Effects of afferent signals from the extraocular muscles upon units in the cerebellum, vestibular nuclear complex and oculomotor nucleus of the trout

0306-4522/8953.00+ 0.00 Per8amon Press plc C 1989 IBRO

Neuroscience Vol. 31, No. 2, pp. 529-541. 1989 Printed in Great Bmain

EFFECTS OF AFFERENT SIGNALS FROM THE EXTRAOCULAR MUSCLES UPON UNITS IN THE CEREBELLUM, VESTIBULAR NUCLEAR COMPLEX AND ~ULOMOTOR NUCLEUS OF THE TROUT J. A. ASHTON, C. MILLERET and I. M. L. DONALDSON* Department

of Pharmacology,

University of Edinburgh,

1 George Square, Edinburgh EH8 9J2, U.K.

Ahstraet-The responses of single units in tbe cerebellum, the vestibular nuclear complex and adjacent regions of tbe brainstem and in tbt oculomotor nucleus were studied in dwrebrate, paralysed rainbow trout (Srtlmo g&&r&). Natural vestibular stimulation was provided by horizontal, sinusoidal oscillation of tbe fish and extraocular muscle agerents of the eye ipsilateral to the razording were activated either by passive eye-movement or by electrical stimulation of the trochlear (IV) nerve in the orbit. Unit responses to vestibular and/or orbital stimuli were examined in peristimulus-time histograms interleaved in time. In the cerebellum and brainstem, of I24 units exposed to both types of stimulus, 26 (21%) responded only to vestibular input, 26 (2 1%) were affected only by the orbital signal and 23 (18%) raoeived both signals. The remaining 49 units (39%) responded to mechanical stimulation of the head or body or to vibration; tbey were lahelied “polymodal” and discarded. The recording sites of 56 units were verified by histology; 30 were in the arehellum and 26 in tbc brainstem. Input from the eye muscles had excitatory or inhibitory effects upon the vestibular responses. Tbc effects of the orbital signal were usually phasic but rare tonic responses also occurred. About half (15 of 34) of the units which responded to passive eye-movement showed statistically significant differences in the magnitude of their respoases to horixontal and to vertical eye-movement. More units preferred horizontal movement (11) than preferred vertical passive eye-movement (four). Note that tbe plane of vestibular stim~ation was always boriaontal. In the region of the oculomotor nucleus, of 19 units. five (26%) gave vestibular responses only and three (16%) were affected only by the orbital signal; three units (16%) with polymodal responses were discarded. Of the eight units carrying both signals, histological confhmation that tbe recording site lay in tbe column of cells forming the oculomotor/trocblear nuclei was obtained in four. The responses and interactions were similar to those found in the brainstem. The results present two principal points of interest. 1. They reinforce the accumulating hody of evidence that, in species with widely different oculomotor and visual hebaviour, signals from extraocular muscle proprioceptors reach the v~tib~o-alar system; this, in turn, suggests that these signals may play some rather fundamental role in the oculomotor system. 2. Tbe hypothesis that extraocular muscle affercnts are involved in ocuiomotor control requires that an effect of these signals he apparent at the output of the system-for example, in tbc oculomotor nucleus-though. if the action were to alter the characteristics of the system rather than to act from moment to moment, such effects might not he detected in acute experiments. In fact, the results reported here confirm that such a signal is found in the oculomotor nucleus of a bony fish in acute experiments and that it can alter the effect of vestibular drive to units in that nucleus. thus. the evidence in this paper further supports the hypothesis that a proprioeeptive signal from the receptors in the extrinsic ocular muscles plays a part in tbe control of eye-movement.

The extrinsic ocular muscles (extraocular muscles, EOM) of many species of vertebrate contain stretch receptors25+3’(review of higher vertebrates in Ref. 37). The view held earlier in the century (e.g. by Sherrington”), that afferent signals from these recep tars were likely to be involved both in the control of eye-movements and also in visual perception, later

fell into disbelief (see Carpenter”). We have discussed some of the reasons for this change in opinion in a recent paper’ in which we also explained the reasons *To whom all correspondence

should be addressed.

Abbreviurions: EOM, extrinsic ocular muscles; PEM, pas-

sive eye-movement; PSTH peristimulus-time histogram; VN, vestibular nuclear complex; VOR, vestibule-~ular reflex.

which led us to form the hypothesis that afferent signals from the EOM might gain access to the oculomotor system at the level of the vestibular nuclear complex (VN). We have now shown that, in an amphibian’ which makes few eye-movements except in association with head-movement, and in the cat3 whose oculomotor and visual behaviour is much more complex, the vestibular nuclei receive signals from the EOM and that these signals can influence the response of central units to concomitant vestibular stimuli. In the pigeon also, which has quite complex oculomotor behaviour, we have found that EOM afferents. reach the cerebellum and vestibular nuclei.‘2 The vestibular nuclear complex projects both monosynaptically and through polysynaptic path-

530

J. A.

ASHTON

ways to the oculomotor nuclei3’ Vestibular pathways convey position and velocity information in such a way that head or body movements in one direction are accompanied by eye-movements of approximately equal amplitude and speed but opposite in direction. These smooth, reflex, movements comprise the slow component of the vestibulo-ocular reflex (VOR). In unpublished experiments on an amphibian (giant toad), whose vestibular nuclei we have shown to receive EOM afferents,5 we found evidence for these afferent signals also reaching the oculomotor (III nerve) nuclei where they were able to modify the effect of vestibular drive to oculomotor cells. Although these results suggest that muscle afferent information might be able to modify the VOR, in the experiments on the toad the vestibular signals were induced by electrical stimulation of the VIII nerve which does not provide a physiologically meaningful vestibular signal. The experiments reported in the present paper were undertaken to try to establish whether. in a vertebrate whose oculomotor behaviour is intermediate in complexity between that of amphibia and cats, the EOM send signals to the vestibulo-oculomotor system; positive findings would lend further support to the idea that such afferent signals play a part in oculomotor control. We wished, also, to look for further evidence of effects at the output of the oculomotor system, in the oculomotor nuclei. A bony fish, like the amphibian, can he used as a decerebrate preparation without compromising the connections to and from the orbit. The rainbow trout was chosen as a teleost which is easily available in specimens of a suitable size. Though the EOM of the trout do not seem to have been examined for the presence of stretch receptors, several other teleost species are known to have such receptors in their EOM; these include perch, pike and roach.25,26.32Central effects of stretching EOM have also been found in the goldfish.20 The finding2’ that cells are labelled in the fused trigeminalifacial (V/VII) ganglion of the carp after injection of horseradish peroxidase into the EOM. presents a striking parallel with the situation in mammals in which the cell bodies of EOM efferents are found in the trigeminal (Gasserian, V) ganglion. ‘“.29.30.37 It seemed probable, therefore. that the eye-muscles of the trout. also, would contain some type of stretch receptor. Preliminary accounts of results in the trout appeared in 19866 and 1987.’ EXPERIMENTAL

