Exp Brain Res (1993) 95:240-250
Experimental BrainResearch 9 Springer-Verlag1993
Evidence for corrective effects of afferent signals from the extraocular muscles on single units in the pigeon vestibulo-oculomotor system I.M.L. Donaldson, P.C. Knox Laboratory for Neuroscience, Department of Pharmacology,Universityof Edinburgh, 1, George Square, Edinburgh EH8 9JZ, UK Received: 28 October 1992 / Accepted: 8 March 1993
Abstract. The role of extraocular muscle (EOM) afferent feedback signals in the control of eye movement is still controversial. We recorded from 106 single units in the vestibular nuclei, oculomotor nuclei and reticular formation of 80 decerebrate, paralysed pigeons. EOM afferents were stimulated by passive eye movement (PEM) during vestibular stimulation by sinusoidal oscillation in the horizontal plane. We found that EOM afferent signals profoundly modified the vestibular responses of 91 (86%) of the single units recorded. As well as using PEM to simulate eye movements similar to saccades, we moved the eye in a manner which mimicked the slow phase of the vestibulo-ocular reflex (artificial VOR, AVOR). We have found evidence that, as well as providing signals closely related to the parameters of eye movement, PEM alters the vestibular responses of cells during AVOR in a manner which suggests that EOM afferent signals may play a corrective role in the moment-to-moment control of eye movement in the vestibulo-ocular reflex. Key words: Extraocular muscle afferents - Oculomotor control - Vestibulo-ocular reflex - Pigeon
Introduction The oculomotor system is said to be one of the best understood neuromuscular systems (Huber 1988). The vestibulo-ocular reflex (VOR), which produces compensatory eye movements in response to head rotations, has been extensively investigated, and is often believed to be both well understood and relatively simple (see Carpenter 1988, but also Collewijn 1989 for an alternative view of the VOR). However, in discussion both of the oculomotor system in general and of the VOR in particular, one familiar element common in other neuromuscular and reflex systems is almost always left out of the account, the role of afferent proprioceptive feedback signals Correspondence to: I. M. L. Donaldson
from the effector muscles: in the case of movements of the eye these are from the extraocular muscles (EOM). This omission is not an oversight; arguments have been advanced which claim that proprioceptive feedback may not be necessary for oculomotor control, at least from moment to moment (see Robinson 1981 for example), and the demonstration that, in the monkey, there is no monosynaptic stretch reflex in the extraocular muscles (Keller and Robinson 1971) has been taken to add weight to the general proposition that orbital proprioception plays no part in oculomotor control. However, it is well known that vertebrate eye muscles contain stretch receptors: only in a few species, including Man (Cooper and Daniel 1949), do these include true muscle spindles; in most species there are simple spirals and a specialised type of tendon organ (see Spencer and Porter 1988, for review). Furthermore, these receptors are known to supply information to the central nervous system (Cooper et al. 1953; Cooper and Fillenz 1955; Cooper et al. 1955; Ashton et al. 1984a). In an extensive series of experiments, we have stimulated EOM afferents by passive eye movement (PEM) and investigated the effects of their signals on units in central nuclei (Ashton et al. 1984a,b, 1988, 1989). In a number of species with quite different oculomotor behaviour, we have shown that EOM afferent signals can profoundly modify the vestibular responses of cells in the vestibular nuclei (Ashton et al. 1984b, 1988, 1989), reticular formation and oculomotor nuclei (Donaldson and Knox 1990, 1991). More recently, we have shown that the electromyogram of the extraocular muscles in the pigeon is also modified by EOM afferent signals during the VOR (Knox and Donaldson 1991a) and have found that these modifications of the output of the VOR are closely related to the parameters of the PEM, such as direction, speed and amplitude. This indicates that the EOM afferent signal conveys specific information to central nuclei concerned with the control of eye movement. In our most recent experiments (Knox and Donaldson 1991a), we have uncovered evidence that, if the eye is moved in a manner which mimics the slow phase of the
241 V O R , the afferent signals thus p r o d u c e d cause changes in the o u t p u t of the V O R which a p p e a r to be corrective. T h u s we f o u n d that, d u r i n g the V O R , w h e n one eye was m o v e d too fast - that is faster t h a n w o u l d be required to c o m p e n s a t e for head m o v e m e n t - the o u t p u t of the reflex (as evidenced by the E M G of the E O M of the other eye) was reduced. O n the o t h e r h a n d , w h e n the eye was m o v e d too slowly the o u t p u t increased. We were a n x i o u s to investigate this effect at the level of single u n i t s in the b r a i n stem a n d to c o m p a r e the effects observed with slow P E M , similar to that d u r i n g the slow phase of the VOR, with those observed with the type of eye m o v e m e n t which we have previously e m p l o y e d a n d which is m o r e similar to saccadic m o v e m e n t s . The results q u o t e d a b o v e a n d those which we n o w report suggest that it is time to re-evaluate the role of E O M p r o p r i o c e p t i o n , b o t h in the o p e r a t i o n of the V O R a n d in o c u l o m o t o r c o n t r o l in general. We have given a p r e l i m i n a r y a c c o u n t of some of the present results ( K n o x a n d D o n a l d s o n 1991b).
normal
~ ~ernent~mposed \\
Stimuli The left eye was moved passively by an electromagnetic servo-controlled device which acted upon a stalk carried by an opaque contact lens held in position on the cornea by suction (Ashton et al. 1984a; Donaldson and Knox 1990). Two different types of eye movement were imposed upon the left eye. One, which we have labelled pseudo-saccadic (PS) consisted of a rapid movement of the eye (15~ at 115~ from the central resting position to a new eccentric position where it was held for 200 ms before being moved back to the central position. In the second, we moved the eye in a manner which mimicked the slow phase of the normal VOR ('artificial VOR', AVOR). When the head rotates in space the eyes counter-rotate to compensate for the head movement (Fig. 1, "normal"). Given the standard vestibular stimulus used in our experiments (peak head velocity 22~ and assuming perfect compensation, the eyes would be expected to rotate at 22~ in the opposite direction to the head. In the experiments reported here the preparation was paralysed, so in the absence of PEM there were no eye movements. However, during "artificial VOR" (AVOR), the left eye was moved sinusoidally phase-locked to the table and thus to the vestibular stimulus. When the eye was moved passively at the same speed as
//~
.......
