Eye movements evoked by superior colliculus stimulation in the alert cat

Brain Research, 106 (1976) 349-363

© ElsevierScientificPublishingCompany,Amsterdam- Printed in The Netherlands

EYE M O V E M E N T S

EVOKED

349

BY SUPERIOR COLLICULUS STIMULA-

TION IN THE ALERT CAT

A. ROUCOUX* AND M. CROMMELINCK Laboratoire de Neurophysiologie, Facult~ de M~decine, Universit~ de Louvain, B-1200 Brussels (Belgium)

(Accepted September 11th, 1975)

SUMMARY

(1) The electrical stimulation of the superior colliculus (SC) in the cat evokes exclusively conjugate and contraversive eye saccades. (2) Their maximum velocity is markedly higher than that of spontaneous saccades. (3) The stimulus parameters (intensity, frequency, pulse width) have but little effect on the characteristics of the saccades. (4) In the anterior half of the SC, corresponding to the projection of the 12-15 central degrees of the retina, amplitude and direction of saccades depend exclusively on the position of the electrode. (5) In the posterior half, corresponding to the projection of the peripheral retina, saccades are 'goal directed' and the position of the goal is determined by the location of the electrode. (6) An increase in stimulus train length produces, in the anterior part, a succession of identical saccades and, in the posterior part, a goal fixation. (7) Taking all these data into consideration, a model of the foveation process in the cat is proposed.

INTRODUCTION In 1870, Adamiik2 had already showed that contraversive eye movements could be evoked by electrical stimulation of the superior coUiculus (SC) in the cat. Stimulation of the lateral parts produced downwards eye movements and stimulation of the medial parts, upwards movements. Moreover, these evoked movements were, in some cases, accompanied by head turning. Various authors confirmed these resuits10-12,e2. In 1945, Apter 4 studied the retinotopic projection on the cat's SC. In * Charg6 de Recherches, F.N.R.S. (Belgium).

350 1946, she established the existence of a correspondence between the retinotop,c projection and the oculomotor output 5. At the same time, Hess et al. t3 described, m a more accurate manner, the motor effects of electrical stimulation of the SC m the free cat. These motor effects mainly consisted in contraverslve e2~e and head movements, the directlon of whmh varied in function of the stimulation site. These movements were qualified by these authors as v~sual grasp reflexes, their central control being mainly achieved by the SC. This idea was again maintained by Akert m 19493. Later on, Hyde and Eliasson 16 as well as Hyde and Eason 15 undertook, for the first Ume, a study of latency, velocity and type of eye movements evoked by long stlmulatmn trains in anaesthetized or enc6phale isol6 cats. Most generally, the movements were of a saccadic nature and ~goal directed', t.e. always directed towards the same point in the visual field. More recently, H o p f et al. 14 and Schaefer 18, operating on the alert or decortlcated cat, as well as Syka and Radii-Weiss 24, confirmed Hess et al.'s past results. Finally, Straschill and Rieger z3, in the spinally anaesthetized cat, described saccadic-like and goal-directed evoked eye movements as well as a correspondence between the sensory and motor maps. As far as the monkey is concerned, two studms have been carried out. Robinson ~7 established that stimulation of SC in unanaesthetized animals produced conjugate and contraversive saccades, the direction and amphtude of which solely depended on the site of stimulation and not on the initial eye positron. The m o t o r and sensory maps fitted very well together. On the other hand, Schiller and Stryker el presented analogous data. On the basis of all these observations, it becomes evident that the SC plays an important role in the orienting reflexes and especially in foveation. Nevertheless, a difficulty arises m comparing the results achieved in cat and monkey. The exclusive existence of goal-directed movements in one case and of movements independent of the initial eye position in the other, can only be attributed either to differences of methodological nature or to an actual functional difference. The aim of this series of experiments was to solve this problem. METHODS

