Properties of central vestibular neurons fired by stimulation of the saccular nerve

Brain Research, 143 (1978) 251-261 ~) Elsevier/North-Holland Biomedical Press

251

P R O P E R T I E S OF C E N T R A L V E S T I B U L A R N E U R O N S F I R E D BY STIMUL A T I O N OF T H E S A C C U L A R N E R V E

V. J. WILSON, R. R. GACEK*, Y. UCH1NO, A. J. SUSSWEIN The Roekefeller University, New York, N. Y. 10021 and University d Massachusetts Medical Sehool, Worcester, Mass. 01605 (U.S.A.)

(Accepted July 7th, 1977)

SUMMARY In 4 cats all vestibular afferents in one labyrinth except those innervating the saccular macula were transected and allowed to degenerate. 23-53 days after the initial surgery the central connections of the remaining saccular nerve were studied under chloralose anesthesia. Stimulation of the saccular nerve evoked N1 field potentials in the ipsilateral lateral and descending vestibular nuclei; little or no field potential activity was seen in the superior nucleus. The distribution of field potentials overlapped with that of neurons of origin of the vestibulospinal tracts. Forty-two neurons in the ipsilateral vestibular nuclei, many in the lateral nucleus, responded, often monosynaptically, to stimulation of the saccular nerve with single or double shocks; some of the neurons projected to the spinal cord. All saccular-fired neurons were tested for commissural actions by stimulation of the contralateral vestibular nerve. Many were facilitated, almost none were inhibited. In agreement with earlier work, we conclude that commissural inhibition may be a property of the canal system only.

INTRODUCTION Relatively little is known about the properties of central vestibular neurons receiving input from the maculae, compared to the wealth of information available about properties and projections of vestibular neurons receiving input from the semi-circular canals. We have recently developed a chronically denervated preparation in which only one macular nerve remains intact that can be stimulated in isolation in acute experiments 23. Because selective preservation is achieved more reliably for the * Address as of July 1, 1977: Department of Otolaryngology, Upstate Medical Center, Syracuse, N.Y. 13210, U.S.A.

252 saccular than for the utricular nerve, we have now made use of the chronic saccular preparation to study several aspects of the central projection of the saccutar nerve. We have examined the pattern of distribution of saccular-evoked activity within the vestibular complex and have compared this distribution with that of neurons giving rise to the vestibulospinal tracts. This emphasis on the vestibulospinal projection was designed to extend previous work, which demonstrated that stimulation of the saccular nerve evokes short-latency potentials in neck extensor motoneurons '-'3. In addition we have tested whether saccular driven neurons are subject to commissural inhibition. This inhibition is prominent in the canal system where it was first studied 17, but does not seem to affect tilt-sensitive neurons in the lateral nucleus ~s. In this paper we will show that field potentials evoked by stimulation of the saccular nerve are distributed in the vestibular nuclei in reasonable accordance with anatomical observations on the termination of saccular afferents. We studied saccular input to a number of vestibular nucleus neurons, some of them projecting to the spinal cord. Very few neurons were inhibited by stimulation of the contralateral vestibular nerve. METHODS

A natomical procedures The surgery for selective preservation of the saccular nerve inchronic cats has been described recently 23 and will be reviewed only briefly. The entire superior division of the vestibular nerve, as welt as the posterior ampuUary nerve, were transected at the level of Scarpa's ganglion under pentobarbital anesthesia: the posterior canal nerve was transected somewhat proximal to the ganglion. The cochlear nerve was left intact. After the acute experiments the animals were perfused with l0 ~ formalin following saline flush. The temporal bones were processed by Rasmussen's Sudan black B technique to stain the myelinated nerve fibers before decalcification. The membranous labyrinths with the attached nerve components were dissected, embedded in t5~o gelatin and sectioned at 15 #m on a freezing microtome. Each labyrinth was reconstructed from serial sections to determine the results of the nerve transections.

