly to this region.

In macaques, this nucleus receives both saccular and utricular input, although the utricular input is more substantial. The MVeMC is at the center of major output pathways of the vestibulomotor activity. Neurons mediating both the horizontal (McCrea et al., 1987b) and vertical (Graf and Ezure, 1986; McCrea et al., 1987a) VOR and vestibulospinal neurons are found in this area (Boyle and Pompeiano, 1980; Boyle et al., 1992). Lateral vestibular nucleus neurons described as driven monosynaptically from the saccular nerve potentially actually reside in the MVeMC, as reports are imprecise about the anatomic location of these cells (Uchino et al., 1997; Sato et al., 2000). The projection of all of the end organ afferents to the MVeMC provides an opportunity for interaction between utricular and canal inputs on central neurons.

The ascending tract of Deiters originates in the MVeMC and rostral MVePC and projects to the ipsilateral oculomotor nucleus subdivision that drives the medial rectus muscle (Baker and Highstein, 1978; McCrea et al., 1987b). Neurons in this tract carry combined angular and linear VOR signals that are likely involved in disconjugate eye movements, because their gain changes with viewing distance (Chen-Huang and McCrea, 1998). Many of these neurons behave like eye head velocity cells, which are known to carry both rotational and translational sensory information (Angelaki et al., 2001) and may be floccular target neurons (Lisberger et al., 1994). Other reports have noted type I pause vestibular pause neurons in the ascending tract of Deiters (McCrea et al., 1987a). The MVeMC also has many neurons projecting to the contralateral abducens (McCrea et al., 1987a). Thus, utricular inputs to this region are likely used to generate the linear VOR.

Spinal vestibular nucleus.

This nucleus projects to the spinal cord, the contralateral vestibular nuclei, and the vermis (Boyle and Pompeiano, 1980; Carleton and Carpenter, 1983, Newlands et al., 1989; Thunnissen et al., 1989; Boyle, 1993; Xiong and Matsushita, 2001). The rostral portion of the nucleus also participates in vestibulo-ocular pathways (Graf and Ezure, 1986). It receives input from the contralateral fastigial nucleus (Sugita and Noda, 1991), ipsilateral vermis (Carleton and Carpenter, 1983; Thunnissen et al., 1989), contralateral vestibular nuclei (Carleton and Carpenter, 1984; Newlands et al., 1989), and the spinal cord (Rubertone and Haines, 1982). Uchino et al. (1997) found saccular driven cells in the SpVe, some with monosynaptic input from one side of the striola and disynaptic convergence from the other side of the striola.

In macaques, the SpVe is most heavily innervated by saccular afferents. These afferents are found throughout the SpVe. Utricular projections to this nucleus are found rostrally, ventrally, and laterally. Several authors have described similar specific peripheral innervation patterns. In gerbils (Newlands et al., 2002), pigeons (Dickman and Fang, 1996), squirrel monkeys (Naito et al., 1995), and cats (Imagawa et al., 1995), utricular afferents were found primarily to project to the lateral portion of the nucleus. A strong saccular input was seen throughout the nucleus in these same species (Kevetter and Perachio, 1986; Naito et al., 1995; Dickman and Fang, 1996; Imagawa et al., 1998). Many lateral vestibular spinal tract neurons to the lumbar cord arise in the caudal SpVe (Boyle et al., 1992). The heavy saccular projection throughout this nucleus supports the role of the saccule in vestibulospinal righting reflexes.

Nucleus y.