Measures were carried out on 3 adult cats. Under Nembutal anaesthesia, a head holder made of two stainless steel tubes embedded into a dental cement crown was attached to the skull. In addition, two other devices were implanted: a chamber through which microelectrodes could be inserted stereotactically into the superior colliculus and 4 subcutaneous electrodes for electrooculographic recording (EOG). These electrodes were made of a piece of platinum wire recessed in a plastic cup filled with an Agar-NaC1 9 ~o gel and screwed to the bone; two of these were located at the outer canthi of both eyes for horizontal E O G derivation and another pair above and below the left eye for vertical derivation. Restraining method

During the experiment, the cat lay comfortably on a table with its head fixed in

351 a Horsley-Clarke stereotactic apparatus by means of two pairs of bars fitted into the steel tubes. The animal quickly got accustomed to this restriction of head movements and presented no sign of discomfort for experimental sessions of 1-2 h.

Calibration of the EOG recording EOG potentials between the two pairs of electrodes were amplified by two Tektronix 5A22N differential DC amplifiers and stored on magnetic tape (P.I. F M recorder). The central position of the eyes in the orbit was determined by noting on the tape, with a marker, along with the eye movements recording, the moments when the eyes were fixed upon a target placed in front of the animal in the longitudinal axis of its body, at a distance of 80 cm. This procedure was repeated at times during the recording session and allowed us to compensate for any d.c. shift. The calibration proper was achieved by attracting the cat's gaze to points situated at 10 and 20 ° away from the centre of its visual field, on the horizontal and vertical meridians, and noting these fixation periods on the tape. It has been found that a deviation of 10° in the horizontal plane corresponded most frequently to a shift of 400 #V and, in the vertical plane, to 200 #V. The linearity of the EOG was fairly good within 25 ° of eye deviation in either direction and the resolution was about 2°.

Stimulation and histology During all the stimulation sessions, the animals were maintained, by a subcutaneous injection of 0.5 mg/kg amphetamine, in a rather constant state of alertness. Stimulus current was delivered through conventional microelectrodes (glass coated platinum, Haer 30-10-2 type). Rectangular pulse trains were used. Pulse width was 0.5 msec and frequency 400 Hz. Train length varied from 30 msec to several seconds. The cathodal stimulus current was constantly monitored and never exceeded 500 #A. At the end of the experiments, the cat was anaesthetized and marks were made by passing a current through steel electrodes inserted stereotactically within the explored colliculus. The brain was perfused with a solution of ferrocyanide and the resulting blue spots were used to reconstruct the electrode tracks and determine the location of the stimulated sites. RESULTS

Type of evoked eye movements The stimulation of a site in the SC, by a 50 msec pulse train, evoked exclusively a conjugate eye movement of saccadic type. No smooth, vergence, monocular or nystagmic movements could be evoked. Fig. 1 shows a series of horizontal evoked saccades of increasing amplitude compared to a series of spontaneous saccades of similar amplitude. It appears that the essential difference between the two series of saccades is their velocity: evoked saccades are much faster than spontaneous ones. Fig. 2 illustrates the maximum velocity-amplitude relationship of both spontaneous and evoked saccades recorded in the same animal in a similar state of alertness. The maximum velocity of evoked as well as of spontaneous saccades is linearly related to

352

//

L

R

A 250 ms

L

R B

k_ Fig. 1. Example of 4 horizontal spontaneous saccades (A) compared to 4 horizontal evoked saccades

of similar amplitude (13)performed by the same cat in the same state of alertness. L, left; R, right.

their amplitude. Slopes of regression lines are of 30.8 ° • s-l/° for evoked saccades, and of 6.8 ° • s-l/° for spontaneous ones. Origins are approximately identical.

Threshold and latency Saccades appeared as all-or-nothing phenomena. The intensity which evoked a saccade for 100 ~o of stimulations (pulses of 0.5 msec width delivered at 400 Hz during 50 msec) was chosen as the threshold. The threshold had a value of 200/~A in the superficial collicular layers and dropped to 30-20 #A in the deep layers. A minimal value of 15/zA was found. Latency from the onset of the stimulation train to the beginning of the evoked saccade ranged from 20 to 30 msec. It was however fairly constant for each site of stimulation, though slightly decreasing when the intensity of stimulation was raised up to 2.5 times threshold and then remained unchanged in spite of a further intensity increase. Any modification of other stimulation parameters did not influence latency.