Acute experiments Initial surgery was performed under halothane (Fluothane, Ayerst Laboratories) anesthesia which was replaced, usually after the implantation of stimulating electrodes in the labyrinth, by a chloralose (initial injection 50 mg/kg, i.v.). At the same time the cats were paralyzed by intravenous injection of gallamine triethiodide (Flaxedil, American Cyanamid) and artificially respired; bilateral pneumothorax was performed routinely. Femoral blood pressure was, when necessary, maintained between 80 and 120 m m Hg by intravenous infusion of metaraminol bitartrate (Aramine, Merck) in normal saline, 80 #g/ml. Rectal temperature was maintained between 36 and 38 °C. Two fine silver wires, insulated except for the spherical tip, were implanted in each labyrinth to stimulate the vestibular nerve bipolarly in. For sumulafion of the saccular nerve on the operated side one electrode was usually placed near the saccular

253 macula, the other near the whole vestibular nerve. Stimuli delivered to these electrodes consisted of constant current, 0.1 msec, rectangular pulses. The middle part of the cerebellum was aspirated and glass micropipettes with resistances of 1-2 ME~, filled with 2 M NaCI saturated with Fast Green FCF, were inserted into the vestibular nuclei for recording of field potentials and extracellular unit activity. When the effect of contralateral vestibular nerve stimulation on single unit activity was studied, this was usually done by processing data on line with a Digital Equipment Corporation PDP 11-45 computer. The computer was programmed to accept data for poststimulus time (PST) histograms. Many locations at which good field potentials were seen or unit responses were studied were marked by electrophoretic injection of Fast Green dye from the recording electrode 2°. Dorsal laminectomy was performed for placement of electrodes into the C1-C3 region. These electrodes consisted of sharpened tungsten wires, insulated to 0.5 mm from the tip, and were used for monopolar stimulation of the lateral and medial vestibulospinal tracts (LVST and MVST); stimuli consisted of 0.2 msec rectangular pulses. The placement of the electrodes near the known location of the two tracts TM 1:3 was performed while recording antidromic field potentials in the lateral vestibular nucleus and in the medial nucleus or MLF. In one experiment (Sac 34) a three-electrode array was placed at C5-6 to stimulate vestibulospinal axons at this level. At the end of the experiment the position of the tip of each spinal stimulating electrode was marked by passing 20 #A of cathodal current through the electrode for 15 sec. The animal was sacrificed and the brain stem and spinal cord were removed and fixed in 10~;i formol saline. 100/~m frozen sections were cut in the plane of the electrode tracks. The locations of lesions in the spinal cord were determined in thionin-stained sections. The brain stem sections were stained by the method of Kliiver and Barerra, and histological reconstruction was used to determine the recording locations in the vestibular nuclei. RESULTS Satisfactory results were obtained from 4 cats with postoperative survival times after partial vestibular denervation ranging from 23 to 53 days: SAC 27 (53 days), SAC 28 (48 days), SAC 34 (23 days) and SAC 36 (34 days). In all these animals the superior vestibular division was completely degenerated. In SAC 27 and 28 the posterior ampullary nerve was also degenerated. In SAC 34 and 36 peripheral posterior ampullary nerve fibers remained, presumably because of the proximal location of the transection relative to the ganglion. Field po ten tials

Stimulation of the vestibular (saccular) nerve evoked small, slowly rising depolarizations (N1 potentials) in the vestibular nuclei, similar to those observed earlier in acute 6 and chronic 23 cats upon stimulation of the saccular nerve. N1 threshold ranged from 70 to approximately 150 #A. The latency of the foot of the N1 potential was 0.7 1.9 msec (Fig. 2C1; mean ~ 1.1 + 0.3 S.D.); peak latency was 1.2-2.5 msec

254

A

B

. 0

:+ i

i!i!::!

O

0

tl

,

o oo ~ ~

s

p~ o,:) i;~:

®

o

.

.

