Orienting-related eye-neck neurons of the medial ponto-bulbar reticular formation do not participate in horizontal canal-dependent vestibular reflexes of alert cats

Orienting-related eye-neck neurons of the medial ponto-bulbar reticular formation do not participate in horizontal canal-dependent vestibular reflexes of alert cats

Brain Research Bulletin, Vol. 38. No. 4, pp. 337-347, 1995 Copyright ~ 1995 Elsevier Science Inc. Printed in the USA. All rights reserved 0361-9230/95...

1MB Sizes 0 Downloads 25 Views

Brain Research Bulletin, Vol. 38. No. 4, pp. 337-347, 1995 Copyright ~ 1995 Elsevier Science Inc. Printed in the USA. All rights reserved 0361-9230/95 $9.50 + .00

Pergamon 0361-9230(95)00106-9

Orienting-Related Eye-Neck Neurons of the Medial Ponto-Bulbar Reticular Formation Do Not Participate in Horizontal Canal-Dependent Vestibular Reflexes of Alert Cats TOSHIHIRO KITAMA, ALEXEJ GRANTYN1 AND ALAIN BERTHOZ Laboratoire de Physiologie de la Perception et de/'Action, C.N.R,S., Collage de France, 15 rue de I'Ecole de Medecine, 75270 Paris Cedex 06, France [Received 9 March 1995; Accepted 23 May 1995] ABSTRACT: Ponto-bulbar reticular formation neurons, including identified reticulospinal neurons, were studied in alert, head-fixed cats. Orion'dng-related neurons of "eye-neck" type (ENNs) were selected on the basis of qualitative correlations of their discharges with visually triggered eye saccades and electromyographic activity (EMG) of dorsal neck muscles. It was tested whether ENNs participate both in visually tdggered gaze shifts requiring eye-head coordination and in gaze-stabilizing movements, such as vestibulo-ocular and vestibulo-collic reflexes (VOR, VCR). Firing patterns were studied dudng passive sinusoidal rotation (0.2-1.0 Hz; 2.0-21.5 (leg peak-to-peak) in the horizontal plane. Responses to electrical stimulation of the superior colliculus and the vestibular nerve were recorded to assess the convergence of tectal and vestibular synaptic inputs. The same methods were applied to a control sample of neurons with discharges apparently "unrelated" to orienting movements. ENNs did not show any modulation of firing rate correlated to compensatory VOR or VCR during passive sinusoidal rotations. Among "unrelated" cells, the fraction of modulated units was close to that reported for reticular neurons projecting in the medial reticulospinal tract. Phasic and sustained components of ENN bursts were associated with anticompensatory movements induced by rotation, such as quick phases, ocular beating field shift, and the increase of EMG activity in neck muscles acting in the direction of passive rotation. Monosynaptic excitation from the contralateral superior colliculus was observed in 92.3% of ENNs, but only 2 out of 17 tested showed an excitatory response to vestibular nerve stimulation. In the control group of "unrelated" neurons the proportions of monosynaptic tectal and excitatory vestibular nerve inputs were, respectively, 75.6 and 71.4%. It is concluded that ENNs are specifically related to active gaze shifts, derived from either visual or from head velocity inputs. Rhombencephalic connections of vestibular nuclei to these neurons appear to be quite weak. Parallel inputs from the mid- or forebrain must be assumed to explain their firing patterns during rotation-induced anticompensstory gaze shifts. ~rdhin the studied range of frequencies and amplitudes of passive rotation, ENNs did not participate in the vestibulo-collic reflex. It is therefore unlikely that reticular neurons controlling orienting eye-neck synergies act also as a premotor pathway for gaze-stabilizing movements.

INTRODUCTION The medial ponto-bulbar reticular formation (RF) contains populations of functionally specific neurons, such as, for example, the medium-lead excitatory and inhibitory bursters and the omnipause neurons controlling the generation of saccadic eye movements [cf. 18,30]. Similarly, RF neurons discharging in relation to orienting eye-head synergies may represent a distinct population [13,42]. The discharges of these cells are correlated with ipsilaterally directed eye saccades and positions of fixation, as well as with dynamic and static components of the ipsilateral neck muscle activity. A majority of such "eye-neck" (EN) neurons receive monosynaptic excitatory input from the contralateral superior colliculus, project to the spinal cord, and establish collateral connections in the lower brain stem [13,14]. Of course, neurons whose discharges are related to eye and head movements are not the only examples of functional specificity within the RF. A number of other, more or less specific populations presumably involved in different kinds of skeletomotor activity have been described as well [cf. 15,41]. Besides the evidence for the functional specificity, arguments have been advanced in favor of a more generalized organization of the RF. They are based on the observations of a broad afferent convergence [1,10,33,40] and widely divergent connections of single neurons [19,37,39]. It has been suggested that some RF neurons could participate in different types of movements or postural adjustments, depending on the weighting of afferent inputs in a given behavioral situation [23]. With respect to the control of head movement during gaze shifts, experiments in decerebrate cats indicated that some reticulospinal neurons (RSN) could be engaged in two antagonistic patterns of motor behavior, such as stabilization of gaze in space during passive displacements of the head and body (vestibulocollic reflex, VCR) and active gaze shifts during orienting [35]. The latter authors hypothesized that the population of such RSNs acts as "the principal mediator of tectally elicited orienting responses, as it does for vestibular reflexes elicited by semicircular canal stimulation" [35, p. 247]. Because orienting-related "eyeneck" RSNs in alert cats [13,14,42] are localized in the same region of the RF as RSNs, which presumably integrate tectal and

KEY WORDS: Reticulospinal neurons, Superior colliculus, Vestibular nerve, Vestibulo-ocular reflex, Vestibulo-collic reflex, Cat.

To whom requests for reprints should be addressed. 337

338

KITAMA, GRANTYN AND BERTHOZ

vestibular inputs [35], it can be supposed that these two cell types actually belong to one population. However, the absence of active gaze shifts in decerebrate preparations makes inconclusive any comparison with the data obtained in alert animals. This study has been undertaken to test whether RSNs and other RF neurons discharging in relation to orienting movements of alert cats receive synaptic input from the vestibular nerve and can eventually contribute to gaze-stabilizing reflexes, such as the VCR and the vestibulo-ocular reflex (VOR). For this purpose, activity of the same neuron has been recorded during visually triggered or spontaneous gaze shifts and during passive sinusoidal rotation in the horizontal plane. Neurons were tested for convergence of tectal and vestibular nerve inputs using electrical stimulation of these structures. The study has been reported in the form of an abstract [20].