Influence of stimulation parameters Modification in pulse width or in frequency did not markedly influence the characteristics of the evoked saccades. Nevertheless, an increase of stimulus intensity up to 2-2.5 times threshold produced, at a given site, a slight increase of amplitude. This amplitude still increased if stimulation was further intensified but, then, the direction of the saccade changed. This can be attributed to the fact that the volume increase of the stimulated nervous tissue induced a phenomenon of spatial summation. We were able to ascertain such a summation by stimulating two colticular sites simultaneously. The amplitude of the saccade evoked by a 1.5 times threshold intensity was chosen as typical for the site under consideration. Below 20 msec of train duration, no saccade could be evoked. Above 60 msec, the eye moved in a particular way described hereafter. In practice, a train of 50 msec was adopted as a standard value.

353 de9 S'1 Maximum VetocJty

1200.

/

IOO0

[/ ~

y = 45 + 30.8x N = 120

/

80C

6OO

400 200 /

y =40 * 6,8x ~

N =400

Amp[ ' 5 ' 10 ' 1; ' ; 0 ' 2 ' 5 '

~:) ' ; 5 '

40'4;

'

deCl

Fig. 2. Amplitude-maximum velocity relationship of a sample of saccades evoked by stimulation of

several collicular sites. Each circle represents the mean maximum velocity of 10 saccades of similar amplitude. Vertical bars are the standard deviation from the mean. The linear regression line fitted to these observations is compared with the regression line representing the amplitude-maximum velocity relationship of spontaneous saccades recorded in the same aroused cat. N is the total number of observations.

Amplitude and direction of evoked saccades Amplitude and direction of evoked saccades were essentially determined by the stereotactic coordinates of the penetration, in a way that differed for the anterior and for the posterior part of the SC. Anterior half. The saccades evoked at one particular site of stimulation were contraversive, of nearly constant direction and amplitude regardless of initial eye position. Fig. 3 shows the saccades evoked at different initial eye positions by stimulation of a site located at A: 4.5, L: 2.5 in the left SC. The mean amplitude is 7.5 ° with a S.D. of 2 ° and the mean angle above the horizontal is 30° with a S.D. of 9°. For eccentric initial eye orientations the amplitude is generally smaller and the direction changes as follows: when the eye approaches the upper limit of its m o t o r range, the saccade becomes more horizontal and, similarly, becomes vertical for extreme initial positions to the right. Fig. 4 also illustrates this phenomenon: horizontal and vertical components remain unchanged when successively evoked by a repetitive stimulation until the eye approaches the limit. At 20 ° to the right, the vertical component seems to get saturated:

354

2l J

L. . . .

S

R

/

f D Fig. 3. Saccades evoked by shmulatlon of a site in the anterior half of the left SC. Each saccade ~s represented by a vector, the origin of which is the initial eye position and the length of which is proportional to its amplitude.

the evoked saccade then becomes horizontal. Similarly, at 20° upwards, the eye seems to reach a limit and no further saccade can be evoked. When the electrode progressiveIy penetrated through the various collicular layers, direction and amplitude of the evoked saccades did not change markedly, although a clear lowering of the threshold was observed. For each penetration, we calculated the mean amplitude and direction of the saccades evoked in intermediate layers: this gave us a sort of 'standard' saccade for each collicular site which was then reproduced on a dorsal view of the SC (Fig. 5A). When considering the anterior part of the SC, it appears that upward saccades were evoked near the midline, downward

t

x

\

'~

R

D

lS

Fig. 4. Saccadcs evoked try repetitive stimulation (50 mscc train duration) of a point in the anterior half of the left SC. Black bars indicate the stimulation periods.