M

4

::?' O :,i! O~!!!i:.!!i!ili:.i:ii o

6

-l~0o~vii!iiiil

ii?!::

- 1200~V

+" t ~

I

9

0

I

I

I mV

SAC28

/-~ \,~..j) SAC.34

I

msec

200~V Q

O

©

msec

8

0

0

O

LVST

0

® VST

Fig. 1. Distribution of orthodromic and antidromic field potentials on a schematic dorsal view o f the vestibular nuclei. A : amplitude of field potentials evoked by stimulation of the saccular nerve and recorded at various sites in the vestibular nuclei. Results of three experiments mapped separately. B: amplitude of antidromic field potentials evoked by stimulation of LVST and MVST at C1 :~. Results of three experiments pooled. For both A and B potential amplitudes were measured as shown in insets: size of circles is proportional to these amplitudes. Abbreviations:D, descending nucleus; L, lateral nucleus; M, medial nucleus; S. superior nucleus. Hatched lines indicate approximate boundaries between nuclei.

255 (Fig. 2C~; mean - 1.5 ± 2 S.D.). Maximal N1 amplitude did not exceed 500 #V). In three experiments systematic tracking was performed in three of the vestibular nuclei: the lateral (Deiter's), descending and superior. Anatomical observations indicate that saccular afferents distribute mainly to the lateral and descending nuclei, whereas the superior nucleus receives canal afferents only .~,19. Significant early activity in the superior nucleus would therefore indicate sparing of posterior canal afferents or sprouting of saccular afferents into the nucleus. Activity in the lateral and descending nuclei would confirm previous anatomical observations physiologically. We did not examine group Y, where saccular afferents terminate 5 and where saccular evoked field potentials were seen by Hwang and Poon 6 and did little tracking in the medial nucleus. The results of three experiments are summarized in Fig. 1A. Little or no field potential activity was recorded in the superior nucleus. Those potentials that were recorded were very small and may have partly been due to volume conduction from more active sites. There were always potentials that started at the border between the superior and lateral nuclei, were present in the lateral nucleus, and extended some distance into the descending. In one experiment (SAC 36) tracks were made very caudally and activity was recorded in the descending nucleus as far as 3 mm caudal to the lateral nucleus. Fig. I B shows the distribution, in the same experiments, of antidromic field potentials evoked by electrodes placed in the upper cervical cord in the vicinity of the LVST and MVST. There is reasonable separation of the two tracts in this areal1,13. The two groups of field potentials are therefore probably due mainly to antidromic stimulation of the two tracts, but some cross-contamination may be present. In agreement with previous observations on field potentials and single neurons 1,'-',~, MVST field potentials distributed more medially than LVST field potentials, but there was significant MVST activity in the lateral nucleus, it is clear from Fig. 1 that the distribution of sacculus-evoked field potentials overlaps with that of the antidromic fields of both vestibulospinal tracts. Re,v~onses of single neurons 14. Responses to stimulation of the saccular nerve, Forty-two neurons responded to stimulation of the saccular nerve with a mean threshold of 2.2 times N1 threshold (2.2 N1T). Two types of response patterns were observed and samples of these patterns are shown in Fig. 2. The neuron in Fig, 2A responded to a single shock, security of transmission gradually increasing as the stimulus strength was raised from threshold (1.8 N1T). Responses of this and similar neurons were usually not enhanced if double instead of single shocks were used; occasionally the response to the second stimulus was weaker. The response of a second type of neuron (18/42) was facilitated if double shocks were used. Facilitation was often shown by greater likelihood of firing to the second shock than to the first. In many instances the two stimuli were equally likely to evoke a response, but the second stimulus produced more synchronized firing at a shorter latency. The neuron in Fig. 2B illustrates all of these phenomena : increase in firing probability, shortening of latency and synchronization. Our sample probably

256

Z

iii . Z

44

4

•

@

044

@404" •4 <0 4044 ~ @ 4 0 0 4 4 044444•444 444 0 4 4 4

LP

o

4~

8

0 ,,t 44.1 04 •

4-

4 ~1-

E

@0~ 44

@@4004 4110444 4

0440044 @4<044 < 0 4 0 4 4 4 " 04 ~4

O"

,q.

.o

c o

rY~

.=_

| ,,11~

E

! '

'

ditMll

L

/ tt)q!dt

I !

¢

.

.

.

.