ABBREVIATIONS Abd-N: abducens nucleus; BDN: burster-driving neuron; EMG: electromyogram; EN: eye-neck synergy; ENN: eye-neck reticular formation neurons (reticulospinal and with unidentified projection); HRP: horseradish peroxidase; P: phasic; PS: phasicsustained; QT: quasi-tonic; RF: reticular formation; SC: superior colliculus (c: contralateral, i: ipsilateral); VCR: vestibulo-collic reflex; V.n.: vestibular nerve; VOR: vestibulo-ocular reflex.

MATERIALS AND METHODS

Surgical Procedures and Chronic Implantations Chronic experiments were performed on nine adult cats. Surgical procedures were conducted in aseptic conditions under pentobarbital anesthesia (initial dose 35 mg/kg IP). Supplementary doses of pentobarbital (on the average 3 - 5 mg/kgha) were administered when necessary, as indicated by acceleration of the heart rate, which was continuously monitored. Postoperative infection was prevented by treatment with local antiseptics combined with intramuscular injection of an antibiotic with prolonged action (Extencilline). To reduce the duration of barbiturate anesthesia the surgery was divided in two sessions, separated by at least 10 days of recovery period. The following implantations were performed during the first session: (1) Search coil for monocular recording of eye movements. (2) Two pairs of silver ball electrodes on the round window and adjacent bone for stimulation of the vestibular nerve on both sides. (3) Bipolar intramuscular electrodes made of multistrand, Teflon-insulated wires to record the electromyogram (EMG) of the following neck muscles: m. obliquus cap#is cranialis, m. Iongissimus cap#is and m. splenius. (4) An acrylic crown, fixed to the skull and embedding three stereotaxically oriented screws for painless head fixation during recordings. The preparation for the chronic experiment was completed in the second surgical session by adding the following implants: (1) Varnish-insulated needle electrodes for stimulation of the superior collicululi (SC). The positioning of the electrodes in the deep layers was guided stereotaxically, with final adjustment by searching the point of the lowest threshold for eye movements or neck muscle contractions. (2) A recording chamber, cemented between the tentorium and the occipital crest, giving access to the lower brain stem through intact cerebellum. (3) In two cats floating needle electrodes were inserted in the ventral funiculi of the spinal cord at C2 level.

izontal plane. The head was fixed with nose-down inclination of 21 °. lntraaxonal or extracellular unit recordings were made with glass microelectrodes filled either with 3.8 M NaCI or 10% solution of horseradish peroxidase (HRP) in buffered 0.5 M KC1. Cell firing was recorded during either spontaneous or visually evoked eye movements and attempted head movements. The standard visual stimulus consisted of a small disk moving at constant velocity along the horizontal meridian of a tangent screen placed in front of the animal. To obtain more vigorous orienting responses, differently shaped objects or food were introduced in the visual field. After these visuo-motor tests, vestibular stimulation was provided by passive whole-body rotations in the horizontal plane, both in the light and in darkness. The frequency and amplitude ranges of sinusoidal oscillations of the servo-controlled turntable were 0.2-1.0 Hz and 2.0-21.5 °, respectively. Some neurons were tested only with manual rotation of the turntable. Filtered neuronal spikes, raw neck EMG activity, horizontal and vertical eye positions, and angular position of the turntable were recorded on magnetic tape and analysed off-line according to previously described procedures [13].

Identification of Synaptic Inputs, Projection, and Location of Neurons Synaptic inputs were evaluated by recording responses to electrical stimulation of the tectum and the vestibular nerves (V.n.) on both sides. Square current pulses of 0.2 ms duration with intensities of 30-500 #A were used for stimulation of the SC. Single pulses or 2 - 5 repetitive stimuli at 200 Hz were used for stimulation of the V.n. Their intensities (30-500 #A) were above thresholds (30--100/.tA) for eliciting monosynaptic responses of secondary vestibular axons and disynaptic field potentials in the abducens nucleus (Abd-N). Stimulus spread to the facial nerve occurred with currents 4.7-7.3 times the thresholds specified above. During initial tests the stimulus strength was gradually increased, until the appearance of facial twitches. If reticular neurons did not respond at these intensities, further increase of current (up to 9 × threshold) usually remained ineffective. The method of implantation of stimulating electrodes did not ensure a selective stimulation of the horizontal canal nerve. Although the stimulus-evoked eye movements were close to horizontal, activation of afferents from other canals and otoliths could therefore be possible. In two animals, neurons projecting to at least the second spinal segment were identified by antidromic stimulation. Collision tests were used to prove the antidromic conduction (Fig. 2A b,c, 3A a,b). In other experiments RSNs were identified by intraaxonal injections of HRP. Projection to the spinal cord was considered as confirmed if labeled axons were followed caudally to the level of or beyond the pyramidal decussation. Technical details of HRP injection and histochemical procedures have been described previously [ 16]. Coordinates of each recorded unit were calculated relative to a permanent stereotaxic reference on the skull. Intra-axonal HRP injections were made in all animals, and locations of injection sites were used to correct errors inherent to the stereotaxic method. After correction, cell locations were plotted on standard parasagittal maps of the lower brain stem (Fig. 1). RESULTS

Cell Locations Experimental Protocol During recording sessions cats were alert, with limb and body movements restrained by a bag that prevented lateral body displacements during sinusoidal rotations of the turntable in the hor-

The explored region extended from 1.4 mm rostrally to 3.5 mm caudally of the Abd-N (Fig. 1 A,B). A majority of tracks were made through the caudal half of the caudal pontine reticular nucleus and the rostral third of the gigantocellular nucleus. For

T E C T O - V E S T I B U L A R C O N V E R G E N C E ON R E T I C U L A R NEURONS

A%T

,

Rv/Rpm ,,'

339

. ?d-N .



" o ~.'"-t~.,.,,~.

,,' RPo

/

-

/

~" "',,...~'~""

/' -

._'.',.~..