headvelocity
......
desired' eye velocity
Materials and methods All of the units described here were recorded in decerebrate, paralysed and artificially ventilated pigeons. The preparation and techniques of stimulating and recording were identical to those which we have described previously (Donaldson and Knox 1990). Briefly, glass-coated tungsten microelectrodes were directed vertically downward through the anterior folia of the cerebellum. Recordings were usually made from the left brain (ipsilateral to the eye being moved) but, on some occasions, the electrode was inserted into the right brainstem. Extracellular recordings were made from single units responding to vestibular stimulation produced by sinusoidal oscillation of the bird in the horizontal plane (usually ___8~ at 0.4 Hz).The recording sites were marked by electrolytic lesions (20 gA for 10 s), and at the end of the recording session the brain was fixed in situ by transcardiac perfusion with buffered formalin (Eden and Correia 1981). After further fixation, 50 gm serial, frozen, parasagittal sections were cut and stained for Nissl substance. Sections were projected and drawn and electrode tracks reconstructed. Structures were identified with the help of the atlas of the pigeon brain by Karten and Hodos (1967).
artificial VOR
//
actual eye velocity
~
velocityerror illustrations show only one half-cycle of sinusoid
Fig. 1. Illustration of the principle of the artificial VOR (AVOR). The normal panel illustrates compensatory counter-rotation of the left eye during the normal VOR. The artificial VOR panel illustrates how eye movement imposed on the left eye in the paralysed pigeon can be adjusted to provide, from above down, excessively high eye velocity, excessively low eye velocity and eye velocity too low and also in the wrong (anti-compensatory) direction. The velocity error in each case is the hatched area
the table but in the opposite direction (ie 180~ out of phase with the table), this stimulus was called compensatory AVOR. Using this approach, various errors of phase, velocity (as illustrated in Fig. 1) or direction relative to head movement could also be imposed on the eye. As in our previous experiments (Donaldson and Knox 1990; Donaldson and Knox 1991), responses were collected into sets of peristimulus time histograms (PSTHs) which were interleaved in time.
Analysis of data Sinusoids of the frequency used for vestibular stimulation were fitted to the records of table position and to the PSTHs of the responses of the units. Since the required frequency was known, an analytical method (modified from that of Arzi and Magnin 1989) was used to solve the differential equations which describe a sinusoid with a minimum sum-of-squares deviation from the data. Some units fired during only part of the vestibular cycle; for these, only the bins of the PSTH corresponding to the response were used to fit the sinusold to the PSTH (see Melvill Jones and Milsum 1970 and Hulliger et al. 1977 for discussion). From the best-fitting sine for table position the corresponding sinusoid for table velocity was derived by differentiation. Because the head moved with the table, table velocity was equivalent to head velocity. The phase of the vestibular response was defined as the difference between the phases of the
242
24 sweeps, bin width 10 msec. sweep duration 2.5 sec
no eye movement peak eye velocity 115 deg/sec
b
'0.5
'1.0
'1.5
'2.0
q2.5
b
'0.5
'1.0
'1.5
'2.0
'2.5
Otg3
d
Fig. 2. a, c Peristimulus time histograms (PSTHs) of the responses of a unit in the reticular formation to vestibular stimulation alone by horizontal, sinusoidal oscillation at _+8~ 0.4 Hz. b The response to a combination of the vestibular stimulation and pseudosaccadic (PS) passive eye movement (15~ at 115~ d Vestibular stimulation was accompanied by sinusoidal movement of the left eye at the same speed as the table to produce a compensatory AVOR. Dotted sinusoids indicate table position with downward deflection corresponding to movement towards the left side (ipsilateral to the recording site); solid trapezoid~ (b) and sinusoid (d) indicate position of the left eye which is being moved passively. Note the marked reduction in the unit's vestibular response when either PS (b), or compensatory AVOR (d), eye movement is combined with the vestibular stimulus, e A reconstruction of a parasagittal section of the brainstem (anterior to left, superior upwards) containing the recording track of the unit whose responses are illustrated above and whose recording site was marked by an electrolytic lesion at C1
movement 1 eyevlociy22degJscc b
'1.0
'0.5
'1.5
, o
'2.0
'2.5
b
'0.5
see.
e
'1.0
'1.5 sec.
~
m e d i a l vestibular nucleus
sinusoids fitted to the response and to table velocity (with phase lead corresponding to occurrence earlier in time and assigned a negative value) and the gain was given by (amplitude of sine fitted to response)/(amplitude of sine fitted to table velocity) in units of (impulses/s) per (~ The compensatory AVOR requires the eye to be moved 180~ out of phase with the table. In the illustrations of the effects of phase errors in eye movement on the responses, these errors are expressed relative to the phase of the compensatory AVOR so it is most convenient to assign the phase value of zero to the compensatory AVOR. Tests of statistical significance of the effects of PEM on the vestibular responses were made as described previously (Donaldson and Knox 1991) using a method derived from that of DSrrscheidt and based on the binomial theorem (D6rrscheidt 1981; Ashton et al. 1984a; Donaldson and Knox 1991).