355 B

Anterior

Al'4erlor

5ram

\\

o" .lo"

..___

~

.lo*

\ \

-

/

Lateral

/

.i/// /

/

/

/

j/ Posterior

0

1

2

3

/*

5rr.'n

D

Anterior

5

mm

IIo

--

,.....<-\

t J' I

3

,2

1

"<

t

/ j , , f~. 0

0

1

2

3

t,

5

mm

Fig. 5. A: dorsal view of the right SC. Mean direction and amplitude of saceades evoked at different sites in the anterior part of the SC (above the cross marks). In the posterior part, saccades evoked when the initial eye position is zero. Saccades are represented as in Fig. 3. B: retinotopic projection upon the SC modified from Feldon et al. 9. Vectors represent the theoretical evoked eye saccades established according to a simple foveation model (see text). C: motor map constructed with smoothed contours of equal amplitude (lateromedian lines) and of equal direction (anteroposterior lines). The cross marks separate the anterior and the posterior part. D: border between anterior and posterior parts of the SC projected upon the contralateral visual field.

saccades in lateral regions while horizontal saccades were distributed on an anteroposterior line approximately in the middle o f the SC. Moreover, the saccades o f the smallest amplitude were e v o k e d in rostral points. The smallest amplitude observed for this type o f e v o k e d saccade was 3.5 ° and the largest 12-15 °. The largest deviation from the horizontal was 80 ° upwards and 45 ° downwards. A lengthening o f pulse train produced an increase of the amplitude of e v o k e d saccades though direction and velocity were left unchanged. If the pulse train was long

356 u

-'-"~,

\ T = 450 T : 70

m

m

T= 250

o

SI

o

S2

T ms

\

~

doo L

m

T= 8 0 0

Q2

51= 100 04s

0

0.2

0.4s

Fig. 6. Temporal interaction of saccades evoked by two identical stimuh (Sl, S~) of 50 msec duration (black bars) given at the same site in the anterior half of the left SC. T ms is the delay between the onset of each stimulation train. $1 = 100 means that one single train of 100 msec has been delivered.

enough (in many cases 400-500 msec), the velocity decreased progressively as the eye came close to its mechanical limit of rotation. This type of movement evoked by long stimulation trains may be considered as a sequence of saccades more or less tied to one another. In a few cases indeed, a series of successive saccades separated by short intervals of fixation were evoked. This assumption is further supported by the observation illustrated in Fig. 6. Two successive stimulus trains of same intensity and lasting 50 msec each were delivered through the same electrode; each train evoked one saccade. These two pulse trains were separated by an interval T msec, varying from 450 to 50 msec. When T = 50, both stimulus trains followed one another without a break and the evoked saccade was similar to the saccade evoked by one single train of 100 msec (St = 100). One may notice that this last saccade actually represented the total sum of the amplitudes of the two saccades evoked by a different 50 msec train. The fusion of the saecades appeared as the duration of the interval decreased. This also shows that the relative refractory period following stimulation was shorter than 50 msec. It was however difficult to measure this value with precision in view of the relative inefficacy of stimulations shorter than 50 msec. Finally, in the case of a stimulus train length of several seconds, the eye rapidly reached its mechanical rotation limit. During such long stimulations, spontaneous movements were not completely suppressed. At the end, the eye often came back to a more central position; nevertheless, it is difficult to consider this phenomenon as an 'off movement' because of the varying time interval between the end of stimulation and the appearance of the return saccade. Posterior half. For a particular stimulat]on site, direction and amplitude of evoked saccades varied in function of the initial eye orientation in such a way that

357 25 (

tJ

/ L .....

/

/

f

D Fig. 7. Saccades evoked by stimulation of a site in the posterior half of the right SC. Each saccade is represented by a vector, the origin of which is the initial eye position and the length of which is proportional to its amplitude.

the eye always tended to gaze at the same area of the visual field; in other words, the movements were 'goal directed'. The goal was contralateral and the evoked saccades were always contraversive: saccades varying in direction beyond 180 ° were never observed. Fig. 7 illustrates saccades evoked from different initial eye orientations by the stimulation of a site located at A: 1.5 and L: 1.5. The saccades carry the eye towards a roughly circular surface in the visual field; this surface has an approximate radius of 6 °. The centre of this surface is situated in the lower left quadrant at 13° under the horizontal meridian and at 18° to the left of the vertical meridian. Unfortunately, initial positions situated very far away from the goal were rare, due to the fact that it was difficult to attract the animal's gaze in a direction diametrically opposite to that where stimulation had carried it. N o clearly ipsiversive saccade was observed though vertical saccades were noticed. When the eye stared at the goal, stimulation was ineffective. In the same way, repeated stimulation evoked one goal directed saccade, followed by a goal fixation or, in some cases, by a series of small movements within the goal area. Fig. 8A illustrates those phenomena. A long lasting stimulus train (500 msec for example) also evoked a saccade followed by a goal fixation (Fig. 8B). In most cases, the fixation went on after the end of the stimulation and no off or rebound movement appeared. Moreover, in the course of stimulation periods of several sec-