7/

B Xll

?",,, . "~





°~_~~RT

X-\ / \ ~

FIG. I. Topographical distribution of sampled neurons in projection on the parasagittal plane. (A) Orienting-related eye-neck neurons (large dots) and "'unrelated" neurons with other types of behavior (small dots). (B) Neurons showing facilitatory responses to stimulation of the vestibular nerve (large dots). Cells responding at presumably di- or polysynaptic latencies and activated from ipsi- and/or contralateral side are represented together. Stars: two orienting-related neurons that responded to V.n. stimulation. Small dots: unresponsive neurons. Orientation of maps is shown by arrows on the upper left (D: dorsal; R: rostral). Dorsal surface of the brain stem is delineated by two contours, corresponding to midline (thin line) and 1.2 mm laterally (thick line). Abbreviations: Abd-N: abducens nucleus; CT: trapezoid body; NRT: nucl. reticularis tegmenti pontis; OI: inferior olive; RGc: nucl. reticularis gigantocellularis; RPc: nucl. reticularis pontis caudalis; RPo: nucl. reticularis pontis oralis; Rv/Rpm: nucl. reticularis ventralis and paramedianus; XII: hypoglossal nucleus. two-thirds of the units included in Fig. 1 the locations have been precisely determined. These were neurons with action potentials corresponding to usual criteria of somatic extracellular recording, some neurons recorded intracellularly in the soma, and those labeled with HRP. The remaining neurons were recorded intraaxonally but close to the cell body, as indicated by electrotonitally attenuated EPSPs. Intra-axonal records without EPSPs have been discarded. Recording sites represented in Fig. 1 A,B were located between 0.3 and 2.0 mm from the midline. In the group of somatic recordings, 75% of cells were in the medial RF, the remaining ones within the paramedian tracts. Such a distinction cannot be made for intra-axonal recordings with EPSPs because the actual distance to soma is uncertain.

Neuronal Samples Defined by Behavioral Criteria Neurons related to orienting movements. In this sample of 71 RF neurons, 26 were classified as orienting-related EN neurons. Their selection was based on the previously described qualitative criteria [ 13,14]. The main characteristic of EN neurons consists in a highly reproducible generation of burst discharges concomitantly with ipsilateral saccades, especially when the latter are

centrifugal and accompanied by contractions of ipsilateral neck muscles. According to the temporal characteristics of bursts, EN neurons can be subdivided in phasic, phasic-sustained, and quasitonic [14]. A more detailed description of typical firing patterns of the latter two groups of EN neurons will be given below. The locations of the studied EN neurons are shown in Fig. I A by large dots. Four of them were phasic (P), 16 were phasicsustained (PS), and 6 were quasi-tonic (QT). Because the scope of the study has been to search for modulation of firing rate correlated to natural vestibular stimulation in the relatively low frequency range, conclusive results were expected only from neurons generating sustained discharge components. Firing during sinusoidal rotation tests has therefore been analysed only on cells of PS and QT types (n = 22). Among these, 12 neurons (4 QT and 8 PS) were tested for the presence of spinal projection by antidromic stimulation or by intra-axonal HRP injections, and 10 (3 QT and 7 PS) were identified as RSNs. The orienting-related activity was not different between EN-RSNs and EN neurons with unidentified projections. All PS and QT neurons were therefore pooled together and will be referred to as ENN (Eye-Neck neurons, including RSNs and reticular neurons with unidentified projection).

340

KITAMA, GRANTYN AND BERTHOZ

A a

b

c

2 msec

C

B

-

Ev

~

~-~---

[ 10deg up down

[

right 10deg left

HEAD

0.5Hz

right

R-LONGISSIMUS

[

10deg

left

imp.Is 100 1

2

L lsec

50

lsec

FIG. 2. Functional characteristics of an orienting-related quasi-tonic eye-neck RSN recorded on the right side. (Aa) Response to a double shock stimulation of the contralateral superior colliculus (SCcl, 2) at intensity 1.7 X threshold. Note that monosynaptic activation is reliable only after the second shock. (Ab) Antidromic response to spinal stimulation at C2 level (Sp), given after the second shock of SCc stimulation. (Ac) Collision of antidromic with orthodromic spikes when the latter precede spinal shock by about 2 x antidromic latency. Only second collicular stimulus is shown (SCc2). This neuron did not respond to V.n. stimulation of both sides. (B) Activity of the neuron during visually triggered gaze shift accompanied by phasic-tonic contraction of neck muscle. (C) Activity during sinusoidal rotation in the light (0.5 Hz, 6.9 ° peak-to-peak) while attracting the cat's gaze to the right. See text for explanation of motor events corresponding to Bursts 1 and 2. Arrangement of records (from top to bottom): Vertical (EV) and horizontal (EH) eye position (interrupted line: straight ahead position in the horizontal plane); Turntable position (HEAD, in C only); Rectified and integrated (time constant 5.0 ms) EMG of right longissimus capitis muscle, with zero level shown by horizontal line; Instantaneous firing rate (FR) of the neuron.

Control sample of "unrelated" neurons. ENNs were compared to RF neurons sampled without any particular behavioral criterion (n = 45). In a majority of these neurons any changes of activity coinciding with orienting eye movements and/or EMG events were clearly absent. Some of them were silent neurons; others generated spontaneous bursts separated by periods of silence or discharged continuously. After establishing the absence of firing rate modulation during visuo-motor tests, we did not pursue the search of other possible behavioral correlates. A small number of neurons (n = 9) showed inconsistent relationships to orienting movements. In these cells firing rate increments during eye-neck synergies were observed only from time to time during the recording. The relationships of firing with motor parameters and the directional preference could change over time. The behavior of these cells did not correspond to identification criteria of ENNs, and they were included in the sample of "unrelated." " U n r e l a t e d " neurons were intermingled with ENNs, as shown by small dots on the map of Fig. IA. Out of 12 neurons

tested for the presence of spinal projection, 7 were identified as RSNs by antidromic stimulation and 1 by HRP labeling.