-'~

8ko

4* I

t-.~

,--

¢~

Fig. 2. Responses of single neurons to stimulation of the saccular nerve. A : neuron that responds well to a single stimulus. In 1, stimulus to the saccular nerve at 1.8 NiT precedes a 30 ~A stimulus to the MVST (two upward arrows). When the neuron responds to the orthodromic stimulus (spike retouched) the antidromic stimulus is blocked by collision. In 2, the neuron responds securely on top of an N1 potential to a stimulus 2.8 NIT. BI..~: neuron responding much more fre,quently to the second of two shocks to the saccular nerve at 3.3 NIT. In the PSP histogram the stimuli are indicated by two upward arrows. C: latencies of the foot (1) and peak (2) of the sacculus-evoked N1 potential in the 4 vestibular nuclei, and latencies of responses of single neurons (3).

257 underestimates the fraction of neurons showing facilitation because we did not always search for neurons with double-shock stimulation. Thus, neurons firing only to a second shock would have been missed. The earliest latency of the response of each neuron, whether evoked by single or double shocks, is shown in Fig. 2C3. Latency varied rather widely, but most of the responses were between 1.2 and 2 msec. There was no tendency for mean latency of neurons responding best to the first stimulus to be earlier than mean latency of neurons responding best to the second. Comparison of the latency of neuron firing to the latency of the foot and peak of the Na potential (Fig. 2C1,~) suggests that firing with latencies of 1.7 msec or less was monosynaptic. The long tail of the distribution of N1 latencies makes it difficult to interpret whether neuron firing later than 1.7 msec is mono- or polysynaptic. Fig. 2C3 shows that the sample consists of 26 neurons in the lateral nucleus (these were found throughout the nucleus rostrocaudally; approximately two-thirds were located ventrally), 8 in the descending, 6 in the superior and 2 in the medial. There are presumably many lateral nucleus neurons both because there is good saccular input to this nucleus (Fig. IA) and because a big fraction of our tracks was made there. The few neurons in the superior nucleus were found despite the almost complete absence of field potentials in this cell group. B. Neurons projecting to the spinal cord. Sixteen neurons were activated antidromically by stimulation of the spinal cord at C1 or C3. In one case the antidromic response was not clearly visible in the antidromic field potential, which nevertheless blocked the response to labyrinth stimulation by refractoriness. In all other instances the antidromic response was clearly visible (as in Fig. 2A1) and its latency could be measured. Antidromic latency ranged from 0.5-1.4 msec, but was mostly between 0.5-0.7 msec. Some of these neurons projected to more caudal levels: in the one experiment in which we stimulated at C5-6, one of three neurons tested projected to this level. Fifteen neurons either responded antidromically only to stimulation with the medial or lateral spinal cord electrode, or else there was a difference in threshold between the electrodes of at least 3:12,15 . On this basis the axons were classified as being in the MVST or LVST. All 8 LVST neurons were in the lateral nucleus. Of the 7 MVST neurons, 5 were in the lateral nucleus and 2 in the descending (cf. refs. 2 and 15). The neurons projecting to the spinal cord responded to stimulation of the saccular nerve at latencies of 1.2-2.4 msec and several showed facilitation with double shocks. If the time for orthodromic firing is added to the antidromic latency for each neuron, the resulting time for impulses to reach the upper cervical cord is 1.9-3.2 msec for the LVST, 1.8-2.4 msec for the MVST. C. Commissural action on sacculus-driven neurons. All 42 neurons were tested for commissural effects evoked by stimulation of the contralateral vestibular nerve, most with the aid of PST histograms (for example, Fig. 3). Table I shows the pattern of effects that was caused by the conditioning stimuli, which usually consisted of triple shocks. By far the most common action (50 ~ ) was facilitation, which is illustra-

258 A 005

o[ 0

-

~ _ ~I__..~JD

0.05 "~.