358 A 3o*u

R

B

30 °L

°i

3O*

R

1$

Fig. 8. A: saccades evoked by repetltwe stimulation (tram duration: 50 msec) of a site in the posterior half of the left SC. B: eye movements evoked at the same point by a 500 msec stimulus. Black bars indicate stimulation periods.

onds, the eye did not always remained goal-fixed and spontaneous saccades sometimes occurred. When the electrode was progressively introduced into this posterior part of the SC, the goal position did not vary markedly with depth although, as it has been noted before in the case of the anterior part, there was a perceptible lowering of the threshold at the level of the deep layers. For each penetration, the mean goal position attained by the saccades evoked at the level of the intermediate layers was calculated. On the dorsal view of the SC illustrated in Fig. 5A (posterior part), the saccade evoked when the eye is initially gazing at the centre of the visual field has been indicated for each penetration. The most eccentric goals were observed when stimulating the caudal part of the SC; goals situated above the horizontal meridian were observed when stimulating medially, while lateral stimulations evoked saccades pointing at goals in the inferior quadrant. Goals located on the horizontal meridian are distributed on an oblique line, from front to rear and towards the median line, in the prolongation of the line upon which horizontal saccades evoked in the anterior half of the SC are distributed. The maximum eccentricity of goals is 23 °, the minimum one is 13°. Motor map. As appears from Fig. 5A, evoked saccades are distributed across the SC in a regular manner, with an anteroposterior amplitude gradient and an approximately lateromedian tilt gradient. It has to be recalled here that, for the posterior half, it is referred to saccades evoked when the eye happens to be in the initial zero position. If, on the basis of these data, smoothed contours connecting various saccades of same amplitude and similar tilt are traced on a dorsal view of the SC, the map shown in Fig. 5C is obtained. The cross marks represent the boundary between the two collicular regions noticeable because of the characteristics of evoked saccades. These marks plotted on the motor field, represented in space, demarcate two zones: one, central, is approximately circular and its radius varies from 12 to 16°; the other, peripheral (Fig. 5D). The central zone corresponds to the anterior part of the SC

359 where evoked saccades do not depend on initial eye position; the peripheral zone corresponds to the anterior part where saccades are goal-directed. Comparison between motor and sensory maps Our results confirm the existence of a precise oculomotor topographic organization at the level of the SC. This motor 'map' coincides with the sensory or retinotopic map already described by several authors. This is illustrated by Fig. 5. The retinotopic map (Fig. 5B) was modified from Feldon et alP. The vectors drawn there represent the eye saccades one would obtain assuming that the stimulation of the SC at a given site evokes a saccade, bringing the gaze to the point of the visual field which is retinotopically projected on the same collicular site. Amplitude and direction gradients of such 'theoretical' saccades are very similar to those of saccades observed by us. The main differences between the theoretical and experimental maps are the following: (1) the central area (within a radius of 5°) of the retina is not so extended in the experimental map; (2) the motor counterpart of the ipsilateral visual field projection has not been observed and (3) the saccades of the largest amplitude (rarely exceeding 25° from the zero position) were evoked at sites corresponding to the projection of the retina's extreme periphery (30-70 ° of eccentricity). This could be related to the mechanical limits of the cat's ocular motricity which is about 25° from the centre. Peripheral goals would then not be within the 'oculomotor range'. Evoked head movements Several times we observed that stimulation of the posterior part of the SC by a 50 msec pulse train produced a contraction of contralateral muscles of the neck and, when the head was set free, its contralateral rotation actually occurred. This head movement did not appear when the anterior part was stimulated, at least when stimulus trains did not exceed 50 msec. Other effects of stimulation Besides eye saccades and head movements, stimulation of the SC also produced a movement of both pinnae, evoked only in the most caudal parts of the structure. On the other hand, stimulation of the tegmentum and of the central grey matter evoked various effects: mydriasis, pinnae, face and limb movements, aggressive reactions or signs of pain and, at least, centering eye movements. DISCUSSION