Activity Patterns During Visuo-Motor Behavior and Natural Vestibular Stimulation In the description of relationships between cell activity and motor events during natural vestibular stimulation we shall use the following terminology. Slow eye movements in the direction opposite to that of passive whole-body rotation (VOR) will be called "compensatory." They help to stabilize the visual image on the retina. When the head is fixed relative to the body, an attempted "compensatory" head movement is manifested by an increase of EMG activity of neck muscles on the side opposite to the direction of passive rotation (VCR). Motor events that underly gaze shifts in the direction of rotation (commonly referred to as "looking where you g o " ) will be called "anticompensatory." These are quick phases of the vestibular nystagmus,

TECTO-VESTIBULAR CONVERGENCE ON RETICULAR NEURONS

shifts of the average eye position (beating field) in the direction of rotation, and increments of EMG activity in muscles that would turn the head in the same direction. Quasi-tonic ENNs. Spike activity during natural vestibular stimulation was studied in 6 QT-ENNs, three of them identified as RSNs. Observations were qualitatively the same on all cells. They will be illustrated by an example of one representative neuron (Fig. 2). As shown in Fig. 2Aa, the neuron received a monosynaptic excitatory input from the contralateral SC. In this particular case spikes at monosynaptic latencies (0.85-1.2 ms) were consistently evoked only by the second of the paired shocks. Antidromic invasion from the cervical spinal cord (latency 0.7 ms) (Fig. 2Ab) was confirmed by collision test (Fig. 2Ac). Spike responses could not be induced by stimulation of the ipsilateral SC and of the V.n. on both sides. During visuo-motor tests the neuron fired at a low and irregular rate when the eyes were close to midposition (Fig. 2B). Its discharge rate increased together with the increments of ipsilateral horizontal eye position and of the level of ipsilateral neck EMG activity. The eye position threshold at which the sustained firing became clearly developed was about 5 ° ipsilaterally. During dynamic portions of gaze shifts to large eccentricities (e.g., from 2.9 to 19.2° in Fig. 2B) the neuron fired at much higher rates. As can be seen in Fig. 2B, phasic discharge components could sometimes occur synchronously with saccades and sharp increments of neck EMG. However, such a temporal correlation could not be proven because of a strong irregularity of firing rate. The example of Fig. 2B shows also that spike activity at a reduced and irregular rate could persist during eccentric eye fixations lasting up to 2 s. A reproducible activation during ipsilateral gaze shifts involving the eye-neck synergy, absence of clearly identifiable movement-related burst components and persistence of the discharge during prolonged eccentric eye fixations are the properties that allowed to classify this neuron as a QT-ENN. Passive sinusoidal head rotation at or below 1 Hz, either in the light or in the dark, usually induced quick phases and periodic shifts of average eye position in the direction of rotation. Such shifts of the ocular beating field usually coincide with the increase of neck EMG activity on the side toward which the animal is rotated, that is, opposite to that of the compensatory VCR 142,43]. The latter, consisting in the EMG modulation with frequency-dependent phase lead relative to contralateral head position [4,5,11], has been observed only rarely. This occurred when passive oscillations did not induce nystagmus and anticompensatory eye deviations (Fig. 2C). However, besides the presence of compensatory VCR, a second condition had to be fulfilled to detect the corresponding modulation of QT-ENN discharges. Because these cells fire tonically only when the eyes are deviated to the ipsilateral side, conclusive tests bad to be made while attracting the cat's gaze to ipsilateral eccentricities larger than the eye position threshold of a given neuron. This was done by pretenting in the periphery of the visual field various objects to which the animal had not yet been habituated. Obviously, such tests could be done only in the light. Fig. 2C shows an example of recording from the QT-EN-RSN during the VOR without quick phases, accompanied by a compensatory modulation of the neck muscle activity (VCR). The neuron did not show any firing that would be phase-locked to sinusoidal oscillations, even though the eccentricity of eye position and the level of EMG activity were well above the threshold for recruitment of this cell during visuo-motor tests (Fig. 2B). VCR or VOR-related modulation was absent in all six neurons of QT type. However, in one neuron this negative result was judged to be inconclusive, because a sufficient eye eccentricity in its ON-direction could not be attained.

341

The specimen record of Fig. 2C shows that during vestibular stimulation cell activity was associated only with rapid gaze shifts reaching large horizontal eye eccentricity. In one case (Burst 1) the neuron was recruited toward the end of rightward rotation when two beats of nystagmus shifted the eye position to the right, that is, in anticompensatory direction. At the same time right longissimus EMG showed a fairly abrupt increment. Because this activation began before the end of rightward rotation, it can be interpreted as an attempted anticompensatory head movement. In the second case (Burst 2) cell activity was associated with a large rightward saccade and EMG burst during spontaneous refixation of gaze on the attracting object on the right. Other QT-ENNs behaved similarly to the illustrated one.

Phasic-Sustained ENNs In the sample of 16 PS-ENNs, 11 neurons (including 7 identified RSNs) were tested by natural vestibular stimulation. A representative example is given in Fig. 3. This neuron responded at monosynaptic latency (1,0 ms) to stimulation of the contralateral SC and antidromically to spinal cord stimulation (Fig. 3A). Visually triggered gaze shifts to eye positions larger than 8 - 9 ° were accompanied by phasic bursts, which were synchronous with ipsilateral saccades and with fast increments of neck EMG activity (Fig. 3B, Bursts 1-3). When postsaccadic eye positions were larger than 14°, the initial phasic bursts were followed by a sustained discharge with a slowly decaying rate (Fig. 3B, Burst 4). During passive sinusoidal rotation (Fig. 3C), accompanied by intermittent anticompensatory movements, discharge patterns showed a similar dependence on eye position. At smaller eccentricities, phasic bursts were generated during ipsilateral quick phases (Bursts 1-3,5,6). At larger eccentricities, the firing became sustained (Burst 7). The behavior of PS-ENNs was therefore qualitatively the same during visually triggered gaze shifts and anticompensatory gaze shifts occuring during natural vestibular stimulation in the dark or in the light. Because PS neurons usually had high eye position thresholds for sustained activity, only three of them could be tested during compensatory VCR combined with a sufficiently large horizontal eye eccentricity. None of them showed any activity that would correspond to the slow modulation of the EMG under these conditions. Rotationrelated activity was not observed also on the remaining eight tested neurons. Because eye eccentricities required for the development of long lasting discharges could not be attained, negative results from these neurons have not been taken in consideration. Sample of "unrelated" RF(S)Ns. In the sample of 45 cells defined as "unrelated" to orienting movements, 34 were tested by natural vestibular stimulation. One of the neurons showing firing rate modulation correlated to sinusoidal horizontal oscillations is shown in Fig. 4. This neuron responded to contralateral SC stimulation at a minimal latency of 1.3 ms (range 1.3-3.0 ms), which suggests a disynaptic transmission (see below). Spike responses to ipsilateral stimulation were absent. Polysynaptic excitatory responses were elicited by V.n. stimulation on both sides. During visuo-motor tests this neuron fired spontaneously with an irregular rate of about 15 spikes/s. Ocasionally it generated discharges of phasic-sustained type during ipsilateral gaze shifts. However, transient activation with contralateral saccades was observed as well. As shown in Fig. 4B, sinusoidal rotation in the light induced the VOR, the VCR, and a modulation of cell discharge. Firing rate increased during contralaterally directed rotation (Type II response) and fell to zero during rotation in the opposite direction. The peak of firing rate coincided approximately with the maximum of rightward eye position during VOR. It was, however, delayed with respect to the peak of VCR activity in the ipsilateral neck muscle.