In many mammals, the ventral portion of nucleus y receives saccular input (Gacek, 1969; Fredrickson and Trune, 1986; Kevetter and Perachio, 1986; Didier et al., 1987; Gstoettner et al., 1992; Naito et al., 1995), whereas the dorsal portion does not. The ventral portion, not the dorsal portion, participates in the vestibular commissural system (Gacek, 1978) and projects to the ipsilateral vestibular nuclei (Rubertone et al., 1983). In contrast, the dorsal portion of nucleus y receives the densest floccular outflow (Langer et al., 1985a) and projects to the oculomotor (Gacek, 1977; Highstein and Reisine, 1979; Steiger and Büttner-Ennever, 1979) and trochlear (Gacek, 1979) nuclei by means of the brachium conjunctivum. In rhesus macaques, Langer et al. (1985b) refer to the ventral portion of nucleus y as the supravestibular nucleus and demonstrate afferent projection to the flocculus (ipsilateral greater than contralateral). Ipsilateral but not contralateral nucleus y projections to the flocculus and bilateral nucleus y projections to the uvula were reported in galagos (Rubertone and Haines, 1981). The physiologic role of ventral nucleus y is not clear. Presumably, this nucleus could influence postural and ocular motor reflexes through extensive connections in the vestibulocerebellum and other vestibular nuclei.

There have been few reports of other vestibular nerve afferent projections to nucleus y. In squirrel monkeys, no utricular or canal projections were seen to nucleus y (Naito et al., 1995). In cats, one of the 16 intracellularly labeled regularly firing lateral canal units terminated in nucleus y; none of the six irregular fibers were traced to this nucleus (Sato et al., 1989). In gerbils, injections into canal ampullae did not result in labeling in nucleus y (Kevetter and Perachio, 1985, 1986; Kevetter et al., personal communication) but utricular afferents to nucleus y and the infracerebellar nucleus have been reported (Newlands et al., 2002). Didier et al. (1987), working in guinea pigs, found that only the posterior branch of the saccular nerve projected to the nucleus y group, whereas fibers traveling with the anterior branch did not. In the current study, both the saccular and utricular afferents projected to ventral nucleus y. As in our previous study in gerbils, dorsal nucleus y did have an occasional utricular terminal (Newlands et al., 2002).

Interstitial nucleus of the eighth nerve.

Little is known about this nucleus, other than its input from vestibular afferents. Input from the cerebellar cortex of the vermis has been reported in galagos (Haines, 1977) and the nucleus projects to the flocculus (Rubertone and Haines, 1981; Langer et al., 1985b) and nodulus (Rubertone and Haines, 1981) in primates.

Neuronal tracer studies have demonstrated, although inconsistently, input from all of the vestibular end organs. HRP tracer studies of the canal and otolith nerves in squirrel monkeys demonstrated little input from canals that was generally rostral and a heavier input from the utricle caudocentrally and saccule caudally (Naito et al., 1995). Similarly, we saw primarily lateral interstitial nucleus of the eighth nerve labeling in macaques. Intracellular studies in cats have demonstrated branches from afferents of all canals (Mannen et al., 1982; Sato et al., 1989) and the saccule (Imagawa et al., 1998). In gerbils, input arises from all end organs (Kevetter and Perachio, 1986; Newlands et al., 2002; Kevetter et al., personal communication).

Cell group l.

Group l is a small group of cells just lateral to the LVe that has been distinguished in macaques (Brodal, 1984), cats (Brodal and Pompeiano, 1957), galagos (Rubertone and Haines, 1982), and squirrel monkeys (Naito et al., 1995). Little is known about the connections of this area, but in cats, the cells appear to project to the spinal cord (Pompeiano and Brodal, 1957). For this reason, this cell group has been thought of as being functionally a part of the LVe (Pompeiano and Brodal, 1957). In squirrel monkeys, terminals from utricular and saccular afferents were seen in cell group l (Naito et al., 1995). In the current study, we saw a group of smaller cells beginning between the rostral pole of the SpVe and the inferior cerebellar peduncle extending up to the entry zone of the eighth nerve. This group of cells is ventral to nucleus y and ventral and caudal to the lateral, caudal pole of SuVe. These cells are heavily labeled after injection of either the utricular or saccular macula, although the utricular input appeared stronger.

Cochlear nucleus.