0 j

nnnon

nnn

Inn

~]n nn _ _ _ J ~

n

rl

n

n

~n nnn

.......~ .....

n~n

B

~m~

o nn~ n

FIO

n

nnn ........j~nnn_o~_iL_:o_~

0.15 F C

0.1

5 msec

0.05 -

Fig. 3. Commissural facilitation of a neuron in Deiters' nucleus. This neuron responded monosynaptically to stimulation of the ipsilateral saccular nerve (latency 1.2 msec). A: spontaneous activity. B: effect of three stimuli to the contralateral vestibular nerve at 1.9 NIT. C: three contralateral stimuli at 3.3 NxT. Upward arrows in C show time of stimuli.

ted in the histograms of Fig. 3. In this neuron the threshold for facilitation was near 1.9 N I T (Fig. 3B); at 3.3 N I T there was considerable facilitation (Fig. 3C). F o r 17 n e u r o n s in which it was measured the m e a n threshold of facilitation was 3.0 N1T (range 1.5-4.5). A few n e u r o n s (12~o) were inhibited, with a mean threshold (n - : 4) of less t h a n 1.8 N i T ; i n h i b i t i o n was less complete than u s u a l l y observed in the canal system. With a few other n e u r o n s ( 1 0 ~ ) the effect was uncertain, and quite a few n e u r o n s ( 2 9 ~ ) were unaffected. N o significant difference was f o u n d between the fraction of facilitated, inhibited a n d unaffected n e u r o n s for the three nuclei studied (:Z2 test); however, the small size o f the sample makes it likely that all but the most obvious differences would have been missed. N i n e spinal-projecting n e u r o n s in the lateral and descending nuclei were facilitated by contralateral vestibular nerve stimulation, one was inhibited. TABLE I Commissural actions Nucleus

Number oJ

Inhibition

Facilitation

Uncertain

3 1 l 0 5 (12)

15 4 2 0 21 t50)

3 I 0 0 4 (10)

No effect

neurons

Lateral Descending Superior Medial Total

26 8 6 2 42 (100%)

< 2 2 12 (29)

259 DISCUSSION In an earlier paper on the chronic preparation with selective sparing of macular branches of the vestibular nerve, we have presented arguments that the activity evoked by stimulation of these branches is not due to spread to other structures or to plasticity '~3. These arguments are reinforced by the present results. Thus, field potentials evoked by stimulation of the saccular nerve were recorded in the lateral and descending vestibular nuclei but not in the superior. This finding is in reasonable agreement with experiments in which the exposed sacculus was stimulated 6 and with expectations based on anatomical studies ~,19which show that the lateral and descending nuclei receive saccular afferents while the superior nucleus receives only canal afferents. These observations indicate that no significant numbers of canal afferents were spared by the original surgery, and argue against extensive sprouting of the intact saccular afferents into areas previously innervated by canal afferents: if such sprouting had occurred it is likely that we would have found more labyrinth-driven activity in the superior nucleus. The latency of the foot of the N1 potential evoked by stimulation of the saccular nerve (0.7-1.9 msec, mean 1.1) is comparable to that reported by Hwang and Poon 6, 0.9-1.5 msec, and only slightly later than that of field potentials usually observed in the lateral nucleus (0.8-0.9 msec, ref. 24) when the whole vestibular nerve is stimulated. On the other hand the latency of the N1 peak (mean 1.5 msec) is considerably longer than that of the typical N1 potential, usually 1.0-1.1 mseca4, 24. This longer latency must be due to the small size of many saccular afferent fibers 5. The latency range that we suggest represents monosynaptic firing of single neurons, 1.0 to about 1.7 msec, overlaps with but is later than that of Hwang and Poon ~ (1.0-1.4 msec). Judging from Fig. 2A of their paper, our N1 potentials were smaller than theirs, suggesting a weaker input. This difference in input may be due to different anesthetics: their animals were anesthetized with ketamine and ours with chloralose. Difference in strength of input could be responsible for the small difference in latency of neuron firing. Most of the neurons that we studied were in the lateral nucleus, although some were in the descending and even in the superior nuclei. Several neurons in the superior nucleus responded at latencies later than 1.7 msec (Fig. 2C3) and they may have been fired disynaptically. On the other hand some responses could have been due to stray saccular fibers or to an occasional surviving canal afferent. A disynaptic input from the sacculus to the superior nucleus would be consistent with earlier observations on utricular influence on neurons in this nucleus. Some neurons respond weakly to lateral tilt12; this response is presumably due to polysynaptic connections from utricular receptors. A number of properties of saccular-driven vestibular neurons shed light on properties of saccular-evoked PSPs, mainly contralateral IPSPs and sometimes ipsilateral EPSPs, recorded in dorsal neck motoneurons 23. Thus, mean threshold for firing of vestibular neurons in the present experiment, near 2 NIT, is similar to the typical threshold of synaptic potentials in motoneurons. In the previous study it