Results
Single units were r e c o r d e d f r o m 80 p i g e o n s ; 106 units which r e s p o n d e d to h o r i z o n t a l v e s t i b u l a r s t i m u l a t i o n
'2.0
'2.5
were t h e n tested with P E M . R e c o r d s from units which r e s p o n d e d to P E M b u t also r e s p o n d e d to m e c h a n o s e n s o ry s t i m u l a t i o n of structures a r o u n d the eye were d i s c a r d ed, a n d will n o t be discussed further here. T h e v e s t i b u l a r r e s p o n s e s of 91 of the 106 (86%) were clearly m o d i f i e d b y p s e u d o - s a c c a d i c (PS) P E M . These units were t h e n tested with the s e c o n d t y p e of P E M which m i m i c k e d the slow p h a s e of the V O R (AVOR); F i g u r e 2 shows the effect of these two types of P E M o n a unit l o c a t e d in the r e t i c u l a r f o r m a t i o n d e e p to the m e d i a l v e s t i b u l a r nucleus (Fig. 2e) a n d w h i c h r e s p o n d e d to i p s i l a t e r a l (leftward) r o t a t i o n . PS P E M p r o d u c e d a large i n h i b i t o r y m o d i f i c a t i o n of the v e s t i b u l a r response. A l t h o u g h P E M t o w a r d s the beak, the tail a n d d o w n w a r d all h a d a m a r k e d effect on the v e s t i b u l a r r e s p o n s e (statistically significant r e d u c t i o n s c o m p a r e d with the c o n t r o l response, P < 0.005), P E M d i r e c t e d t o w a r d s the tail (Fig. 2b) h a d the largest effect w h i c h differed significantly (P < 0.005) f r o m b o t h the c o n t r o l r e s p o n s e (Fig. 2a, " n o eye m o v e m e n t " ) , a n d the o t h e r two test P S T H s (not shown). A set of P S T H s was
243
a
24 sweeps, bin width 10 msec. sweep duration 2.5 sec no eye-movement
eo
6 deg/sec
~~ i A ~
0 c/)
16 deg/sec
22 deg/sec
28 deg/sec
33 deg/sec
, ~,~l....
,.,
39 deg]sec
41 deg/sec
l
'0.5
,,~,,'1.0~
~
~, ~ '1.5
'2.0
1'2.5
'0.5
:1.0 '1.5 SEX3.
see.
'2.0
'2.5
Fig. 3. a Eight PSTHs of the responses of a unit in the medial vestibular nucleus to sinusoidal horizontal oscillation (_+8~ at 0.4 Hz) alone (no eye movement) and during AVOR with the peak eye velocities marked against the panels. The response at 22~ represents that during compensatory AVOR when the head and eye speed are equal but in opposite directions. Note the changes in the size of the vestibular responses as peak eye velocity changes, from too small through compensatory to too large. b A plot of the data from a in which the ratio of the gain of the vestibular response at each eye velocity to that at 22~ (compensatory AVOR) is plotted against the peak eye velocity. The line is the linear regression (P < 0.01). When the eye was moved more slowly than required to compensate for the head (table) movement the gain of the response increased; when the eye moved too fast the gain decreased
b 140 |
120 ~5~ 100
,N
80 60
..= r o 9 '~
40
20
r = 0.84; 0.001
0 0
T 10
i 20
r 30
40
50
maximum velocity of eye-movement (deg/sec) then collected in which the left eye was moved in a manner which mimicked the normal VOR. The peak table velocity during the sinusoid was 22~ Therefore, the eye was moved at this speed (22~ but in the opposite direction to produce a compensatory AVOR. Once again there was a control PSTH with the vestibular stimulus alone and the eye held at a central position (Fig. 2c). Compensatory A V O R caused a clear inhibitory modification of the vestibular response (Fig. 2d); the gain of the vestibular response was reduced from 45 (impulses/s) per (~ to 21 (impulses/s) per (~ that is to 47% of the control value. In 70 units a comparison was made between the vestibular response when the eye was not moved and the
vestibular response during compensatory AVOR. The vestibular responses of 27 of the 70 units (38%) were reduced by more than 10% during A V O R ; those of 10 units (14%) were increased by more than 10%. Seventeen of the 70 units tested in this way were located in the abducens nucleus and, of these, the vestibular responses of 9 (53%) were reduced by more than 10%, while only one unit showed an increase greater than 10%.