All these experimental data suggest that, in the cat, the SC constitutes a sensorymotor integrating centre, playing an important part in the 'foveation' reflex2°, or 'visuelle Greifreflex'la. This foveation model of the SC has been best illustrated by Schiller and Stryker21 in the monkey, showing that deep collicular neurones possess a superimposed sensory and motor field; these cells indeed discharge most vigorously before and during a saccade which brings a stimulus falling within their receptive field onto the fovea. Applied to the anterior half of the SC of the cat, this simple fovea-

360 tion model 17,20 gives a key to a general interpretation of our data. Evoked saccades appear, indeed, as all-or-nothing phenomena; their direction and amplitude are spatially coded at the level of the SC, i e. they are exclusively determined by the location of the electrode and independent of initial eye position. The error, simulated by stimulation, which the evoked saccade tends to annul, is defined with respect to the centre of the retina and is thus coded in a retinal coordinate system. This explains why a long lasting stimulus produces a succession of identical saccades, often t~ed together, bringing the eye to its mechanical limit, since central stimulation places the whole system in an "open loop' condition. Nevertheless, the foveatlon process does not account for the characteristics of the saccades evoked in the posterior half of the SC which receives the peripheral retinotopic projection ( > 12-15°). In fact, it is no longer the amplitude and the direction of saccades which are spatially coded, but the goal position, i.e. the final position of the eye. Now, the simulated error is no longer defined in a retinal coordinate system but with respect to the axis of the animal's head, i.e. m a head coordinate system. This involves the existence of a coordinate transformation process taking place somewhere between the retinotopic input and the motor output, calling upon information concerning the position of the eye in the orbit. This information could have a proprioceptive origin 1, or could be an efference copy 7. Hence, central stimulation no longer places the system in an 'open loop' condition and a repeated or long lasting stimulus will evoke a single saccade followed by a goal fixation. How could this double mechanism be interpreted? One could possibly imagine that a visual stimulus acting on the central part of the retina ( < 12-15 ° of eccentricity) produces a single fixation saccade On the other hand, a visual stimulus falhng on the periphery of the retina brings about a combined eye and head movement in the direction of the stimulus. Such a head movement, indeed, can be induced exclusively by the stimulation of the posterior half of the SC. Moreover, it should be emphasized that a head movement towards a visual stimulus must be programmed in a head coordinate system. In man, Yarbus 25 shows how fixation saccades of more than 15° are often accompanied by a corresponding head rotation. Besides this, it may be recalled here how Bizzi e t al. 6 have shown m the monkey that, in the case of conjugate head and eye movement, the motor command is almost simultaneous and uses probably the same information error. Quite a large number of stimulation experiments in cat's SC have already been carried out. Hyde and Eason 1~ and Straschill and Rieger 2a however were, to our knowledge, the only authors who actually measured the characteristics of evoked movements. Like Straschill and Rleger z3, we have established a narrow correspondence between the sensory retinotopic and oculomotor maps. This had already been ascertained by Apter 5. On the other hand, there are numerous points of disagreement between Hyde and Eason's 15 and Straschill and Rieger's 23 results and ours. In the enc~phale isol6 cat, Hyde and Eason 15 described disconjugate evoked movements, the temporal course of which is very &fferent from that of spontaneous saccades. The velocity reaches its maximum value 125 msec after the beginning of the movement (40-150 ° • s -1) and