342

KITAMA, GRANTYN AND BERTHOZ

'~

B

,,

~'~

!i

!

0.5mV 1msec

C

1 i

~

[

...............

HEAD

i lOdeg t down

! i

right 10~1 fl

"----~------~- . . . . . . . . . . . . . . . . . . . .

! . . . . . . ~-----" . . . . . . . . . . . . . . . . . . . . . . .

1__. e g

: .

i=

[20deg

:

' ~0.Z3Hz

'i

12 13

4

2OO I imp,Is

15 16 7

L ,il

0 1see

FIG. 3. Functional characteristics of an orienting-related phasic-sustained RSN recorded on right side. (Aa, Ab) Monosynaptic response to single shock stimulation of the contralateral superior colliculus (SCcl) followed by spinal stimulation at C2 level (Sp). Collision of antidromic with orthodromic spikes is shown in Ab. Note the presence of the negative field potential after spinal stimulation. (B) Cell activity during visually triggered eye movements and accompanying neck muscle contractions. (C) Activity during horizontal rotation in the light (0.33 Hz, 13.1° peak-to-peak) while attracting cat's gaze to the right. Arrangement of records and other conventions as in Fig. 2. Individual bursts are referred to in the text by their numbers. Vertical broken lines are drawn through the onsets of those saccades and quick phases that are synchronous with bursts,

A clear modulation of firing during horizontal oscillations was observed in 5 out of 34 (14.7%) tested neurons. Modulation was of Type II in one (Fig. 4B) and of Type I in other four. Among five modulated cells, four had a spontaneous background activity, and the remaining one was a predominantly silent neuron generating only sporadic bursts. This might indicate that spontaneous firing represents a characteristic feature of neurons modulated by passive head movements, as it has been reported in decerebrate cats [36]. However, this does not fit our observations in alert cats: a sufficient background activity was present in 16 out of 34 "unrelated" RF(S)Ns but, as noted above, only 4 of them were modulated during sinusoidal rotation.

Responses l~duced by Superior Colliculus and Vestibular Nerve Stimulation In addition to the behavioral tests, the convergence of tectal and vestibular afferents has been examined using electrical stimulation of the SC and of the V.n. For monosynaptic SC responses the upper limit of EPSP latency was set at 1.0 ms, that of spikes at 1.2 ms. In responses to V.n. stimulation, EPSPs, when present, had a slow onset that made uncertain the measurements of their latencies. The latencies of spikes showed fluctuations in all cells, their minimum values ranging from 1.25 to 10.0 ms. On the basis

of preceding studies [31,32], reproducible spike responses at latencies -< 2.0 ms were judged to be disynaptic. Within the whole sample, ENNs and "unrelated" neurons taken together (n = 71), all cells were tested by contralateral SC stimulation and 62 by stimulation of the SC on both sides. V.n. responses were studied on 59 cells (47 tested from both sides and 12 unilaterally). Locations of neurons that responded to V.n. stimulation are shown in Fig. 1B (large dots). They were encountered over the whole explored region and, in contrast to ENNs (Fig. 1A), did not show any particular concentration in the immediate vicinity of the Abd-N. Quantitative data on the responsiveness and convergence in the samples of ENN and "unrelated" neurons are summarized in Table 1 and Fig. 5. Neurons related to orienting movements (ENNs). In the group of ENNs, monosynaptic EPSPs and/or spikes were elicited by contralateral SC (SCc) stimulation in 24 (92.3%) cells (Fig. 5A). Only two cells responded disynaptically. One of them received unilateral input from SCc (Table 1, column 2), and the other was activated from both sides and therefore was included in Column 3 of Table 1. Asymmetry of tectal input, that is, excitatory effect from contralateral side only, has been previously described as a typical feature of ENNs [13]. This property was tested by SC stimulation on both sides for 25 cells. Sixteen of them did not respond to ipsilateral stimulation (SCi), and 5 showed an IPSP.

TECTO-VESTIBULAR

CONVERGENCE

A

ON

RETICULAR

NEURONS

343

B

SCc

[up

- - ~ "

Ev

10deg

down

v

lr'"

"t

10deg left

Sp

l o. mv lrnsec

EH

,~

~'

RWD [ 1LOODdeg/s

R - LONGISSIMUS

~(

L

V.n.-c 50 imp.Is 5rnsec

25

V.n.-i

HEAD

1 Hz

right lOdeg left

5rnsec

1 sec

FIG. 4. Functional characteristics of one of the five reticular neurons that showed firing rate modulation during passive horizontal oscillations but no relationship to orienting eye-neck synergies. (A) From top to bottom: spike responses to contralateral superior colliculus stimulation (in disynaptic latency range); antidromic response to spinal stimulation (latency 0.7 ms); responses, in polysynatic latency range, to contra- (V.n.-c) and ipsilateral (V.n.-i) vestibular nerve stimulation. (B) Modulation of firing rate during horizontal rotation in the light (I .0 Hz, 7.0 ° peak-to-peak). Neuron is located on the right side, modulation is of Type II. Arrangement of records and conventions as in Figs. 2, 3, except that eye velocity (EH) trace (third from top) is added, and turntable position record ( H E A D ) is at the bottom.