260 was also observed that double shocks to the saccular nerve were often more ellective in evoking PSPs than were single shocks. We have found that firing of many vestibular nucleus neurons is facilitated by double shocks, and this could account for the greater effectiveness of double shocks in evoking PSPs. Finally, calculated arrival times of sacculus-evoked impulses at C1-C~ are 1.9-3.2 for LVST fibers, 1.8-2.4 lbr MVST fibers. Adding a segmental delay of 0.5-1.0 msec to these times essentially accounts for the latency of saccular-evoked synaptic potentials in neck motoneurons (Fig. 5 in ref. 23) without the need to postulate a spinal interneuron in the pathway. Since the MVST contains inhibitory fibers3,Z2, 25 the MVST neurons we studied, whose axons may be in the contralateral or ipsilateral MLF, are likely to be involved in producing IPSPs in contralateral neck motoneurons. The LVST is excitatory '~'-'. and some of the LVST neurons in our sample are presumably involved in the production of EPSPs in ipsilateral neck motoneurons. Fibers in both tracts may extend to more caudal levels and participate in sacculus-evoked vestibulospinal reflexes acting on limb muscles4, 21. Second-order neurons that are excited by ipsilateral horizontal rotauon are typically inhibited by contralateral rotation, and a similar situation exists tbr neurons with input from vertical canals. Such type I neurons are usually inhibited at short latency when the contralaterat vestibular nerve is stimulatedS,9,1L This commissural inhibition is quite precisely organized, so that there is excitation from one and inhibition from the other ampullary nerve of each coplanar canal pair 7. In contrast to commissural inhibition of canal second-order neurons. Deiters' neurons affected by lateral tilt, i.e. mainly by stimulation of the utricle, are facilitated by stimulation of the contralateral vestibular nerve is. The facilitation has a high mean threshold of about 3.3 N1T and a long latency, and appears ~o involve pathways through the reticular formation. As described in Results a large proportion of affected sacculus-related neurons in the lateral and descending nuclei was facilitated. and only few neurons were inhibited by contralateral vestibular nerve stimulation. The mean threshold of facilitation was 3.0 N1T. It is very likely that the facilitation is comparable to that of Shimazu and Smith ts and is mediated through the reticular formation. It therefore seems a reasonable generalization that commissural inhibition, which enhances the response of second-order neurons 9 is a property of canal- but not macula-related central neurons. ACKNOWLEDGEMENTS Supported in part by Grants NS 02619 and NS 08451 from the National Institutes of Health. A.J.S. was a recipient of N.I.H. Postdoctoral Fellowship 5 F32 NS 05011.

REFERENCES 1 Akaike, T., Comparison of neuronal composition of the vestibulospinal system between cat and rabbit, Exp. Brain Res., 18 (I973) 429-432.