PROCEDURES

Preparation

Rainbow trout (Sa/mo gairdneri) were anaesthetized by immersion in a solution of benzocaine. 50p.p.m. in tap water. an abdominal vein was cannulatcd and the fish was mounted in a trough with the head held by a clamp on the upper jaw and stabihzed by bars which pressed on each side of the skull. Cooled. aerated. water containing benzocame

el al

(50p.p.m.) was then circulated through the mouth to pass over the gills into a trough whence it was returned to the aerator and recirculated. A dorsal craniotomy was made to expose the brain and the fish was decerebrated by crushmg the cerebral hemispheres and destroying the brain rostra1 to the optic tecta. In some experiments both optic nerves were also cut. Curare (tubocurarine I mglkg. i.v.) was then given and the circulating fluid was replaced by cooled. aerated water without benzocaine. The decerebrate preparation was maintained for the rest of the experiment without further anaesthesia. The dura mater was opened over the cerebellum and the optic tecta and the cerebrospinal fluid whrch welled up was allowed to clot to form a protective layer over the brain. No further protection was usually necessary and the preparation was stable enough to permit satisfactory single unit recording. Srimularion The trough and head-holder were mounted in a frame attached to a heavy, horizontal. servo-controlled turntable and natural vestibular stimulation was provided by sinusoidal oscillation at 0.38 Hz at amplitudes between k 15’ and k 20’ peak-to-peak in the horizontal plane. In the goldfish,Y another teleost, the gain of the VOR is constant at about -0.8 when tested with sinusoidal oscillatron of amplitude &20” between 0.125 and 0.5 Hz and there IS little phase-shift of the eye relative to the head at these frequencies. Extraocular muscle (EOM) afferents of the left eye (ipsilateral to the side of recording) were usualI> activated by passive eye-movement (PEM), as we have described previously,‘” except that the opaque contact-lens which applied movement to the eye was glued to the cornea in the present experiments. Amplitudes of PEM between 3 and 13’ at 175-275:/s were used. These amplitudes” and velocities” are within the normal range for the goldfish In some experiments PEM was not used but one or both trochlear (IV) nerves were exposed in the orbits and arranged, in a pool of mineral oil. for electrical stimulation The stimuli were three pulses 100 ps wide at a frequency of 2OOHz and an amplitude approximately twice the voltage threshold required to produce a central response. The apparatus was controlled by a laboratory computer on-hne as we have described previously.3 Recording

Glass capillary rmcroelectrodes, filled with 3 M NaCI. were inserted in the frontal plane. either through the cerebellum to reach the vestibular nuclear complex. or through the left optic tectum close to the midline, to reach the oculomotor (III/IV) nuclei. The responses of single units to vestibular stimulation alone. to PEM or electrical stimulation of the IV nerve alone, and to combinatrons of vestibular and orbital stimulation were collected as perisumulus-time histograms (PSTH) interleaved m ume. as we have described previously.‘.‘~” At the end of a successful recording track the shaft of the microelectrode was usually broken to leave the distal part in the brain and thus mark the recording track. Hisrologj

At the end of the experiment a pin was inserted to mark the recording region and the brain was fixed by transcardiac perfusion with 10% buffered formalin. After further fixation, the brain was embedded in celloidon or gelatin with the marker still in sifu. The marker was then removed and 50-pm sections were cut in the frontal or parasagittal plane and stained with Cresyl Violet for Nissl material Recording tracks were reconstructed using the marker track and the glass fragments of the broken-off microelectrodes. Further analysis qf responses to passive eye-morlemenl The excitatory responses to PEM from PSTHs of units which responded only to PEM. and the excitatory “inter-

531

Effects of extraocular afTeren signals in trout action” responses from units which received both vestibular and PEM signals were examined for evidence of preference for movement of the eye along one. rather than another, orbital radius. usiq the method whicr we have described previously.‘” Briefly, the responses to deflection along each radius were estimated by counting the total number of impulses in a time window (a certain number of PSTH bins) which contained the whole response. as judged by eye. The same size of window was used for each of the PSTHs to be tested in a group of four. The sizes of these responses were then compared in pairs using a statistical test devised by Diirrscheidt’4 and based on the binomial theorem. A unit was regarded as having “planar” preference if its response to PEM along one radius differed signiikantiy from that along at least one other radius in an orthogonal plane. The criterion of statistical significance was P d 0.025. This definition corresponds to that which we previously used for “radially-selective” responses to PEM in the visual cortex.’ RESL‘L’T.5

Responses in posterior brainsrem and cerebellum In the experiments whose object was to record from the brainstem in the region of the vestibular nuclei. 124 units were isolated which responded to vestibuiar stimuli or to an orbital signal. or to both vestibular and orbital stimuli. In all, 26 units (21%) responded only to the vestibular stimulation produced by horizontal. sinusoidal oscillation of the fish; 26 units (21%) were affected onl) by the orbital signal produced by pas-

sive movement of the left (ipsilateral) eye (PEM, 22 units) or by electrical stimulation of the fourth (trochlear) nerve in the orbit (four units). There were 23 units (19O/,) which received both a vestibular signal, since they were affected by horizontal oscillation of the fish, and an orbital signal since either (a) they were excited by this signal alone, or, (b) the orbital signal modified their vestibular response; often both (a) and (b) were true. The remaining 49 units (39%) responded to mechanical stimulation of the head or body (for example brushing of the skin, tapping the skull), or to vibration or tapping of the head-holder. Although many of these units also responded when vestibular or orbital stimuli were delivered, it is not possible to be certain of the effective source of input. Since these units often appeared to respond to more than one type of stimulus they were labelted “pofymodal”. They are considered further in the description of control experiments, below. The types of response are summarized in Table 1. Location of recording sites. On the basis of their depth from the surface of the cerebellum and from histological reconstruction of their recording tracks, 30 units were located in the cerebellum (in the corpus cerebelli see Ref. 27) and 26 were located in the brainstem in the region of the vestibular nuclei. Table 2 shows the types of responses obtained in each of those groups of units.