361 then decreases exponentially to a value close to zero after 800 msec. Such movements are obtained by stimulations lasting at least 2 sec. In spinally anaesthetized cat, by means of stimulation lasting from 1 to 10 sec, Straschill and Rieger 2a evoked, with values of intensity close to the threshold, slow movements, 'saccades decelerating into slow movements' and also nystagmus, which we never did. These diverging points could be explained by a difference either in the animal's condition, or in the length of the stimulus train used. The fact that the state of alertness deeply affects the characteristics of spontaneous saccades has been demonstrated 8. In this context, it is quite possible that an enc6phale isol6 cat or a spinally anaesthetized animal displays an oculomotor activity, the characteristics of which are altered. Indeed, Straschill and Rieger 2a refer to the appearance, in some cases, of spontaneous nystagmus, a type of movement never observed in the intact animal. Moreover, according to our observations, it appears that movements evoked by stimulations lasting more than 50 msec, as regards the anterior part of the SC, represent in fact a succession of saccades bringing the eye to its mechanical limit. The velocity of such movements decreases progressively. Considered as a whole, these movements are comparable to those described by Hyde and Eason 15 using trains of 2 sec, and by Straschill and Rieger 2a also using trains of more than 1 sec. Straschill and Rieger 2a evoked by suprathreshold stimulation a majority of conjugate saccadic eye movements, the duration-amplitude relationship of which coincided with that of the spontaneous saccades performed by the same unanaesthetized animal. The saccades evoked in our experiments have a much higher velocity. The objection could be raised that such velocities are due to the stimulus intensities chosen (20-200 #A), which may seem rather high compared to Straschill and Rieger's (0.6-5 #A) 23. Now, of the one hand, we observed that the maximum velocity was almost independent of the stimulus intensity above the threshold and, on the other hand, with intensities equal to the ones used by Straschill and Rieger 23, we never could evoke saccades, even during very long stimulation periods. Such differences in threshold values can possibly be attributed to technical factors. Finally, it should be mentioned that our thresholds and our intensity range correspond to values used in the monkey 17,21. Hyde and Eason 15, as well as Straschill and Rieger 2a have elicited movements of 'goal directed' type only. Straschill and Rieger 2a occasionally refer to some cases where direction and amplitude of saccades are not very much affected by the initial eye position. It may happen that some goals described by these authors, after stimulation of the anterior part of the SC are, in fact, close to the mechanical limit of rotation of the eye in the orbit. This would explain their finding of a better correspondence of sensory and motor maps if, in a certain number of cases, they consider the position of the eye after the initial saccade of their evoked movements. Yet, some goal positions illustrated by Straschill and Rieger 2a, as regards the posterior part of the SC, correspond to our data. On the other hand, in an observation of the cat reported by Schiller TM, saccades described are exclusively not goal directed 'though not quite as neat as in the monkey'. In the monkey, Robinson ~7 as well as Schiller and Stryker 2~ have demonstrated