TABLE

1

RESPONSIVENESS TO STIMULATION OF THE SUPERIOR COLLICULUS AND VESTIBULAR NERVE OF ENNs (ON THE LEFT) AND NEURONS "'UNRELATED" TO ORIENTING MOVEMENTS (ON THE RIGHT) Eye-Neck Neurons n = 26 I

2

3

"Unrelated" Neurons n = 45 4

5

1

2

3

4

5

NR

Total

1 --1 --

10 16 4 12 3

2

45

Contra Contra SC V.n.~ Contra Ipsi Both sides NR nt Total

I I I

Mono

Mono

Di-Poly-

Both Sides

I (PS) 1 (P) __ 13 6

--__ -I

---2 2

1

4

21

,] I 1

NR

Total

------

1 I 0 15 9

0

26

a

b

Di-Poly-

Both Sides

2 3 3 --

5 7 1 7 1

1 1 1 1 1

I 5 2 -1

8

21

5

9

F

I I I

-

]

I I _1

Columns: 1: numbers of cells with excitatory responses to stimulation of the contralateral superior colliculus (SCc) at monosynaptic latencies ( l a cells tested unilaterally, I b-bilaterally); 2: at di- or polysynaptic latencies; 3: to bilateral stimulation, irrespective of latency; 4: lacking the response. C o l u m n s 1, I b, and 2 contain ceils with no response or inhibition from the ipsilateral SC (asymmetric input from two sides). Letters in parentheses (Column 1) specify phasic-sustained (PS) and phasic (P) types of ENN. Rows: numbers of cells responding to vestibular nerve (V.n.) stimulation on contra- or ipsilateral side or bilaterally, irrespective of latency; NR: unresponsive cells; nt: vestibular nerve stimulation not tested. Numbers of cells showing tecto-vestibular convergence appear inside the dashed frame.

344

KITAMA, G R A N T Y N AND BERTHOZ

Fractions (per cent) 0 I

A

20 L

B

C

V.n. D,P

D

V.n. D,P & SCc M,D,P

E

V.n. D & SCc M

~

I

60 ~

1

80 ~

I

100 i

I

26 45

SCc M

SCc M,D,P & SCi IPSP, NR

I

40

25 35

~

17. 42"k

~

40*

17.

t4O17

FIG. 5. Percentages of cells responding to tectal and/or vestibular nerve stimulation. Black bars: ENNs, hatched bars: RF(S)Ns "unrelated" to orienting movements. Numbers of tested neurons indicated to the right of each column. Asterisks mark differences between samples with significance levels ofp < 0.05 (*) and p < 0.001 (**). Each pair of bars shows percentages of cells having one of the following characteristics: (A) monosynaptic excitatory response to contralateral SC (SCc); (B) monosynaptic (M), di- (D) or polysynaptic (P) excitation from SCc and IPSP or no response (NR) to ipsilateral SC (SCi); (C) di- or polysynaptic response to vestibular nerve (V.n.), ipsilateral, contralateral, and bilateral effects pooled together; (D) convergence of excitatory effects from V.n. and SCc without distinction by latency; (E) convergence of disynaptic excitatory effect from V.n. with monosynaptic from SCc, which is equivalent to convergence of direct (monosynaptic) inputs from second order vestibular neurons and SCc.

umns lb and 2; Fig. 5 B). Bilateral excitatory effects were encountered in nine neurons, of which five responded to SCc stimulation at monosynaptic and four at di- or polysynaptic latencies. Ipsilateral effects were indirect, with exception of one cell responding to SCi at a monosynaptic latency. Out of 42 neurons tested by V.n. stimulation on at least one side, facilitatory responses were recorded in 30 (71.4%; Fig. 5C). Facilitation was contralateral in 9, ipsilateral in 16, and bilateral in 4 neurons (Table 1). Responses in the disynaptic latency range were observed on 8 out of 30 cells (26.7%). For neurons receiving excitatory SCc input, either monosynaptic or indirect, the probability of convergence with V.n. input was 72.5% (Table 1, Fig. 5D). Convergence of monosynaptic SCc and disynaptic V.n. effects was observed on 7 out of 40 neurons (17.5%) receiving input from both structures (Fig. 5E). This fraction can be compared to 1 out of 17 in the ENN group. As indicated above, the overall responsiveness to V.n. stimulation was rather high (71.4%) in the sample of "unrelated" RF(S)Ns, in contrast to a low number of cells modulated by passive rotation (n = 5). Among the latter, excitatory effect of V.n. stimulation was revealed in three cells. However, 18 out of 29 unmodulated cells were also responsive to V.n. stimulation on ipsi- or contralateral sides. Therefore, any relationship appeared to be absent between the probabilities of response to electrical and natural vestibular stimuli. Comparisons between the samples. Comparisons between ENN and "unrelated" samples with respect to proportions of neurons responding to SC and/or V.n. stimulation were compared using the normal deviate (z) (Fig. 5). The probability of a monosynaptic excitatory response to SCc was significantly higher in the sample of ENNs (Fig. 5A) than in the sample of "unrelated" neurons (z = 1.724, p = 0.043). The differences between ENN and "unrelated" samples were highly significant (z > 4.000, p = 0.001) with respect to the overall probability of response to V.n. stimulation (Fig. 5C) and with respect to the probability of tecto-vestibular convergence (Fig. 5D). All other differences, including that describing the asymmetry of contra- vs. ipsilateral effects, were not significant (p > 0.05). DISCUSSION

A clear asymmetry of responses to contra- and ipsilateral stimulations was therefore present in 84.0% of ENNs (Fig. 5B). Only four cells received excitatory input from both sides (Table 1, third column), Effects of V.n. stimulation were tested on 17 ENNs (Table 1), in the same range of stimulus intensities (see Methods) as that used for "unrelated" neurons. Because all of them responded to SC stimulation, this sample was also used to evaluate the probability of the tecto-vestibular convergence. Only two neurons responding monosynaptically to SCc, were also facilitated by stimulation of the V.n. (Table l, first column; Fig. 5C,D). One received contralateral V.n. input at a minimum latency of 1.9 ms; the other showed a weak ipsilateral facilitation with widely varying latencies (range: 4 - 1 9 ms). Both neurons failed to show any VOR or VCR-related modulation during rotation tests. "Unrelated" RF(S)Ns. In this sample, the proportion of neurons that received monosynaptic input from the SCc was 75.6% (Fig. 5A), that is, lower than among ENNs. Longer latency excitatory responses to SCc were recorded in nine cells. Four of them responded to both sides and were included in Column 3 of Table 1. The asymmetry of tectal input was evaluated on 35 neurons tested from both sides, after excluding the two that did not respond to either side. Among neurons receiving either monosynaptic or indirect SCc input, absence of response or IPSPs from SCi was observed in 26 cells (74.3%) (Table 1, Col-