Table 1. Responses of 124 units in posterior brainstem and cerebellum of trout Stimulus Horizontal vestibular stimulation only Orbital signal onl! PEM Electrical stim. Horizontal vestibular stimuiat~on and orbital signal Vest. + PEhl Vest. -r electrical stim. “Polymodal” (see text)

No. units 26 (21%) 22 4>

26(21%)

18 5

23 (19%)

Total Table 2. Analysis of responses of 56 of the units of Table 1 according to location of recording site ~ol~modal responses excluded) Cerebellar vermis Horizontal vestibular stimulation only Orbital signal only PEM Electrical stim. Horizontal vest. stimulation and orbital sipnal PEM Electrical stim.

9 (30%) 7 3>

10 (33%)

8 3>

Ii (37%)

Total, cerebellar units Vestibular nuclear complex of brainstem Horizontal vestibular stimulation only Orbital sinnal onlv PEM Electrical sum. Horizontal vest. stimulation and orbital signal PEM Electrical stim. Total. brainstem units

30 11 (42%) 6 I>

7 (27%)

7 1>

8 (37%) 26

532

J. A.

ASHTOK

et

al.

Table 3. Tynes of vestibular response

TYW II

Type III

2 -

4 4

3 4

2

8

7

5 2

3 3

7

6

TYW

Cerebellar units Vestibular response onlj Vestibular + orbital response Total

I

Brainstem (vestibular nuclear complex) units Vestibular response only 3 Vestibular + orbital response 2 Total

5

Responses to vestibular stimulation. The criterion for response was cyclic modulation, as judged by eye, of the response in the PSTH of the unit’s activity during exposure to table oscillation alone. No attempt was made to quantify the responses, but, as we found previously in the cat,’ there were few instances in which there was any doubt about the presence of a response. Of the 49 responses to vestibular stimulation (26 vestibular only and 23 which also received an orbital signal), 35 could be classified according to the relation between the direction of table-movement and modulation of the

Plot

durotlon

2.5

..0.

24

units’ firing (Table 3; see Ref. 31 for definition of the response types). In units in both cerebellum and brainstem, Type II responses were the most common but Type III were only a little less frequent; this contrasts with our findings3 in similar experiments on cat vestibular nuclei where Type III responses were rare. The response of a unit in the brainstem, dorsal to the magnocellular vestibular nucleus (Deiters). to vestibular stimulation alone is illustrated in Fig. I. This unit’s firing was modulated principally by table movement away from the side of recording and so it

SWEEPS.

SCALE

BAR

50

F. 21E0102

IMPULSES/SEC

VESTIE

VESTIB

se

i00

150

200

550

0

50

100

190

200

25e

50

100

150

200

25E

VESTIB

i

0

50

100

150

200

250

0

Fig. 1.Set of four PSTHs of the responses of a unit dorsal to the magnocellular vestibular nucleus (Deiters) to natural vestibular stimulation by horizontal oscillation of the fish (VESTIB) at 0.38 Hz. peak-to-peak amplitude approximately + IS-. The PSTHs. which are made up of 24 stimulus presentations. have a bin-width of 1Oms and a total duration of 2.5s. The upper traces indicate table position: downwards deflection of the sinusoidal signal indicates table movement towards the left side (ipsilateral to the recordmg electrode). The PSTHs were interleaved in time so that the four histograms were recorded effectiveI> simultaneously. In this case the stimulus was identical for all the PSTHs Note. Type II vestibular modulation. The recording site of the unit is shown at EOl in Fig. 4.

533

Effects of extraocular affercnt signals in trout Plot

dvrotron

.?. 5 ..o.

4B

SWEEPS.

VESTIB

BAR 25 - NASAL

VESTIB

*

IMPULSES/SEC PEM

F. 2788506

rn nr

I R 0

VESTXB

SCALE VESTIB

50

*

UP

10B

158

200

251

PEH

TEHPOR

PEH

Fig. 2. Set of four interleaved PSTHs showing the responses of a unit rccordod deep in the ccrebcllar vermis to vcstibular stimulation alone (VESTIB; horizontal oscillation at 0.38 Hz, peak-to-peak amplitude approximately & 15’) and to vestibular stimulation plus passive eye-movement (PEM) in the horizontal plane and nasal quadrant of the orbit (VESTIB + NASAL PEM); in the horizontal plane and temporal orbital quadrant (VESTIB + TEMPOR PEM) and in the vertical plane and superior quadrant (VESTIB + UP PEM). The upper traces show the vestibular table position (top left) and the table position with the position of the eye-mover superimposed (upper right and lower panels.) The leading edge of the step change in the upper trace indicates dcgcction of the eye away from the primary position and the trailing edge shows its movement back to the primary position. The notched appearance of the step edges is due to quantixation by the analogue-to-digital converter; the movement was smooth. Details of vcstibular stimulation as in Fig. 1, The amplitude of PEM was approximately 13” at 275’1s. Other details as for Fig. I. Note: Type III vestibular modulation; phasic responses to PEM with the largest response in the horizontal plane, nasal orbital quadrant (top right). This unit shows “planar” preference-see text.

is classified as Type II. Its recording site is shown at EOf in Fig. 4. Effects of passive movement of the ipsilateral eye. Eighteen units showed modulation of their vestibular responses when PEM was added to the table oscillation and, in 15 of these units, the effect of PEM was phasic. In most cases there was a phasic increase in firing, sometimes followed by reduced discharge; four units showed purely inhibitory effects of PEM. These effects were similar to those which we have recently described in cat vestibdar nuclei3 Latencies, measured from the beginning of the eye-movement to the beginning of the response, ranged from 9 to 100 ms. For the 37 phasic excitatory responses (22 to PEM only; 15 modulations of vestibular response) the latencies appeared to fall largely into two groups; 17 units with latencies of 20ms or less and 14 units with latencies of 50 ms or more. Figure 2 illustrates phasic, excitatory interactions between vestibuiar stimulation and PEM in a unit

recorded deep in the cerebeilar vermis. There were no obvious differences between responses in the cerebellum and those in the vestibular nuclei. Three units showed sustained, “tonic” effects of PEM; since central responses of this type to stimulation of EOM proprioceptors have rarely been reported, the response of one of the “tonic” units is illustrated in Fig. 3. This figure shows (bottom right) a sustained response to sustained deflection of the ipsilateral eye nasally in the horizontal plane. There was little resting discharge (top left) and little effect when the eye was moved ~rn~rally-ho~zon~lly (top right) or vertically upwards (bottom left); thus, the unit shows a “planar” preference (see below). The unit of Fig. 3 also had a vestibular response; its recording site lay near the descending vestibular nucleus (Fig. 4). Twenty-two units responded to PEM of the ipsilaterat eye but not to horizontal vestibular stimulation; their responses consisted of phasic excitation

534

J. A. ASHTONer al. Plot

duration

2.5

..5.