Effect of alterations in velocity of PEM during A VOR In 29 units we investigated the effect of varying the peak P E M velocity while maintaining the peak table velocity
244 1.2
o
,El
s0001
9
o
Fig. 4. Linear regression of the ratio of the gain of vestibular responses during AVOR (expressed as ratio to the gain with no eye movement, i.e. eye velocity 0~ on peak eye velocity for 216 pairs of observations in 27 units in various locations in the brainstem. Dotted hyperbolas are 95% confidence limits for the line. The correlation coefficient is highly significant (P < 0.001). The gain ratio falls, on average, by about 1% for each l~ increase in peak eye velocity during the AVOR
"~ 0.8
.-= 0.6 o 9 0.4
I
I
l0
20
I
30
40
50
peak velocity of eye-movement (degrees/sec) at 22~ Peak P E M velocity ranged from 6~ through 22~ (compensatory AVOR) to 42~ An example of the effects observed is shown in Fig. 3. The PSTHs show that the vestibular response of this unit was modified by the PEM. When a compensatory A V O R (peak PEM velocity 22~ was applied (Fig. 3a, "22 deg/sec'), there was a clear reduction in the vestibular response compared to the control response ("no eye movement"). As the velocity of PEM increased, the gain of the vestibular response was progressively reduced. If the vestibular response with peak PEM velocity at 22~ is taken to indicate what would be expected during a natural VOR with a gain of - 1.0, then it is clear from Fig. 3b that, when the eye was moved more slowly than would be required to compensate for head movement, there was an increase in the gain of the vestibular response, and, when the eye was moved more quickly than was appropriate, the vestibular response was decreased. The response to vestibular stimulation alone (without eye movement) is treated as occuring at a velocity of 0~ This unit was recorded from the anterior ventral border of the medial vestibular nucleus. In the 29 units tested in this way the gain of the vestibular response in 27 was reduced progressively by increasing velocity of PEM during the AVOR. Linear regressions were constructed of: (ratio of response at various velocities of PEM to control response without PEM) to (velocity of PEM); the correlation coefficients were greater than r = 0 . 7 (r 2 > 0.5) for all the units. Analysis of covariance of the regression data for 15 units with the most widely spread parameters suggested that the regression lines form a group homogeneous in slope with a mean slope of -0.01. Combining the data from the 27 units produced the regression line plotted in Fig. 4, which has a slope of - 0 . 0 1 . The value of the correlation coefficient (r) of 0.67 for 214 degrees of freedom gives P < 0.001 and leaves no doubt about the statistical correlation between gain and eye velocity. Thus, for each degreeper-second increase in velocity of eye movement during the AVOR, there was a reduction, on average, of about 1% in the gain of the vestibular response of these units. Histological reconstruction showed that the 27 units were located in several centres involved in the control of
the vestibulo-ocular reflex, including the abducens nucleus and the medial vestibular nucleus. There was little evidence that alterations in the velocity of PEM during the AVOR produced alterations in the phase of the vestibular responses in the units tested. From Fig. 3 it is clear that the unit fired throughout the same part of the stimulus cycle whatever the velocity of PEM. Careful measurement did reveal small changes in the phase of the responses, of the order of a few degrees, but these shifts showed no systematic relation to the stimulus velocity.
Effect of changing the table (head) peak velocity during the AVOR We examined the effect of altering peak table velocity in a number of units. Eight units were tested over the range of different peak eye velocities at three or four different table velocities. Figure 5 shows a typical result. Reducing peak table velocity caused an increase in the gain of the vestibular response of most units. The unit in Fig. 5a responded to ipsilateral rotations of the head and was located in the medial longitudinal fasiculus. The same range of eye velocities was used at each table velocity. As the peak PEM velocity increased there was the reduction in the gain of the vestibular response that we described above. However, as is particularly clear in Fig. 5a, the effect of PEM was greater at the lower table velocities. Thus, with a peak table velocity of 25~ the maximum reduction in gain was to 50% of the control value (which for the purpose of these tests is the no-PEM condition, peak PEM velocity of 0~ whereas with a peak table velocity of 6~ the gain of the vestibular response was reduced to 40% of the control value. The slope of the fitted regression line was steeper at the lower peak table velocity, suggesting that a given increase in the PEM velocity caused a greater decrease in the gain of the vestibular response at low head velocities. The second example, illustrated in Fig. 5b, is from a unit located in the abducens (VI) nucleus. Once again it is apparent that decreasing the peak table velocity caused an increase in the gain of the vestibular response of this unit. Although
245 the other five units and in one case the gain of the vestibular response was highest with a peak table velocity of 25~
a 50
40
~D
g,
30 .........O..
~
Effect of altering the phase relation between eye movement and table movement
--- ...~
,,~20
I
I
I
I
10
20
30
40
50
peak velocity of eye-movement
symbols peak velocity of table (head) movement 9 --x- . . . .
6
9 ....
~k--
12
19
25
degree/sec
b 30
"D 25q ......... 9
............... 9 ......... 9
ezs3
............ 0 ....... 0 ...... * o
-~ 20
~
151
"4
....
-A--_
9
....
9 .....
~ - - - A
....
10 !
5
0
I
r
I
I
10
20
30
40
50
peak velocity of eye-movement Fig. 5a,b. Effects of interaction between the peak table (head) velocity and the peak eye velocity imposed during the AVOR on the gain of the vestibular response of two units. A standard set of peak eye velocities was combined with four (a) or three (b) table peak velocities (all movements were sinusoidal). Gains are plotted against peak eye velocity and the lines are the linear regressions at the various table velocities. Note the larger gains of the vestibular responses without eye movement at the lower table velocities (points at 0~ on ordinate) and the greater slope of the regressions on eye velocity for the lower table velocities
eye movement apparently had little effect at a table velocity of 25~ at the lower table velocities the same peak eye velocities caused a clear reduction in the gain of the vestibular response. Of the eight units tested in this way, five showed this same pattern of effect. Two of these were located in the abducens nucleus and one in the medial longitudinal fasciculus; one was located in the medial vestibular nucleus and one in the central nucleus of the dorsal medulla. In the remaining three units there were the same reductions in the gain of the vestibular response, but the effect of the alteration of the peak table velocity was not as clear as in
In a number of units we altered the phase of P E M relative to the vestibular stimulus while keeping the peak velocities of the eye and the table equal. Phase lead indicates occurrence earlier in time. In some units phase errors resulted in systematic changes in the gain of the vestibular response, though the results were less clear than those produced by altering the velocity of PEM. The unit whose response is illustrated in Fig. 6 was located in the oculomotor (III) nucleus. From both the PSTHs (Fig. 6a) and the plot of the gain of the vestibular response (Fig. 6b) it can be seen that, when the eye was moved with a phase lag relative to the table, the gain of the vestibular response of the unit was reduced (Fig. 6a, phase +60, +40, +20). Also, when the eye moved in exactly the opposite direction to the table (phase 0, compensatory AVOR) but with the same speed, the gain of the vestibular response was reduced from control levels. When the eye was moved with a phase lead, whilst there was still a reduction from the control, this reduction was not as great (Fig. 6a - 2 0 , - 4 0 , - 6 0 ) as when there was phase lag. A further five units responded in a similar manner; two were located in the abducens nucleus. Four units were held for long enough to test them over a wider range of phase errors. Figure 7 shows the effects on two units of moving the eye with progressively larger phase errors. Note that each plot is a composite of two sets of PSTHs; this accounts for some of the jitter. Note also that in these examples the peak velocity of the PEM was 31~ greater than that for the table. In both examples phase lag of the PEM of up to 100 ~ relative to the vestibular stimulus produced decreases in the gain of the vestibular response. As the phase lag decreased and became a phase lead the gain of the vestibular response increased. Phase leads, while still reducing the gain of the vestibular responses, did not reduce it as much as phase lag. Once again, as for the effect of velocity errors, the main effect was on the gain of the vestibular responses in these units. As can be seen from Fig. 6a, the phase of the responses remained unaltered.