Several studies have described saccular innervation of the cochlear nuclei. Kevetter and Perachio (1985, 1986, 1989) described this projection in gerbils. Single unit labeling and injection of HRP into the macula demonstrated terminations in the anterior and posterior portions of the ventral cochlear nucleus and the dorsal cochlear nucleus. Most of these terminations were in the granule cell domain, especially in the subpeduncular corner between the anteroventral cochlear nucleus and the floccular peduncle of the cerebellum. Burian and Gstoettner (1988) also showed saccular afferents terminating between the dorsal and posteroventral cochlear nuclei. Gstoettner et al. (1992), also in guinea pigs, demonstrated saccular projections to the dorsal cochlear nucleus. In the current study, we see clear evidence of a saccular projection in rhesus monkeys to the VCP that is strongest at the medial edge of this nucleus abutting the inferior cerebellar peduncle. We did not, however, examine the cochlea for spread. Given the lack of label in the cochlear nerve and other parts of the cochlear nuclei, contamination seems unlikely. Acoustically responsive saccular nerve fibers have been demonstrated in cats, although the role of such fibers and of the saccular afferent projection to the cochlear nucleus remains obscure (McCue and Guinan, 1994).

External cuneate nucleus.

Projections to the superior part of the ECu have been irregularly described originating from several end organs. In gerbils, this projection has been described from the utricle (Newlands et al., 2002) and saccule (Kevetter and Perachio, 1985, 1986). In cats, Siegborn and Grant (1983) reported this projection from the lateral canal but not the utricle. In guinea pigs, a saccule to ECu projection has also been reported (Didier et al., 1987). We find clear projections to this nucleus from the saccule and utricle in macaques. The ECu carries proprioceptive input from the neck and forelimb to the thalamus and cerebellum in monkeys (Rustioni et al., 1979; Pearson and Garfunkel, 1983). Thus, this nucleus may be a center for integration of proprioceptive and otolith inputs that relay a signal related to the animal's position space to higher levels in the brain (Jensen and Thompson, 1983; Anastasopoulos et al., 1991).

Cerebellar nuclei.

Based on degeneration, Carpenter (1960) described cerebellar-projecting vestibular afferent fibers to the fastigial nucleus of cats, whereas Brodal and Høivik (1964) and Korte and Mugnaini (1979) saw fibers passing through but were not able to determine whether they were fibers of passage or terminals. In rats, Mehler and Rubertone (1985), by using degeneration techniques, noted fibers to the caudal fastigial nucleus only. HRP studies in gerbils have demonstrated saccular projections to the lateral and interposed cerebellar nuclei (Kevetter and Perachio, 1986) and utricular projections to the fastigial and anterior interposed cerebellar nuclei (Newlands et al., 2002). Lateral and anterior canal afferents terminate in the interposed nucleus as well (Kevetter et al., personal communication). All of these projections described in gerbils are on a minute scale.

In the current study, projections to the fastigial nuclei were seen bilaterally from the utricle and unilaterally from the saccule. Although a stray terminal from the saccule was seen in other cerebellar nuclei, these terminals are rare. Our study agrees with others that vestibular afferent input to the cerebellar nuclei is weak, and primarily involves the fastigial nucleus. Thus, complex otolith responses recorded in the fastigial nucleus are likely manifestations of polysynaptic inputs (Zhou et al., 2001).

Vermis.

Many studies have demonstrated a substantial vestibular afferent input to the vermis, but few have been designed to distinguish the otolith contribution. Injection of tritiated amino acids into the labyrinth demonstrated heavy labeling of mossy fiber rosettes in the granular layer of the ipsilateral uvula and nodulus and lighter labeling in the deep aspects of vermal folia I, V, and VI (Carleton and Carpenter, 1984). Similar findings were reported in rabbits (Gerrits et al., 1989) where heavy labeling was noted in the ipsilateral and medial third of the contralateral vermal lobules X and IXd after injection of WGA-HRP and tritiated leucine in the vestibular ganglion. Weak labeling was seen in lobules I, II, and the fissures between lobules IV and V and V and VI.