24

SWEEPS.

CONTRG,

1

50

I00

I n I II 159

200

0

250

PEM

1I, 1111I

NASAL

I

0

IMPULSES/SEC

F. 26E0202

1

11 0 UP

SCALE BAR 25 TEMPORAL PEM

,,

l,linLL& 50

100

150

200

I

II

I IIUI

I

50

100

150

200

25e

50

100

150

200

25E

PEM

j 250

0

Fig. 3. Set of four PSTHs showing the responses of a unit in the region of the descending vestibular nucleus lo PEM alone. Top left: CONTROL (no stimulus). The upper traces show the position of the eye-mover: PEM amplitude and velocity were approximately 13’ and 275:/s, respectively. Conventions otherwise as for Fig. 2. Note: sustained response to PEM in the horizontal plane. nasal orbttal quadrant (bottom right) and absence of response to horizontal movement in the temporal quadrant (top right) and to vertical movement. The unit shows “planar” preferencc for the horizontal plane. The recording site of this unit is shown at E02 in Fig 4

and were similar to the phasic. excitatory interactions described above. “Planar” prqferences. Most units were tested with PEM along three orbital radii which were, usually: vertically-up. nasal-horizontal and temporalhorizontal. All the eye-movements, which started and finished with the eye approximately centred in the orbit, consisted of an excursion out and then back along a particular radius. Nineteen of 36 units tested in this way gave similar responses to movements along the three radii but the other 17 units showed statistically significant differences in the magnitude of the responses to PEM along orthogonal radii when these were tested as explained in Experimental Procedures. Interestingly, of these 17, 12 units preferred movement in the horizontal plane to vertical movement, and only five preferred the vertical to the horizontal plane. Preferences were significant at P Q 0.005 for nine units and at P C 0.025 in the other eight. The unit of Fig. 2 has “planar” preference since its responses in the horizontal plane (top right, bottom right) differ significantly from its response to upward movement. In this case the nasal and temporal responses differ significantly from each other, so there is some further specificity for the arc

through which the eye moves in the horizontal plane (see Discussion). Responses to electrical stimulation in the orbit. In a

few experiments. the effects of electrical stimulation of the trochlear (IV) nerve in one or both orbits were examined to help to localize the site of receptors responsible for the orbital signal, as we have described previously.‘.5 Responses to electrical stimulation of the trochlear nerve alone and interactions between the effects of electrical, and those of vestibular stimuli, were found, as Tables 1 and 2 show. The significance of these results is discussed belon. Similar control experiments to eliminate spurious responses were carried out for the observations on the cerebellum and brainstem and for those on the oculomotor nuclei; they are described at the end of the section on the oculomotor nuclei. Responses in oculomotor nuclei

In these experiments the recording electrode descended through the posterior part of the left optic tectum. across the optic ventricle then the cerebellar valvula. to pass close to the IV nerves where they cross the surface of the midbrain, then into the region of the oculomotor (III) and trochlear (IV) nuclei in

Effects of extraocular a&rent signals in trout

535

Table 4. Analysis of responses of 19 units in trout midbrain in the region of the oculomotor (IIIiIV) nuclei No. units Stimuius Horizontal vestibular stimulation only Orbital signal only PEM Electrical stim. Horizontal vestibular stimulation and orbital signal Vest. + PEM Vest. + electrical stim. “Polymodal” (see text)

5 (26%) 2 1>

3 (16%)

4 4>

8 (42%) 3 (16%) 19

Total

the midbrain. Figure 6 shows a reconstruction of two recording tracks of which the posterior passed through the oculomotor nucleus. The arrival of the electrode in this region could often be detected by the change in background activity as the recording tip left the cerebellar valvuia and entered the midbrain.

I

F26 7H

Fig. 4. Reconstruction of a transverse section of the trout brainstem to show a recording track (Track E) which is in the left side of the brain. The section is somewhat oblique to the long axis of the brainstem with the left side of the figure (as viewed) more rostral. Thus the electrode track passed just posterior to 8m and 8t which are visible on the kft of the figure but not on the right. VIIIn, eighth nerve; v4. fourth ventricle: 8m(D). magnocellular vestibular nucleus (kiters): 8d. descending vestibular nucleus: 81. tangential vestibular nucleus: 6r. abducens (VI) nucleus. rostra1 part; Sd. descending trigeminal tract; EOI. EO2. recording sites of the units whose responses are illustrated in Figs 1 and 3. respective+.

Units were tested for responses to vestibular and lo orbital stimuli as described in the previous section. As Table 4 shows, 19 units were recorded in the region of the oculomotor nuclei but, of these, three were “poiymodal” as defined above. Of the remaining units, five responded only to horizontal vestibular stimulation and three only to the orbital signal; the remaining eight units received both vestibuiar and orbital signals. The recording sites of four of the latter were identified histologically in the column of ceils forming the III and IV nerve nuclei. Figure 5 shows the response of a unit in the ocuiomotor (III) nucleus to vestibular stimulation (top left) which produced cyclical modulation of firing. The other panels show that there was powerful inhibition of the response when the IV nerve of the ipsiiateral orbit was stimulated electrically at various times during the vestibular stimulus. The IV nerve of the ipsilateral orbit arises, of course, in the IV nerve nucleus contralateral to the recording site. The recording site of the unit of Fig. 5 was at A02 in Fig. 6. In Fig. 7 is shown the response of a unit whose recording site was identified histologically in the posterior part of the III nucleus or. possibly. the anterior part of the IV nucleus (the column of ceils is continuous). This unit had a vestibular response and it was additionally excited by PEM of the ipsilateral eye. Its vestibular response (Fig. 7, top left) shows a small peak at about the time of maximum table velocity (100 on abscissa. corresponding to 1 s); this peak was greatly increased when PEM in the horizontal plane and the nasal quadrant was added (top right); PEM vertically upwards and in the temporal horizontal plane had much less effect (lower panels). Contra/ experiments. It was important to eliminate, as possible causes of the responses which were found, sources of signal other than those from vestibular stimulation due to table oscillation and those from orbital afferents. To eliminate possible visual signals, the optic nerves were cut in several experiments; the responses to PEM were not abolished and the results in these experiments did not appear to be different from those with the optic nerves intact. Thus visual stimulation can be excluded as an effective stimulus. A number of units which apparently responded to

536

J. A. Plot

dvrotion

2.5

..e.