Locations from which units were recorded Units were recorded in various structures throughout the brainstem as indicated in Table 1. We recorded from both of the motor nuclei involved in horizontal eye movement (III and VI) and the responses of all the units there were modified by PEM. This confirms our previous finding that all of the units located in the abducens (VI) nucleus are affected by PEM (Donaldson and Knox 1991). In
246
a
24 sweeps, bin width 10 msec. sweep duration 2.5 sec no eye-movement
,n
l
i
i
l
"~
9 v,,~
i
phase +60 deg
l li ~ni[ Ill I L ! !1~I Ill i ~lnI~t~LIl~
Inill~llllli][llII i1~1inn i
phase +40 deg
i
i
i
i
~
u
I
i
i
i
phase +20 deg
i
II~fl ~1~1 i
FI
i
i
i
i
i
i
~
i
t'q
9 (.,i r~
Ii1 II nni 411 nun nn
In
.
QJill
phase -40 deg
II IIII
0
'1.0
'0.5
i
'1.5
'2.0
'23
phase -60 deg
1.++., b
'03
see.
'1.0
'1.5
'2.0
'2.5
see.
b J
gain when no eye-movement Fig. 6. a A set of eight PSTHs similar to those of Fig. 3 but in which the peak eye velocity was always 22~ thus producing an AVOR compensatory in speed for the head movement, and in which the phase of the eye movement was varied from 60~ phase lead (-60 ~ to 60~ phase lag (+ 60~ relative to the AVOR compensatory in both speed and phase, taken as phase 0~ Phase lags resulted in greater reductions of gain than did phase leads but all phase errors caused some reduction in the gain ratio, which is expressed in this figure relative to the gain during compensatory AVOR at phase 0o
150 c~
t00 2 ~n
50
@
o
lead I
I
t
-60
-40
-20
0
-80
lag 0
I
I
I
20
40
6O
80
phase of eye-movement during AVOR (deg)
Table 1. Histological location of 66 units Vestibular nuclei
Medial vestibular nucleus Other vestibular nuclei
12 3
Motor nuclei
Abducens (VI) Oculomotor (III)
23 5
Reticular formation
Central nucleus of dorsal medulla Nucleus reticularis parvocellularis Other reticular nuclei
8 1 10
Other locations
Medial longitudinal fasciculus Cerebellum
3 1
o t h e r l o c a t i o n s there were s o m e units which r e s p o n d e d to the v e s t i b u l a r stimulus b u t w h o s e responses were n o t m o d i f i e d by P E M of either type. R e s p o n s i v e units were also l o c a t e d in the m e d i a l v e s t i b u l a r nucleus a n d the u n d e r l y i n g r e t i c u l a r f o r m a tion. T h e r e is a n a r e a of r e s p o n s i v e n e s s to c o n t r a l a t e r a l r o t a t i o n which lies d e e p to the a n t e r i o r m e d i a l v e s t i b u l a r nucleus, the c e n t r a l nucleus of the d o r s a l m e d u l l a , w h e r e once a g a i n (see D o n a l d s o n a n d K n o x 1990) a n u m b e r of units w h o s e r e s p o n s e s were m o d i f i e d b y b o t h types of P E M were located.
247 ~ " 100
[]
9
[] []
80 O
[] 60
[]
0
[]
[]
xA
[]
N 40 []
9N 20
lag
lead
O O
I
[
0 -200
0
-100
200
100
phase of eye-movement during AVOR (deg)
' ~ " 100
g
[] [] []
[]
[]
O
0
x~ "~
[]
6o
[]
20
lag
lead
I
I -200
Fig. 7. Plots for two different units of the ratio of gain during AVOR (eye movement always at the same speed, 31~ but at various phase angles relative to head movement) against phase of eye movement. Note that, in this figure, the ordinate is the ratio (gain)/(gain when there is no eye movement), because the data are derived from two sets of interleaved PSTHs whose common stimulus condition is that of no eye movement
[] []