362 that s t i m u l a t i o n of the SC evoked only eye saccades, the direction and a m p h t u d e of which are absolutely i n d e p e n d e n t from initial eye position. Hence, m th~s ammal, the o r g a n i z a t i o n is s~mdar to that observed by us m the a n t e r i o r half of the cat's SC. As a conclusion, the mechamsms of c o m m a n d of fixation saccades seem to be more complex m the cat's SC t h a n m the monkey's. In the latter species, head movem e n t p r o g r a m m i n g is probably elaborated m a n o t h e r structure. ACKNOWLEDGEMENTS We wish to t h a n k Professor M. Meulders for his c o n s t a n t s u p p o r t a n d advice a n d Dr. J. M. G o d f r a i n d for his helpful criticism of the manuscript. W e are also grateful to Mrs. F. Langhendries, Miss F. Ameels, Mr. C. Hendrick, J. Schouppe a n d C. S t o q u a r t for their technical assistance. REFERENCES 1 ABRAHAMS,V. C., ANDROSE,P. K., Projections of extraocular, neck muscle, and retmalaffe rents to superior colhculus in the cat: their connections to cells of origin of tectospinal tract, J. Neurophysiol., 38 (1975) 10-18. 2 ADAMUK,E., Obgr die Innervatlon der Augenbewegungen, ZbL reed. Wiss., 8 (1870) 65-67. 3 AKERT,K., Der visuelle Greifreflex, ttelv, physiol, pharmacol. Acta, 7 (1949) 112-134. 4 APTER, J. T , ProJection of the retina on superior colliculus of cats, J. Neurophysiol., 8 0945) 123-134. 5 APTER, J. T , Eye movements following strychninizatJon of the superior colhculus of cats, J. Neurophysiol., 9 (1946) 73-86. 6 BIzzl, E, KALIL, R. E., MORA~O, P., AND TAGLIASCO,V., Central programming and peripheral feedback during eye-head coordination in monkeys, Bibl. ophthal. (Basel), 82 (1972) 220-232. 7 CROMrC~L1NCK,M., AND ROUCOUX,A., Properties of superior colliculus neurons during oculomotor activity in paralyzed cats, Arch. int. Physiol., 81 (1973) 477~,93. 8 CROMMEEINCK,M., AND ROUCOUX,A., Characteristics of cat's eye saccades in &fferent states of alertness, Brain Research, 103 (1976) 574-578. 9 FELDON,S., FELDON,P., AND KRUGER,L., Topography of the retinal projection upon the superior colliculus of the cat, Vision Res, l0 (1970) 135-143. l0 FERRIER,D., The Functions of the Brain, Smith Elder, London, 1886. l 1 GOLDLOWSKI,W , l-,es centres sous-corticaux du regard et des mouvements assoc~6s des globes oculalres, Tray. clin. Wab HerE. Univ, Cracovia, 1936, 98 pp (quoted by J. E. Hyde and S. G Eliasson, 1957). 12 GODEOWSKI, W., Experlmentelle Untersuchungen fiber die durch Reizung des Zwischen-und Mittelhirns hervorgerufenen assoziierten Augenbewegungen, Z. ges. Neurol. Psychiat., 162 (1938) 160-t82. 13 HESS, W. R., BURGI, S., UND BUCHER, V., Motorische Funktion des Tektal- und Tegmentalgebietes, Mschr. Psychiat. Neurol., 112 0946) 1-52. 14 HOPE, G., HEELER,H., AND SCHAEFER,K. P., Stimulation experiments on the tectum opt~cum of freely moving decortlcated cats, Electroenceph. clin. Neurophysml., 27 (1969) 620. 15 HYDE, J. E., AND EASON, R. C., Characteristics of ocular movements evoked by stimulation of brain stem of cat, J. Neurophysiol., 22 (1959) 666-678 16 HYDE,J. E., ANDELIASSON,S. G., Brainstem induced eye movements in cats, J. comp. Neurok, 108 (1957) 139-172. 17 ROBINSON,D. A., Eye movements evoked by colhcular stimulation m the alert monkey, Vision Res, 12 (1972) 1795-1808. 18 SCHAEFER,K. P., Unit analysis and electrical stimulatmn in the optic tectum of rabbits and cats, Brain Behav. Evok, 3 (1970) 222-240. 19 SCHILLER,P. H., The role of the monkey superior colliculus m eye movement and vision, Invest. Ophthak, 11 (1972) 451~[60.

363 20 SCHILLER,P. H., AND KOERNER,F., Discharge characteristics of single units in superior colliculus of the alert rhesus monkey, J. Neurophysiol., 34 (1971) 920-936. 21 SCHILLER, P. H., AND STRYKER, M., Single-unit recording and stimulation in superior colliculus of the alert rhesus monkey, J. Neurophysiol., 35 (1972) 915-924. 22 SPIEGEL,E., AND SCALA, N. P., Ocular disturbances associated with experimental lesions of the mesencephalic central gray matter, Arch. OphthaL, 18 (1937) 614-632. 23 STRASCHILL,M., AND RIEGER, P., Eye movements evoked by focal stimulation of the cat's superior colliculus, Brain Research, 59 (1973) 211-227. 24 SVKA,J., AND RADIL-WEISS,T., Electrical stimulation of the tectum in freely moving cats, Brain Research, 28 (1971) 567-572. 25 YARaUS,A. L., Eye Movements and Vision, Plenum Press, New York, 1967.