48

bHTON

SWEEPS.

VESTIB

er al.

SCALE VESTIB

BAR 100 * IPSIL

IMPULSES/SEC N IV

F. 3QA0204

II

I/

1

0

VESTIB

i

/, 0

50 *

IPSIL

100 N

150

I

280

250

VESTIB

IV

I 50

0

100

150

200

250

i

50

*

IPSIL

100

N

150

200

IV

WNM ‘i 0

I

50

100

150

200

Fig. 5. Set of four interleaved F’STHs of the response of a unit in the left oculomotor (III) nucleus to natural horizontal vestibular stimulation and to electrical stimulation of the trochlear (IV) nerve in the ipsilateral orbit. Upper traces show table position (as in previous figures) in the top left panel. plus marker (downgoing tick) indicating the time of electrical stimulation of the nerve. in the other panels. Conventions otherwise as for Fig. 2. Note: “vestibular” response with marked inhibition by electrical stimulation of the IV nerve. Note also that stimulation of the IV nerve did not excite the unit as the initial response though there appears to be a “rebound” excitation following the inhibition when the stimulus was given about one quarter of the way mto the vestibular cycle (bottom left). The recording site of thts unit is shown at A02 in Fig. 6.

A03’ A04’

F39

250

6D

Fig. 6. Reconstruction of a parasagittal section of the trout brain a little to the left of the median plane. Anterior is to the left of the figure. The units recorded at AO2. A03 and A04 gave responses to vestibular stimulation which were inhibited by IV nerve stimulation. The response at A02. shown in Fig. 5. is typical of these. cbm. chg. cerebellum, molecular and granular layers. vcb. valvula of cerebellum: v3. third ventricle: v4, fourth ventricle: IVn, fourth (trochlear) nerves, 3, third (oculomotor) nucleus (dorsolateral part); 4. fourth (trochlear) nucleus: Track A. B. reconstructed recording tracks: AOI-A04. recordmg sites of four units.

I 25t

531

Effects of extraocular afferent signals in trout Plot

duration

2. 5

..=.

48

SWEEPS.

k0 VESTIB

SCALE VESTIB

VESTIB

+

UP

100

I

150

BAR l

isa

50

NASAL

ice

;0

PEM

VESTIB

*

TEMPO

F53A0104

IMPULSES/SEC PEM

is0

500

lS0

200

PEM

44

t 0

I

I

50

100

150

200

I

250

‘t

0

I

I

S0

100

I 250

Fig. 7. Set of four PSTHs of the responses of a unit in the ocuiomotor nucleus to horizontal vestibular stimulation alone (VESTIB) and to vestibular stimulation plus PEM m the horizontal plane and nasal quadrant of the orbit (VESTIB + NASAL PEM), in the horizontal plane and temporal orbital quadrant (VESTIB + TEMPO PEM) and in the vertical plane. superior quadrant (VESTIB + UP PEM). The recording site of the unit appeared to be at the posterior edge of the left oculomotor nucleus. Conventions as for Fig. 2 Note: upper left-“vestibular” response with small additional peak of excitation about the time of maximum table veloaty (100 on abscissa): additional excitation by PEM in the horizontal plane and nasal orbital quadrant (top right): much smaller effects of PEM in horizontal temporal quadrant and in vertrcal plane (bottom two panels).

could also be made to fire by tapping the turntable or the head-holder; in other cases light mechanical stimulation of the skin of the head or body was effective in drivtng them. All the units which responded to any of these stimuli (some of which also showed “vestibular” responses) were classed as “polymodal” and excluded from further examination. PEM

Natural vestibular stimulation was confined to the horizontal plane and the fish was positioned so that the long axis of its body was horizontal: in this position the lateral (horizontal) semi-circular canals lie approximately in the plane of the horizontal turntable but it is unlikely that stimulation was completely confined to the horizontal canals-a point which may be of some importance in considering “planar” interactions (see below).

The choice of stimuli

Evidence for projection of orbital proprioceptors to the brainstem and cerebellum

The reasons for preferring PEM rather than stretch of individual eye-muscles as a means of activating EOM receptors have been discussed in detail previously.‘.’ The most important points are that PEM mimics the state of affairs during a natural eyemovement much more closely than does stretching a single eye-muscle. and that. using PEM produced by suitable apparatus.‘.’ it is possible to compare the effects of displacement of the eye along different radii. effectively simultaneously.

The observations described in Results and summarized in Tables 1 and 3 seem to show conclusive11 that signals which arise from the orbit during PEM activate units in the trout brainstem and cerebellum. This claim depends. of course. on the elimination of sources other than orbital afferents for the effective signal when the eye is moved passively. Control experiments. described above. were performed to eliminate visual. vibrational and cutaneous sources for the signal. Units which responded IO vibration or