40
0 Q i.
[]
[]
9
-100
0
100
200
phase of eye-movement during AVOR (deg)
Discussion
In these experiments we sought to investigate the interaction between afferent signals from the extraocular muscles (EOM) and the vestibular responses of cells located in central structures which are involved with the control of eye movement, such as the vestibular nuclei and oculomotor nuclei. We have discussed elsewhere the advantages of using passive eye movement (PEM) to stimulate E O M afferents by simulating natural eye movement (Ashton et al. 1984a; Ashton et al. 1988). Our present experiments do not provide information about the details of the peripheral signal from the muscle receptors. It must also be remembered that the eye muscles were paralysed and that the sensitivity of the receptors in unparalysed muscles to a given stretch may be different; it may well be greater, as the muscles will be stiffer. In our previous experiments we used passive eye movements similar to
saccadic eye movements (PS PEM), within the normal saccadic range for the pigeon. The stimulus consisted of a rapid movement of the eye from the centre of the orbit to an eccentric position where it was held before a second, rapid movement brought it back to the central position again. Using this type of stimulus we have shown that the vestibular responses of central units are clearly modified in a manner related to the speed, direction and amplitude of PEM. However, it might be argued that this type of PEM is inappropriate under conditions of sinusoidal vestibular stimulation; the VOR produces a compensatory sinusoidal eye movement at a similar speed, though in the opposite direction, to the head. Thus the expected eye velocity in our experiments would be much lower (22~ than the standard velocity used in the PS tests (usually greater than 100~ Furthermore, we have shown that the PS eye movement which consists of a trapezoid (ramp-hold-ramp) is a complex stimulus. Different units responded to different combinations of some,
248 or all, of the different elements of the trapezoid (Donaldson and Knox 1990). Therefore we developed the artificial vestibulo-ocular reflex (AVOR) procedure (Knox and Donaldson 1991a; Knox and Donaldson 1991b) in which the PEM is both simpler and is a more nearly physiological stimulus with which to investigate the effect of afferent signals from the EOM during the VOR. Our previous results (Knox and Donaldson 1991b) and those reported here show that the presence or absence of the normal compensatory eye movement is important in setting the gain of the vestibular responses of single units in those nuclei involved in the generation of the VOR, the motor nuclei [units were located in both the abducens (VI) and oculomotor (III) nuclei], the vestibular nuclei and the reticular formation. This is shown by the comparison of control histograms (vestibular stimulation alone, no PEM) with those in which the vestibular stimulus was combined with AVOR PEM with the eye being moved at the same speed as the table, but in the opposite direction. This mimics the normal VOR and thus, we assume, produces an afferent signal similar, though from only one eye, to that which would normally be generated by the VOR. Under such conditions there were clear alterations in the gain of units. However, there was a range of sensitivity to the EOM afferent signal. Thus some units were little affected, the gain of others was reduced by up to 60% compared to the gain with no eye movement, and, in a smaller number, the gain increased. A comparable experiment was carried out by Taylor (1965) in the anaesthetised cat; curarisation of the eye muscles, and therefore removal of the EOM afferent signal, caused an increase in the mean discharge rate of abducens motoneurons and a reduction of approximately 20% in the gain of vestibular responses. Given the difference in preparation and technique, it is of interest that here too the effect was on the gain and not the phase of the vestibular responses. The results obtained with velocity errors during AVOR were clear and consistent. When there was no PEM the gain of the vestibular responses of the units from which we recorded was at its maximum; AVOR PEM reduced the gain. If the gain when the eye was moved at the correct speed for compensation is regarded as the normal gain, we found that, when the eye was moved more slowly than required for compensation, the gain of vestibular responses was increased; moving the eye faster than was required for compensation caused the gain of the vestibular response of units to be progressively reduced. Furthermore, simple comparison of the response produced by the compensatory AVOR with that to vestibular stimulation without PEM provided an incomplete picture of the effect of the afferent signal. Thus in some units, in which the gain of the vestibular response was affected by only a few percent during compensatory AVOR, tests at a number of different peak velocities of PEM showed a clear relationship between peak eye velocity and gain. It is also of interest that, for units in such a wide range of brainstem locations, the gain was always reduced as PEM velocity increased. While comparison with only the compensatory AVOR suggested a wide range of sensitivities to the afferent signal, most, if not all,
of the units tested over a range of velocities were affected similarly with a given increase in peak PEM velocity reducing the gain of the vestibular response by proportionately the same amount. Presumably we have recorded units which are located at a number of different points in the VOR control circuitry, yet all showed this same pattern of reduction of gain at higher velocities of passive eye movement. That such large reductions in gain can be obtained at relatively low peak velocities compared with PS PEM suggests that the details of the time course and waveform of PEM, thus the type of PEM, as well as its size, speed and direction, is important. The same effects have been observed at the output of the system, as evidenced by the modification of the EMG of the EOM themselves (Knox and Donaldson 1991a). Reductions of up to 40% were observed in the EMG of the lateral rectus under identical stimulus conditions to those in the present experiments; in the central units examined here much larger reductions were observed - of up to 80% in some cases. While the gain of the vestibular responses is markedly affected by the addition of EOM afferent signals, their phase is little affected. Of course, it is possible to alter both phase and gain of the responses to vestibular stimulation alone by changing, for example, the frequency of the vestibular stimulus (see the Bode plot for abducens units in Donaldson and Knox 1991 and the results of Anastasio and Correia 1988). Again, the small and irregular changes in the phase of the vestibular response produced by the EOM afferent signal (see below) are consistent with earlier reports that removal of the EOM afferent signal in the cat by curarisation did not alter either the phase of single unit responses from the abducens nucleus (in the anaesthetised cat, Taylor 1965) or the EMG recorded from lateral rectus (in the decerebrate cat, Carpenter 1972). This has been taken to indicate that a feedback signal from the EOM afferents is not used to provide the eye position signal necessary for the generation of the slow phase of the VOR (see Carpenter, 1988 for further discussion). However, our present results show that the principal change induced by the EOM afferent signal is in the gain rather than the phase of the responses of units in the vestibular nuclear complex and the abducens nucleus during the artificial VOR (and thus, by extension, presumably during the "natural" VOR). The tests of the effect of AVOR PEM at different table velocities also suggest that afferent signals play a corrective role in the normal VOR. If the table velocity is reduced and the same PEM velocities are used, we produce a larger error between the two signals and thus expect to see greater reductions in the gain of vestibular responses: this is indeed what happens. Because the alteration in the vestibular stimulus produces an increase in gain, the effect of PEM is proportionately very similar at the different table velocities tested. But the effectiveness of the afferent signal is increased at the lower table velocities and so, for a given increase in the PEM velocity over the table velocity, there is a greater reduction in the gain. Phase errors between the passive eye movement and the vestibular stimulus also produced alterations in the gain of the vestibular responses (Fig. 6). Although most