There is strong evidence suggesting that ENNs contribute to signal transformations between the superior colliculus and motoneurons of extraocular and neck muscles during orienting eyeneck synergies [13,15,16]. We explored the question whether neurons of this particular population receive input from the vestibular system, which would be sufficient to induce spike activity related to compensatory gaze shifts (VOR, VCR). Neurons apparently "unrelated" to orienting were studied with the same method and served as a control sample.

Orienting-Related RF(S)Ns Do Not Participate in Compensator), Vestibular Reflexes One of the tasks of this study was to find out whether or not the firing rate of ENNs is modulated during passive whole-body rotation. The answer to this apparently simple question is not easy to obtain because alert head-fixed cats show a strong tendency to look in the direction of rotation. Such anticompensatory gaze shifts resulting, to a major part, from the summation of quick phases of nystagmus [8,9] are accompanied by an increased neck muscle activity [43] and by the discharges of ENNs [42 and this study[. These anticompensatory patterns are in counterphase with the expected VCR-related modulation [4,5,11], which can therefore be completely masked. Even in the absence of anticompensatory movements, the mean eye position during VOR must be shifted from the midline to the side of the recorded

TECTO-VESTIBULAR CONVERGENCE ON RETICULAR NEURONS

neuron, in order that the level of ipsilateral neck EMG becomes sufficient to reveal the VCR-related modulation. These conditions must also be fulfilled to observe tonic firing of QT-ENNs and the sustained components in the discharges of PS-ENNs. Our results have shown that, even when the above requirements for conclusive tests were met, modulation during passive wholebody rotation could not be detected in any of ENNs. In these experiments, the frequency and velocity ranges of the natural vestibular stimulation have been limited to -< 1 Hz and 50°Is. However, a majority of neurons were tested in the range 10-25°ls because higher velocities led to the restlessness of the animal. A VCR-related modulation of neck EMG with such stimulus velocities has been reported in decerebrate cats, although responses in this range were clearly submaximal [5,11 ]. It cannot be excluded that our parameters of stimulation, albeit sufficient to induce the VCR, were below the threshold for the recruitment of ENNs in the VCR-related activity. In such a case, our conclusion that ENNs do not participate in the VCR must be restricted to the lower velocity range of the natural vestibular stimulation. Sinusoidal polarization of the horizontal ampullary nerve has been reported to induce, together with the VCR, modulation of firing in 15.3% and 9.4% of RSNs projecting, respectively, in the medial and lateral reticulospinal tracts [36, Table 2]. In our sample of neurons "unrelated" to orienting movements, the probability of modulation by horizontal rotations was close to these values (14.7%). Modulation was absent in 8 ENNs for which all the requirements for conclusive tests have been fulfilled. Also the remaining nine ENNs did not fire when threshold eye position and/or EMG level were occasionally reached during the VOR and VCR in the absence of phasic anticompensatory events. These observations suggest that ENNs are a functionally specific group of reticular neurons, different from RSNs, which subserve compensatory reflexes. A definitive proof could not be obtained because of a small number of observations: a comparison of probabilities of modulation between the two samples (ENNs and neurons "unrelated" to orienting) did not indicate a statistically significant difference. Nevertheless, the above suggestion is strongly supported by the study of synaptic connections (see below). it remains an open question whether ENNs may be recruited in the VCR-related activity during rotation in planes other than horizontal, lndeed, it has been reported that 69% of RSNs and 65% of unidentified RFNs are modulated by low frequency oscillations about the longitudinal axis [24]. More recently, a modulation of otolithic origin has been found in 56% of RSNs during multi-axis rotations [6]. In the above studies on decerebrate cats, the neuronal samples differed from that of ENNs either by their more caudal location [24] or by the presence of spontaneous background discharge [6]. Therefore, these demonstrations of the sensitivity to otolithic input in many reticular neurons cannot be taken as predicting the same property for ENNs.

Responses to Electrical SC and V.n. Stimulation and the TectoVestibular Convergence Previous studies suggested several kinds of neurons that could link vestibular afferents to the medial ponto-bulbar RF. Intracellular recordings in cats indicated that vestibular nuclei stimulation elicits postsynaptic potentials in a majority of pontine and bulbar RF neurons, about 50% of EPSPs and IPSPs being monosynaptic [32]. These data are difficult to reconcile with anatomical descriptions of relatively weak direct vestibulo-reticularconnections [7,26,29]. In agreement with anatomy, disynaptic latencies were observed in only 31% and 14% of RF neurons responding, respectively, to ipsi- and contralateral V.n. stimula-