These anterograde studies have been complemented by retrograde studies as well. Injection of HRP into the nodulus of five rabbits demonstrated Scarpa's ganglion labeling, whereas injections into lobule VI did not (Alley et al., 1975). Large HRP injections into the vermis in cats also retrogradely labeled Scarpa's ganglion cells bilaterally (Kotchabhakdi and Walberg, 1978). Injection of WGA-HRP into lobules IXd and X retrogradely filled 64% of Scarpa's ganglion cells in rabbits (Barmack et al., 1993). The authors considered 64% a conservative estimate of the strength of the vestibular afferent projection to the vermis. They confirmed the projection by injection of C fragment of tetanus toxin into the labyrinth and observing dense labeling ipsilaterally in vermal areas. Weak labeling was seen in the primary fissure and between the third and fourth lobules. They also injected fluorescent markers into the uvula–nodulus and vestibular nuclear complex and found double fluorescent labeling in more than 50% of Scarpa's ganglion neurons.

Retrograde and anterograde tracer studies in gerbils have confirmed projections from all end organs to the vermis. The posterior canal projected to the uvula and nodulus, whereas the saccule projected to lobules I, V, VI, and, most strongly, the uvula (Kevetter and Perachio, 1986). Utricular, anterior canal, and lateral canal afferents project most heavily to the uvula and nodulus (Kevetter et al., personal communication; Newlands et al., 2002). The afferents in this species make up mossy fibers that terminate on the granule cells and unipolar brush cells (Dino et al., 2001). Injection of BDA into the uvula and nodulus demonstrated that the majority of utricular and saccular afferents to the vermis were dimorphic and arose from the peristriolar zone (Purcell and Perachio, 2001). Didier et al. (1987) noted that projections to the nodulus travel through the posterior branch but not the anterior branch of the saccular nerve in guinea pigs. In perinatal mice, saccular and utricular afferents project primarily to the uvula, subfolium IXd, with little projection to the nodulus, whereas canal afferents project primarily to the nodulus, with less input to IXd (Maklad and Fritzsch, 2003).

The uvula and nodulus have a rich convergent input of vestibular and visual signals (Barmack and Shojaku, 1995). Climbing fibers from the β-nucleus and dorsomedial cell column of the inferior olive carry vestibular inputs related to coding dynamic roll information in the plane of the vertical canals and static tilt information, presumably of utricular origin (Barmack and Shojaku, 1995). Mossy fiber input to the uvula/nodulus arises from the SuVe, rostral and caudal MVePC, and SpVe, whereas the LVe does not project to this area and the MVeMC projections are light in rabbits (Thunnissen et al., 1989; Barmack et al., 1992). These connections were found to be reciprocal and some vestibular nucleus neurons that project to the uvula/nodulus also project to the flocculus (Epema et al., 1990). The convergence of primary and secondary vestibular information in the uvula/nodulus makes this area an ideal candidate for control and modification of spatial orientation tasks. Evidence for the uvula/nodulus in this role is the finding that lesions of the uvula and nodulus in rhesus monkeys eliminate the ability of the otolith system to generate steady-state nystagmus during constant velocity rotation and prevent otolith contribution to the VOR during low-frequency sinusoidal off-vertical axis rotations (Angelaki and Hess, 1995).

Surprisingly, we found light input to lobule IX from the utricular injections. This contrasts with a very strong projection to lobule IXd with a weaker input to lobule X seen after saccular injection. This finding suggests a difference in utricular and saccular nerve projections, although the subtotal nature of our injections raises the possibility that our finding reflects a subtotal sampling of the otolith to vermis projection.

Flocculus and paraflocculus.