249 units which were sensitive to velocity errors appeared to be less sensitive to phase errors; this could be due to sampling effects because of the difficulty of holding cells for the long periods required to perform all the different tests. Velocity effects were usually tested before those of phase and it was not always possible to continue and carry out a complete set of phase tests. Nevertheless, we have presented evidence above that the gain of some units is modified both by the velocity and the phase of PEM relative to the vestibular stimulus. While the unit whose response is illustrated in Fig. 6 shows the effect of phase errors alone, the composite plots in Fig. 7 illustrate the effect on two units of a combination of phase errors and a velocity error. In the latter examples the left eye was moved faster than required for compensation (31 ~ vs 22~ and the addition of phase errors resulted in a larger reduction of the response than that produced by the velocity error alone. This suggests that these two types of error may interact in an additive manner. Thus, when there is a reduction in gain due to a greater-thanrequired PEM velocity, the addition of a phase lag causes a further reduction in gain, while the addition of a phase lead produces an increase in gain. The effect of phase errors is more difficult to understand than the effect of errors in velocity in terms of correction for inappropriate eye movement. The firing of units was fixed relative to the vestibular stimulus, usually with a small phase lead relative to table velocity. Careful examination shows that, during phase lags of PEM on the vestibular stimulus of between 20 ~ and 60~ there is a complicated series of effects. Thus, when the unit whose response is illustrated in Fig. 6 begins to fire, the eye is moving in the correct direction for compensation but at too high a speed. As the unit reaches peak firing, the eye is moving in the correct direction but at too slow a speed. Finally, as the firing of the unit tails off, the eye is again moving in the correct direction but at too high a speed for compensation. The details of the difference between the eye movement required for compensation and the actual eye movement (the PEM) therefore depend on a combination of the phase position of the unit's activity in relation to the vestibular stimulus and the phase of the PEM. Given this complexity, it may be necessary to collect data on a wider range of units in a number of different brainstem centres before drawing firm conclusions. It is not clear why the composite plots in Fig. 7 do not show a consistent pattern of effect over a larger range of phase errors. However, perhaps this is no more than an expression of the fact that, under natural conditions, it seems unlikely that large phase errors would arise. Certainly the results show that the system is more sensitive to the effect of errors of eye velocity than to those of phase and its response to errors is expressed in changes of gain rather than in changes of phase. The question arises as to whether there is a "difference in principle" between the two types of PEM used to stimulate EOM afferents in these experiments. If EOM afferent signals do provide proprioceptive feedback used in the moment-to-moment monitoring of eye movement, then there is indeed such a difference between PS and AVOR PEM. It is generally thought that fast phases are
not initiated at a fixed point in the cycle during sinusoidal head oscillation (Carpenter 1988). If the VOR control system interprets PS PEM as indicating that such a fast eye movement has occurred, there will be no central motor command produced by the pigeon which can appropriately be compared with the afferent feedback resulting from the PS PEM imposed by the experimenter. Therefore, from the point of view of the central control system, the entire afferent signal produced by PS PEM is likely to be treated as an error signal resulting, in effect, from an unintended eye movement. In complete contrast, during the AVOR there will be an on-going sinusoidal motor command in response to the vestibular stimulus, and this command will seek to drive the eye at the same (or nearly the same) speed as the table, but in the opposite direction, to produce the slow phase of the normal VOR. In the paralysed preparation, during the AVOR, we either mimic this compensatory eye movement or produce slightly different eye movements, and hence introduce an error. However, the afferent signal in toto will not be the error; in this case presumably only the difference between the afferent signal and that "expected" from the motor command will be treated as error. This, then, is clearly a major difference, a difference-in-principle, between PS and AVOR PEM. This fundamental difference between the significance for the control system of the two types of eye movement may also explain why it has been difficult to fit the effects PS PEM into a neat corrective scheme, whereas the effects of the AVOR fit such a scheme rather well. While rapid pseudo-saccadic stimuli are useful in indicating that EOM afferent signals convey amplitude, velocity and directional information to central structures, they are certainly inappropriate for studying the functional effects of EOM afferent signals during the VOR. What is of considerable interest, and awaits further study, is what happens when AVOR and PS PEM are combined to mimic the combination of fast and slow eye movements which happens in the VOR in intact animals. The results reported here add to a growing weight of evidence indicating that EOM afferent signals do play a role in the VOR. Of particular interest is the finding that, in the rabbit, removal of the EOM afferent signal by cutting the ophthalmic branch of the trigeminal nerve disrupts the slow phase of the VOR (Kashii et al. 1989). We have now found that, in the pigeon, EOM afferent signals can modify the vestibulo-collic reflex which stabilises the head in space (Hayman et al. 1993): this suggests that EOM afferent signals may play a wider role in the control of gaze stabilization and eye-head coordination. Furthermore, there is recent evidence from psychophysical studies in Man which indicates that EOM afferent signals are involved in the normal human visual system in the accurate location of objects in space (Gauthier et al. 1990). These results follow earlier findings showing that strabismic patients show characteristic deficits in the location of objects after surgery which, it has been argued, disrupts feedback from the EOM (Steinbach 1987). Thus it seems that it is time to reconsider the role of EOM afferent feedback, not only in the control of the VOR, but also in oculomotor control in general. This has implications for our understanding of the oculomotor
250 system, for the models c o n s t r u c t e d to illustrate a n d illum i n a t e its o p e r a t i o n , a n d also, m o r e i m p o r t a n t l y , for o u r u n d e r s t a n d i n g of h u m a n o c u l o m o t o r a n d v i s u o m o t o r control. D a m a g e to the E O M or d i s r u p t i o n of the afferent signals m a y need to be considered as possible causes of disorders of the v i s u o m o t o r system.
Acknowledgements. We are grateful to Mrs J.P. Donaldson for the preparation of all the histological material and to Mrs C. Wollaston for expert technical assistance. The work was supported by grants from The Wellcome Trust and the W.H. Ross Foundation (Scotland), whose assistance we acknowledge with gratitude.