538

J. A. ASHTOS er al.

lo cutaneous stimulation were discarded because. in the presence of such input. it is impossible to interpret the responses which many of these units gave during PEM. The site qf the receptors responsible jar responses to passive eye-movement In previous experiments on the cat’ and toad’ we have used electrical stimulation of the intraorbital part of oculomotor nerves to test the hypothesis that the receptors involved in responses to PEM include receptors in the EOM. Electrical stimulation of intraorbital parts of oculomotor nerves has also been used in many experiments on the central projections of EOM afferents in other laboratories.‘,g As we have pointed out previously.3-’ the most convincing evidence that the effects during PEM are due to signals which arise predominantly, if not exclusively. from receptors in the eye-muscles is that such effects can be produced both by PEM and also by electrical StimuIation of an oculomotor nerve within the orbit. where it carries afferents from the muscle receptors In the trout we found that units in the cerebellum and brainstem could be affected by electrical stimulation of the intraorbital part of the ipsilateral trochlear (IV) nerve. Some units were excited, in others the effect was inhibitory upon the response to vestibular stimulation. Stimuiation of the intraorbital part of the IV nerve will also. of course. activate trochfear motoneurons antidromically and might. therefore. cause activation of recurrent collaterals of these motoneurons. Such collaterals exist in the cat but do not seem to extend beyond the oculomotor nucleus.“.‘” No info~ation appears to be available on the trout. but. in other teleosts. collaterals of oculomotor and trochlear motoneurons have not been found I119It seems safe to conclude. therefore. that the effects on bramstem and cerebellar units produced by electrical stimulation of the IV nerve in the orbtt are indeed due to activatton of muscle afferents. afferents. We also used electrical stimulation of the IV nerve in the orbit in experiments m which we recorded from the oculomotor region. In these experiments there might appear to be a greater risk of confusion due to antidromjc activation since the column of cells which forms the III and IV nerve nuclei is continuous SO that it is sometimes difficult to be certam from a Nissl section whether a recording site is in the most posterior part of the oculomotor. or the most anterior part of the trochlear nucleus. The trochlear nerves cross on leaving the midbrain though. in another teieost at least (carp’9). occasional trochlear neurons project to the ipsilaterai nerve. However. it is most unlikely that any of the effects in our experiments on the ocufomotor region can be explained by, antidromic invasion, Most of the units were in the anterior part of the cell column. and thus m III rather than 1‘1 nucleus on histological criteria: more importantly. all but one of the effects were inhibitory on the responses

of the units to vestibular stimulation. whereas. had we been recording in the IV rather than the III nucleus, and, had the effect been due to one of the rare ipsilaterally projecting trochiear motoneurons. it would, presumably. have been seen as an excitator> response. Since both PEM and electrical stimulation produce centraf responses in the trout and affect the responses of units to vestibular stimulation. the present results provide evidence for the existence of stretch receptors in trout EOM as well as for their projections to certain regions of the CNS. Responses to passive eye -mor;ement

Most of the responses to PEM alone. and most of the effects of PEM upon the vestibular responses. were phasic and were very similar to those which we have previously reported in the cat.’ Three units. however, gave sustained responses; one of these. illustrated in Fig. 3, appeared to be quite specific for PEM in the horizontal plane for movement from the primary position towards the nose. The response had a very long latency and this somewhat complicates interpretation of its possible function. It would be most interesting to be able to test such units with a range of amplitudes and velocities of PEM to try to ascertain whether they carry an eye-position signal: unfortunately the arrangements used in the present experiments did not allow this. One can conclude only that these rare units are potential candidates for carriers of a position signal. ,?$ects L$ the orbital axerent cerebe[~um and restibufar nuclei

signai on units ill

The main purpose of the experiments was to test the hypothesis that. in a teleost fish. the VN receives a signal from EON afferents. The results presented above appear to show that this is Indeed the case W’e were then interested to discover whether there was evidence of specificity of the action of the EOM afferent signal of the kind which we have previously found in the cat visual cortex4 and vestibular nuclei.’ It cat visual cortex we found that units often show a preference for PEM in one plane (e.g. horizontal) rather than in the orthogonal plane and that some units also appear to prefer eye-movement in one quadrant of the orbit (e.g. nasal). to movement in the same plane but the opposite quadrant (temporal). In the cat VN we found that some units appear to have a rather similar type of planar specificity.! In the trout. the present results show that 15 of the 34 units (44%) with PEM responses which were tested showed evidence of planar specificity since their response to movement along at least one quadrant in one plane was significantly different (when tested as described in Experimental Procedures) to the response to movement along at least one quadrant in the orthogonal plane. it is particularly interesting that. of the 15 units. 11 preferred horizontal to vertical eye-movement. The

Effects of extraocular afferent signals in trout most important point is that, of the 10 units whose vestibular responses were modified by PEM, eight showed preferences for horizontal eye-movement, Our experiments selected units which responded to horizontal vestibular stimulation since this was the only plane in which vestibular stimulation was delivered. Though it would be unrealistic to suppose that there was no stimulation at all of vertical canals, units with input from these canals are likely to have been recorded as having no vestibular response. The fact that units selected by the experimental conditions for their responses to horizontal vestibular stimulation are also more likely to be affected by horizontal than by vertical PEM suggests that there is a convergence of signals which is orderly in the sense that many units which carry the horizontal vestibular drive signal will also receive input from the EOM proprioceptors during the resulting eyemovement. Some units also show selectivity for one quadrant of the orbit which is similar to that which we have found in visual cortex.’ We have recently found’? a great preponderance of horizontal specificity in units in pigeon cerebellum and VN. again in units which were selected. by the experimental conditions. for their responses to horizontal vestibular stimulation. The details of these interactions await further experiments of a rather different design to those reported here, which will be best be carried out in a species, such as the pigeon. in which the yield of units per experiment is larger than that in the fish. so that it becomes practicable to build up a much larger sample of each type of response. Responses

in

cerebellum

compared

IO

those

in

brainstem

There were no obvious differences between the responses in the trout cerebellum and those in the VN. Some of the responsive cerebellar units were presumably Purkinje cells since they had complex spikes. Signals from EOM afferents reach the cerebellar vermis in the cat’.‘“.” and the flocculus in the rabbit.?’ However. the pathways by which the EOM afferent signals reach the cerebellum are not known and we have no information about the afferent pathway followed by EOM afferents in the fish. Thus it is impossible to say whether the VN units lie upstream or downstream of the cerebellum or whether these two structures simply share a common source of input from the orbital proprioceptors. Eflects

of orbital

signals

on units

in the oculomotor

nucleus

If the hypothesis that EOM afferent signals play some part in oculomotor control is correct. it follows that effects of these signals should be detectable at the output of the system. Thus there should be effects upon the activity of oculomotor neurons. However. if the action were to alter parameters within the system rather than to act from moment to moment