The presence or absence of afferent projections to the flocculus and paraflocculus is controversial. Degeneration studies suggest such a projection but are inconclusive (Korte and Mugnaini, 1979). Tritiated amino acid injections into the labyrinth of cats demonstrate limited labeling in the flocculus but no labeling in the paraflocculus (Carleton and Carpenter, 1984). Electrophysiologic studies of this region have not provided definitive answers, but in rabbits (Ito et al., 1982) and cats (Shinoda and Yoshida, 1975), data have been obtained to suggest a direct afferent input to the flocculus.

Several retrograde and anterograde tracer studies have examined vestibular afferent projections to the flocculus and paraflocculus. A small number of ipsilateral vestibular ganglion cells were labeled after HRP injection into the flocculus in rats (Blanks et al., 1983). In rabbits, Gerrits et al. (1989) injected WGA-HRP and tritiated leucine into the vestibular ganglion. None of the labeled afferents projected to the flocculus or paraflocculus. These authors believe that earlier retrograde studies in rabbits (Alley et al., 1975) and rats (Blanks et al., 1983) with injection into the flocculus involved spread of the tracer to nucleus y, giving a false-positive result. Also in rabbits, injection of WGA-HRP and fluorescent dyes into the flocculus and C fragment of tetanus toxin into the vestibular ganglion demonstrated essentially no vestibular afferent projection to the ventral paraflocculus in rabbits and a weak projection to the flocculus (Barmack et al., 1993).

Osanai et al. (1999) injected HRP, WGA-HRP, Fast Blue, or diamidino yellow into the flocculus, ventral paraflocculus, and/or dorsal paraflocculus of rats. After flocculus injection, these authors reported vestibular ganglion labeling in several but not all of their experimental animals. In the limited injections to the flocculus, no retrograde vestibular ganglion labeling was seen, whereas in the large injections labeling was seen, possibly from spill into nucleus y. Injection into the dorsal or ventral paraflocculus did not produce labeling in the vestibular ganglion. BDA injection into the flocculus of gerbils labeled peripheral zone dimorphic units in all of the semicircular canals but not afferents originating in the otolith organs (Purcell and Perachio, 2001).

As with these afoveate animals, results of anatomic studies in monkeys have yielded inconsistent results. Langer et al. (1985b) injected HRP into the flocculus of rhesus monkeys and reported no retrograde vestibular afferent labeling. In Macaca fuscata, WGA-HRP injection into the flocculus resulted in labeling in 5–10% of the vestibular ganglion cells, whereas injections into the ventral paraflocculus did not retrogradely label vestibular ganglion neurons (Nagao et al., 1997). These authors contend that only those folia caudal to the posterolateral fissure in monkeys (folia 1–5; after Madigan and Carpenter, 1971) are comparable to the flocculus in other species, including afoveate animals. The rest of the folia (6 through 10) are the ventral paraflocculus. These folia include the area suggested by Miles and Lisberger (1981) to be involved in motor learning in the VOR. They found that injection into the ventral paraflocculus (rostral to posterolateral fissure) did not demonstrate input from the vestibular ganglion. None of the retrograde studies in monkeys include histologic examination of the vestibular neuroepithelium to address which end organ is innervated by vestibular afferents that project to the flocculus or ventral paraflocculus.

We demonstrate evidence of clear, albeit weak, utricular projection to the flocculus and ventral paraflocculus. The role of this region in VOR gain adaptation suggests an interesting potential role for these afferents, particularly in the mediation of the VOR gain appropriate for vergence during the translational VOR. However, the significantly greater inputs to the cerebellar cortex from the MVeMC, MVePC, interstitial nucleus of the eighth nerve, and nucleus y suggest that the disynaptic pathway is more dominant for conveying translational motion information to the flocculus and ventral paraflocculus. The convergence of these two sources of utricular-related inputs to the flocculus regions with visual-mediated signals relayed from the nucleus prepositus hypoglossi and the dorsal cap of the inferior olive (Langer et al., 1985b; Takeda and Maekawa, 1989) could underlie the mechanisms for visual modification of the translational VOR.