References Anastasio AJ, Correia MA (1988) A frequency and time domain study of the horizontal and vertical vestibulo-ocular reflex in the pigeon. J. Neurophysiol 59:1143-1161 Arzi M, Magnin M (1989) A fuzzy set theoretical approach to automatic analysis of nystagmic eye movements. IEEE Trans Biomed Eng 36:954-963 Ashton JA, Boddy A, Donaldson IML (1984a) Directional sensitivity in the responses of units in cat primary visual cortex to passive eye movement. Neuroscience 13:653-662 Ashton JA, Boddy A, Donaldson IML (1984b) Input from proprioceptors in the extrinsic ocular muscles to the vestibular nuclei in the giant toad, Bufo marinus. Exp Brain Res 53:409-419 Ashton JA, Boddy A, Dean SR, Milleret C, Donaldson IML (1988) Afferent signals from cat extraocular muscles in the medial vestibular nucleus, the nucleus praepositus hypoglossi and adjacent brainstem structures. Neuroscience 26:131-145 Ashton JA, Milleret C, Donaldson IML (1989) Effects of afferent signals fi'om the extraocular muscles upon units in the cerebellum, vestibular nuclear complex and oculomotor nucleus of the trout. Neuroscience 31:529-541 Carpenter RHS (1972) Cerebellectomy and the transfer-function of the vestibulo-ocular reflex in the decerebrate cat. Proc Roy Soc Lond B 181:353-374 Carpenter RHS (1988) Movements of the eyes, 2nd edn., Pion, London Collewijn H (1989) The vestibulo-ocular reflex: is it an independent subsystem? Rev Neurol (Paris) 14518-9]: 502-512 Cooper S, Daniel PM (1949) Muscle spindles in human extrinsic eye muscles. Brain 72:1-24 Cooper S, Daniel PM, Whitteridge D (1953) Nerve impulses in the brainstem of the goat. Short latency responses obtained by stretching the extrinsic eye muscles and the jaw muscles. J Physiol (Lond) 120:471-490 Cooper S, Daniel PM, Whitteridge D (1955) Muscle spindles and other sensory endings in the extrinsic eye muscles: the physiology and anatomy of these receptors and of their connexions with the brain-stem. Brain 78:564-583 Cooper S, Fillenz M (1955) Afferent discharges in response to stretch from the extraocular muscles of the cat and monkey and the innervation of these muscles. J Physiol (Lond) 127:400-413
Donaldson IML, Knox PC (1990) Directionally-specific effects of afferent signals from the extraocular muscles upon responses in the pigeon brainstem to horizontal vestibular stimulation. Neuroscience 38:145-161 Donaldson IML, Knox PC (1991) Afferent signals from pigeon extraocular muscles modify the vestibular responses of units in the abducens nucleus. Proc Roy Soc Lond B 244:233-239 D6rrscheidt DG (1981) The statistical significance of the peristimulus time histogram (PSTH) Brain Res 220:397-401 Eden AR, Correia MJ (1981) Improved fixation of the pigeon brain by transcardiac carotid catheterization. Physiol Behav 27:947949 Gauthier GM, Nommay D, Verchier J-L (1990) Ocular muscle proprioception and visual localization of targets in man. Brain 113 : 1857-1871 Hayman MR, Knox PC, Dutia MB, Donaldson IML (1993) Effects of extraocular muscle afferent signals on the electromyogram of pigeon neck muscles during the vestibulo-collic reflex. J Physiol (Lond) 459: 458P Huber A (1988) Electrophysiology of eye muscles. In: Kennard C, Clifford Rose F (eds) Physiological aspects of clinical neuroophthalmology. Chapman and Hall, London, pp 377-390 Hulliger M, Matthews PBC, Noth J (1977) Static and dynamic fusimotor action on the response of IA fibres to low frequency sinusoidal stretching of widely ranging amplitude. J Physiol (Lond) 267:811-838 Karten HJ, Hodos W (1967) A stereotaxic atlas of the brain of the pigeon (Columba Iivia). Johns Hopkins University Press, Baltimore, Md. Kashii S, Matsui Y, Honda Y, Ito J, Sasa M, Takaori S (1989) The role of extraocular proprioception in vestibulo-ocular reflex of rabbits. Invest Ophthalmol Vis Sci 30:2258-2264 Keller EL, Robinson DA (1971) Absence of a stretch reflex in extraocular muscles of the monkey. J Neurophysiot 34:908-919 Knox PC, Donaldson IML (1991a) Afferent signals from the extraocular muscles of the pigeon modity the electromyogram of these muscles during the vestibulo-ocular reflex. Proc Roy Soc Lond B 246: 243-250 Knox PC, Donaldson IML (1991b) Effects of extraocular muscle afferent signals on units in pigeon brainstem during an artificial vestibulo-ocular reflex. J Physiol (Lond) 438:163P Melvill Jones G, Milsum J (1970) Characteristics of neural transmission from the semicircular canal to the vestibular nucleus of cats. J Physiol (Lond) 209:295-316 Robinson DA (1981) Control of eye movements. In: Brookhart JM, Mountcastle VB (eds) Handbook of physiology, sect 1, The nervous system, vol 2, American Physiological Society, Bethesda, Md., pp 1275--1320 Spencer RF, Porter JD (1988) Structural organisation of the extraocular muscles. In: Biittner-Ennever JA (ed) Neuroanatomy of the oculomotor system. Elsevier, Amsterdam, pp 33-79 Steinbach MJ (1987) Proprioceptive knowledge of eye position. Vision Res 27:1737-1743 Taylor A (1965) The role of sensory feedback in the vestibulo-ocular response in cats. J Physiol (Lond) 179:76-77P