539

upon the control of individual eye-movements, the eventual effects on the output might not become manifest for some time and no changes might be detected in activity in the oculomotor nucleus in acute experiments. Although the general opinion in recent years has been that EOM afferents do not affect oculomotor control there have been a number of observations which strongly suggest that EOM afferents may indeed affect the oculomotor system.“*” However, as we recently pointed out,’ these results seem largely to have been ignored. In earlier, unpublished experiments on the toad we found that PEM appeared to affect the excitability of units in the III nucleus to the vestibular drive provided by electrical stimulation of the VIII (vestibular) nerve which, of course, provides a vestibular stimulus but not a physiological one! We wished to find out whether, in the decerebrate fish, there was evidence of an effect of the orbital afferent signal on the vestibular drive provided by the physiological stimulus of oscillation in the horizontal plane. As Table 4 shows, 16 units were found (when polymodal responses had been excluded) which responded to vestibular stimulation or to an orbital signal and, of these, eight carried both a vestibular and an orbital signal. Since only horizontal vestibular stimulation was used, one would expect to find vestibularly-driven units only when the electrode was in the internal rectus representation in the III nucleus: however. the fact that the alignment of the semicircular canals with the horizontal plane was almost certainly imperfect might have made neurons of the superior or inferior recti available also. Given these limitations. there is no means of knowing whether the units which responded apparently only to PEM might also have had vestibular responses to stimulation in another plane. The most significant findings are of four units in which vestibular drive was modulated by PEM. Figure 7 shows an example of one of these which is of further interest because it seems that horizontal PEM in the nasal quadrant was the most effective: thus this unit shows a preference for motion in only one part of the horizontal plane. The evidence that these units lay in the oculomotor nucleus depends upon histological reconstruction of their recording tracks from the glass fragments of the recording pipettes left in place at the end of the experiment. In Fig. 6 only one unit is shown as being within the III nucleus (AOZ. the unit whose response is shown in Fig. 5). However. the reconstruction shows only the limits-as far as they could be defined by the distribution of large cells-of the dorsolateral part of the III nucleus. Teieosts are peculiar in having, close to the midline. a ventral extension of the oculomotor nucleus which extends inferiorly almost to the base of the brain.> It is likely that units A03 and A04. which gave responses almost identical to those of A02. lay within this extension. Without further evidence one

540

J. A. ASHTONer al.

cannot be certain whether the units in the oculomotor nucleus from which we recorded were motoneurons. Once again, the details of the interactions await further experiments. Our purpose for the present is to present evidence that. in the trout, aRerent signals from the EOM reach the oculomotor nucleus as well as the cerebelhtm and VN, and are there able to influence the vestibular responses of cells. It has been suggested 3~1’that a possible action of the orbital signal might be to play a part in maintaining calibration of the performance of the oculomotor control system. and we have pointed out’ that this does not conflict with the possible actions of some type of corollary discharge. The observations reported here, showing that signals from the EOM afferents reach the oculomotor nucleus and there interact with the vestibular drive to units during natural vestibular stimulation, suggest that there may be an action of the EOM afferent signal upon the VOR. though it must be remembered that the passive eye-movements used were in the saccadic range of velocity rather than that of the slow phase of the VOR. The finding of effects of the afferent signals upon the oculomotor nucleus in acute experiments would appear to be consistent with an action on individual eye movements. However, the results would also be consistent with an action to adjust the parameters of the control system provided that the effect of this adjustment developed over a period of a few minutes. In that case there might be no effect of the afferent signal arising during a movement upon that particular eye-movement but an effect would appear upon later, similar, movements. Because we delivered a number of identical passive eye-movements during each collection of PSTHs our experiments would not distinguish between an action on the individual movement and a rapidly developing effect of parametric feedback. It is most interesting that. in the rabbit.” deafferenting the EOM on one side by cutting. or blocking conduction in, the ophthalmic branch of the

trigeminal nerve (which in other mammals is known to carry EOM aflerents to the trigeminal ganglionsee Ref. 37 for summary) caused the gain of the VOR to be reduced by 2040%. Further experiments on the details of the interactions between vestibular and EOM afferent signals at the level of the oculomotor nuclei will be necessary to test the hypothesis that the EOM signal is able to exert some regulatory influence on the behaviour of the oculomotor system during the VOR.

CONCLUSIONS The results of the experiments described here have confirmed that proprioceptors in the extraocular muscles of the trout, a teleost, send afferent signals to the cerebellum and to the VN. These signals are able to modify the responses of centrai c&Is to vestibular stimuli. This adds weight to the hypothesis that EOM afferent signals play a part in oculomotor control. In addition, the hypothesis is greatly strengthened bj the confirmation, which the results provide, of the prediction that effects should be found at the output of the oculomotor system, in the ocuiomotor nucleus. In this nucleus. also. EOM afferents were found to modify the responses of units to vestibular drive. Further experiments the details of the

are now necessary to examine effects of the EOM afferent

signal upon the processing of vestibular input during controlled perturbations of the behaviour of the vestibuto-ocntomotor system.

Acknowledgemenu-This

work was supported b) a grant from the Wellcome Trusr. C.M held Short Term FellowshIps from the European Science Foundation and the Fondation Simone et Cino de1 Duca. We are grateful to Mrs J. P. Donaldson for the preparation of all the hwological material. Mrs M. Prentis provided expert technical assistance. We are grateful IO Dr J. D. Porter for his comments on an earher vewon of the manuscript and for much helpful discussion

REFERENCES 1. Abrahams

V. C. and Rose P. K (1975) Project;ons of extraocular. neck muscle and retinal afferents to supenor colltculus in the cat. their connection to cells of origin of tectospinal tract. 1. Neuroph.vsiol. 38, l&18. 2. Ariens Kappers C. U.. Huber G. C. and Crosb) E. C. (1967) The Comparawe Anaromj of the h’mous SJxtern of Venebrares. Including Man. Warner. New York. 3. Ashton J. A.. Bodd} A.. Dean S R.. Milleret C. and Donaldson I. M. L. (1988) Afferent signals from cat extraocular muscles m the medial vestibular nucleus. the nucleus praeposims hypoglosG and adjacent brainstem structures Keuroscience 26, 131-145. 4. Ashton J. A., Boddy A. and Donaldson I. M. L. ff984) DIrectional selectivity in the responses of units in cat primar!