261 2 Akaike, T., Fanardjian, V. V., Ito, M., Kumada, M. and Nakajima, H., Electrophysiologica| analysis of the vestibulospinal reflex pathway of rabbit. I. Classification of tract ceils, Exp. Brain Res., 17 (1973) 477-496. 3 Akaike, T., Fanardjian, V. V., lto, M. and Ohno, T., Electropbysiological analysis of the vestibulospinal reflex pathway of rabbit. 1I. Synaptic actions upon spinal neurones, Exp. Brain Res., 17 (1973) 497-515. 4 Anderson, J. H., Soechting, J. F. and Terzuolo, C. A., Dynamic relations between natural vestibular inputs of forelimb extensor muscles in the decerebrate cat. 1. Motor output during sinusoidal linear accelerations, Brain Research, 120 (1977) 1 16. 5 Gacek, R. R., The course and central termination of first order neurons supplying vestibular end organs in the cat, Acta Oto-Laryngol., Suppl. 254 (1969) 1 66. 6 Hwang, J. C. and Poon, W. F., An electrophysiological study of the sacculo-ocular pathways in cats, Japan J. Physiol., 25 (1975) 241 251. 7 Kasahara, M. and Uchino, Y., Bilateral semicircular canal inputs to neurons in cat vestibular nuclei, Exp. Brain Res., 20 (1974) 285-296. 8 Mano, N., Oshima, T. and Shimazu, H., Inhibitory commissural fibers interconnecting the bilateral vestibular nuclei, Brain Research, 8 (1968) 378-382. 9 Markham, C . H . , Midbrain and contralateral labyrinth influences on brain stem vestibular neurons in the cat, Brain Research, 9 (1968) 312 333. 10 Markham, C. H., Yagi, T. and Curthoys, I. S., Influence of the contralateral labyrinth on resting and dynamic activity of cat vestibular nucleus cells, Neurosci. Abstr., 2 (1976) 1059. 11 Nyberg-Hansen, R., Functional organization of descending supraspinal fibre systems to the spinal cord. Anatomical observations and physiological correlations, Ergebn. anat. Entwicklungsgesch., 39 (1966) 1-48. 12 Peterson, B. W., Distribution of neural responses to tilting within vestibular nuclei of the cat, J. Neurophysiol., 33 (1970) 750-767. 13 Petras, J. M., Cortical, tectal and tegmental fiber connections in the spinal cord of the cat, Brain Research, 6 (1967) 275-324. 14 Precht, W. and Shimazu, H., Functional connections of tonic and kinetic vestibular neurons with vestibular afferents, J. Neurophysiol., 28 (1965) 1014-1028. 15 Rapoport, S., Susswein, A., Uchino, Y. and Wilson, V. J., Properties of vestibular neurons projecting to neck segments of the cat spinal cord, J. Physiol. (Lond.), 268 (1977) 493-510. 16 Shimazu, H. and Precht, W., Tonic and kinetic responses of cat's vestibular neurons to horizontal angular acceleration, J. Neurophysiol., 28 (1965) 989-1013. 17 Shimazu, H. and Precht, W., Inhibition of central vestibular neurons from the contralateral labyrinth and its mediating pathway, J. Neurophysiol., 29 (1966) 467 492. 18 Shimazu, H. and Smith, C., Cerebellar and labyrinthine influences on single vestibular neurons identified by natural stimuli, J. Neurophysiol., 34 (1971) 493-508. 19 Stein, B. M. and Carpenter, M.B., Central projections of portions of the vestibular ganglia innervating specific parts of the labyrinth in the rhesus monkey, Amer. J. Anat., 120 (1967) 281 318. 20 Thomas, R.C. and Wilson, V.J., Precise localization of Renshaw cells with a new marking technique, Nature (Lond.), 206 (1965) 211 213. 21 Watt, D. G. D., Responses of cats to sudden falls: an otolith-originating reflex assisting landing, J. Neurophysiol., 39 (1976) 257 265. 22 Wilson, V. J., Physiological pathways through the vestibular nuclei, Int. Rev. Neurobiol., 15 (1972) 27 81. 23 Wilson, V. J., Gacek, R. R., Maeda, M. and Uchino, Y., Saccular and utricular input to cat neck motoneurons, J. Neurophysiol., 40 (1977) 63 73. 24 Wilson, V. J., Kato, M., Peterson, B. W. and Wylie, R. M., A single-unit analysis of the organization of Deiter's nucleus, J. Neurophysiol., 30 (1967) 603-619. 25 Wilson, V. J. and Maeda, M., Connections between semicircular canals and neck motoneurons in the cat, J. Neurophysiol., 37 (1974) 346 357.