visual cortex to passive eye movement. Neuroscience 13, 653-66:. 5. Ashton J. A.. Bodd? A. and Donaldson I. M. L. (1984) input from proprioceptors m the extrinsic ocular muscles to the vestibular nucler m the Giant Toad Bufo mar&us. Expl Brarn ReS. 53, 409-419. 6. Ashton 1. A.. Donaldson I. M. L. and Milleret C. (1986) Afferent signals from the extraocular muscles reach the veslibular nuclei and rhe cerebellum in the trout. J. Ph.rsio!.. Land. 376, 56P. 7. Ashton J. A.. Donaldson I. M. L. and Millerel C. (1987) Input from extraocuiar muscle afTerems to oculomotor nuclei m the trout J. Physiof.. Land. 382, 74P 8. Batini C.. Buisseret P. and Kado R. T. (1974) Extraocular proprioceptive and trigeminal projections to the Purkmje cells of the cerebellar cortex. Archs irai. &of. 112, 1-I 7 9 Batini C. and Horcholle-Bossavit G. (1979) Extraocular muscle afferents and vrsual input interactions in the supenor colliculus of the cat. In Rcfle.4 Control q/ Posrure and Mouemenr (eds Grann R. and Pomperano 0.). pp. 335-344. Progress in &urn Research. Vol. 50 Eisevier North-Holland. Amsterdam

Effects of extraocuiar afferent signals in trout

541

10. Bortolami R., Lucchi M. L., Pettorossi V. E., Callegari E. and Manni E. (1987) Localization and somatotopy of sensory cells innervating the extraocular muscles of lamb, pi8 and cat. Histochemical and electrophysiological investigation. Archs ital. Biol. 125, l-15. 11. Carpenter R. H. S. (1988) Movemenrs ojrhe Eyes, 2nd edn. Pion, London. 12. Donaldson 1. M. L. and Knox P. C. (1988) Units in pigeon brainstem whose vestibular responses are modified by passive eye-movement. J. Physioi., Land. 398, 35P. 13. Donaldson 1. M. L. and Long A. C. (1980) Interactions between extraocular proprioceptive and visual signals in the superior coiliculus of the cat. J. Physiol.. Land. 298, 85-110. 14. Dijrrschcidt G. J. (1981) The statistical significance of the petistimulus time histogram (PSTH). Brain Res. 22&397-401. 15. Evinger C., Baker R. and McCrea R. A. (1979) Axon collaterals of cat medial rectus motoneurones. Brain Res. 174, 153-160. 16. Fuchs A. F. and Komhuber H. H. (1969) Extraocular muscle afferents to the cerebellum of the cat. J. Physiol., Land. 200, 713-722. 17. Gemandt 3. E. (1968) interactions between extraocular myotatic and ascending vestibular inguences. Exp/ Neural. 20, 12&134. 18. Graf W. and Baker R. (1985) The vestibufoocular reflex of the adult flatfish. I. Gculomotor organization. J. Neurophysioi. 54, 887-899.

19. Graf W. and McGurk J. F. (1985) Peripheral and central oculomotor organization in the goldfish Carassiw uurarus. J. romp. Neural. 239, 391-401. 20. Hermann H. T. (1971) Saccade related potentials in the tectum opticum and the cerebellum of Carassiuc. Brain Res. M, 293-304.

21. Hermann H. T. and Constantine M. M. (1971) Eye movements in the goldfish. Y&on Res. 11, 313-331. 22. Kimura M. and Maekawa K. (1981) Activity of flocculus Purkinje cells during passive eye movements, J. N~raphy~io~. 46, 1004-1017. 23. Kimura M., Takeda T. and Maekawa K. (1981) Functional role of extraocular muscle afferents in the control of eye movements in rabbits. J. Phvsiol. Sot. Japan 43, 317. 24. Luiten P. G. M. (1970) Pro&ceptive reflex connections ofhead musculature and the mesencephalic trigeminal nucleus in the carp. J. camp. Neurk lSj, 903-912. 25. Maier A.. De Santis M. and Eldred E. (1974) The occurrence of muscle snindles in extraocular muscles of various vertebrates. J. Morph. 143, 397408. . ’ 26. Montgomery J. C. and Macdonald 3. A. (19791 Stretch receptors in the eye muscles of a teleost fish. Experiemiu 36, 1176-1177. 27. Nieuwenhuys R. (1967) Comparative anatomy of the cerebellum. In The Cerebellum (eds Fox C. A. and Snider R. S.). pp. l-93 Progress in Brain Research. Vol. 25. Elsevier. Amsterdam. 28. Northmore D. P. M. (1984) Visual and saccadic activity in the goldfish torus longitudinalis. J. camp. Physiol. 155, 333-340. 29. Porter J. D.. Guthrie B. L. and Sparks D. L. (1983f Innervation of monkey cxtraocuiar muscles. Localization of sensory and motor neurons by retrograde transport of hoseradish peroxidase. J. romp. Neuroi. 218, 208-219. 30. Porter J. D and Spencer R. F. (1982) Localization and morphology of cat extraocular afferent neurones identified by retrograde transport of horseradish peroxidase. J. camp. Neural. 2@4, 56-64. 31. Precht W. (1978) Neuronni Operatlonx in rhe Vesribular System. Springer, Berlin.

32. van Sabussov G. H.. Maslow A. P. and Burnaschewa D. W. (1964) Vergleichend-morphologische und tinige histochemische Beobachtungen an besonderen Rezeptoren der Augenmuskeln bei Wirbeltiern. Anm. hr. 114,S 27-37. 33. Schwarz D. W. F. and Tomlinson R. D. (1977) Neuronal responses to eye muscle stretch in ccrebcllar lobule VI of the cat. Expl Brain Res. 27, 101-I 11. 34. Shairer J. 0, and Bennett M. V. L. (1986) Changes in gain of the vestibulo-ocular ref?ex indu=d by combined visual and vestibular stimulation in the goldfish. &am Res. 373, 164-176. 35. Sherrington C S. (1918) Observations on the sensual role of the proprioceptive nerve supply of the extrinsic ocular muscles. Brain 41, 332-343. 36. Spencer R. F.. Evinger C. and Baker R. (1982) Electron microscope observations of axon collateral synaptic endmgs of cat oculomotor motoneurones stained by intracellular injection of horseradish peroxidase. &urn Res. 234, 423-429. 37. Spencer R. F. and Porter 1. D. (1988) Structural organization of the extraocular muscles. In Reviews in Oculomoror Research. ~euroa~aiom~ qfrhe O~ulorn#~o~S.ssrem (ed. Buttner-Ennever J. A.), pp. 33-79. Elsevier Science, New York. 38. Tomhnson R. D. and Schwarz D. W. F. (1977) Responses of oculomotor neurones to eye-muscle stretch. Can. J. Ph.niol. Pharmac. 55, 568-573. (Accepred

13 February 1989)