345

tion [34]. The latencies of tri- or polysynaptic responses of RF neurons cover a broad range from about 2 - 3 ms to 20-30 ms [35,36]. Most likely, the responses at shorter latencies could be transmitted by the nonsecondary neurons of the vestibular nuclei [32], neurons of the nucleus prepositus hypoglossi [3,25], and burster-driving neurons [21,31 ]. Of particular interest for this study is the comparison of the overall responsiveness to V.n. stimulation between ENNs and other populations of reticular neurons. In cats under chloralose anesthesia, excitatory responses to either ipsi- or contralateral V.n. were evoked in 61-63% of RSNs and unidentified cells [33,34]. A smaller proportion (31%) has been reported in decerebrate cats [35]. The responsiveness was shown to be higher (71.4%) in a selected sample of 14 RSNs responding to sinusoidal V.n. polarization with phase lags similar to those of EMG modulation in the neck muscles [36]. In the present work on alert cats 71.4% of neurons "unrelated" to orienting movements received an excitatory V.n. input. This fraction is larger than reported in some of the previous studies, because of a purposeful search for neurons responding to either V.n. or tectal stimulation, or both. Although such a bias could have exaggerated the differences between our two samples, it can be safely concluded that ENNs differ from RF(S)Ns "unrelated" to orienting by a virtual absence of excitatory responses to activation of vestibular afferents and, consequently, by a very low probability of the rectovestibular convergence. A great majority of ENNs receive a monosynaptic excitation from the contralateral SC and show an asymmetry of effects from both sides, defined as a combination of contralateral excitation with inhibition or absence of response to SCi. Among RF neurons "unrelated" to orienting movements, the proportion of cells excited monosynaptically from the SCc was smaller than among ENNs, but the statistical significance of the difference was not very high. With respect to the asymmetry of responses, the difference between samples was not significant. Therefore, the identification of ENNs by their discharges during active gaze shifts allows to predict rather well the pattern of response to SC stimulation. On the contrary, the presence of the contralateral excitatory input from the SC, even if it is monosynaptic and unilateral, is not sufficient to conclude that the neuron belongs to the tecto-reticulo-spinal system of orienting. Similarly, excitatory responses to electrical V.n. stimulation do not allow to predict whether or not the neuron will participate in vestibular reflexes induced by sinusoidal rotations in the horizontal plane.

Participation of ENNs in Rotation-Induced Anticompensatory Gaze Shifts As suggested by their discharge patterns and connections revealed by intra-axonal staining [14,16], ENNs could likely control the coordination of eye and head movements during visually triggered orienting. They appear to fulfill the same function during rotation-induced anticompensatory gaze shifts. Anticompensatory eye movements (quick phases of the vestibular nystagmus and deviation of eye position in the direction of rotation) have been interpreted in terms of a refixation of gaze on a "'center of interest" whose orbital eccentricity is proportional to the velocity of rotation, till a saturation at 15° [8]. Eye or head position shifts in the direction of rotation have been defined as a robust reaction of orienting toward a new sector in space, as it occurs, for example, during circular locomotion [27]. Because vestibular input is necessary to trigger orienting movements of this type, also the quick phase and eye position-related firing of ENNs during rotation in the dark must depend, directly or indirectly, on the

346

activation of vestibular afferents. At the same time, it may be controlled by the processes of direction-selective attention.

Origins of ENN activity correlated to anticompensatory gaze shifts: Presumed rhombencephalic inputs. In the rhombencephalon, one of the likely sources of excitatory input to ENNs during quick phases are burster-driving neurons (BDN) [21,31]. They are located in the prepositus nucleus and the adjacent medullary RF and project, after crossing, to the contralateral paramedian pontine RF. Here the locations of ENNs (Fig. 1A) overlap those of excitatory and inhibitory saccade-related burst neurons [38]. During passive sinusoidal rotation, BDNs display a Type II modulation of firing rate and burst activity correlated with quick phases to the contralateral side. Thus, during rightward rotation left BDNs might exert facilitatory effect on right ENNs. A similar effect can be expected from "anticompensatory" neurons of the prepositus nucleus [17]. An assumption that these two types of bulbar neurons could be a major source of excitatory input to ENNs during anticompensatory gaze shifts is difficult to reconcile with the present results. Indeed, ENNs have a very low responsiveness to V.n. stimulation, whereas BDNs are reliably driven at disynaptic latencies [31]. A trisynaptic pathway through BDN-type neurons which, in principle, could link primary vestibular afferents to ENNs appears therefore to be very weak, insofar as can be judged from the responsiveness to the V.n. stimulation. The same applies to other known rhombencephalic vestibulo-reticularconnections, which have been discussed above.

Origins of ENN activity correlated to anticompensatory gaze shifts: Presumed supratentorial inputs. Evidence pointing to a low efficacy of rhombencephalic vestibular input to ENNs is in some contradiction with their consistent activity related to anticompensatory gaze shifts during natural vestibular stimulation. An explanation can be sought in the parallel organization of pathways controlling gaze. If connections of BDNs or "anticompensatory" prepositus neurons to ENNs do exist, their facilitatory effects could represent only a small part of the total synaptic input originating outside the lower brain stem. Indeed, it has been demonstrated that, according to their afferent and efferent connections, BDNs may be a component of a pathway acting in parallel with the crossed projection of the SC to the medial pontine RF [22]. Earlier studies suggested that the SC could contribute to the generation of quick phases or other components of anticompensatory gaze shifts during natural vestibular stimulation. An asymmetry of ocular responses to head velocity steps after unilateral SC lesions [12] can be interpreted as an increased facilitatory action of the intact SC on the contralaterally directed quick phases. The disappearance of eye position shift to the side, opposite to the lesioned SC, has also been documented ([12], Fig. 2). It has been reported that saccade-related burst neurons of the monkey SC discharge before and during the quick phases of caloric nystagmus [44]. In such a case facilitatory effect on ponto-bulbar RFNs could be exerted directly through the crossed tecto-reticulo-spinal pathway. However, experiments in alert cats failed to reveal any activity of identified tectoreticulo-spinal neurons during vestibular nystagmus induced by passive rotation in the horizontal plane [28]. In spite of this negative evidence the SC must be kept in mind as a potential source of facilitatory input to ENNs and other ponto-bulbar neurons discharging in relation to quick phases. Forebrain structures are also expected to participate in the generation of gaze error signal during anticompensatory gaze shifts. This is suggested by recent experiments in the rat that demonstrated that a deficit of selective attention after lesions of the prefrontal cortex is accompanied by the disappearance of eye deviations in the direction of passive rotation [2].

KITAMA, GRANTYN AND BERTHOZ

CONCLUSION: EVIDENCE FOR FUNCTIONAL SPECIFICITY OF ORIENTING-RELATED ENNS These results provide further evidence in favor of the specificity of ENNs, as a population of reticular neurons participating in the control of fast orienting eye-head synergies. Their function consists in the generation of premotor signals during active gaze shifts, and, at low frequencies of natural vestibular stimulation, they do not participate in the gaze-stabilizing components of VOR or VCR. Earlier work in decerebrate cats suggested that a population of RSNs that receives convergent tectal and vestibular inputs could act as a premotor pathway to neck motoneurons, common for both the VCR and orienting head movements [35]. The participation of the same population of neurons in two different types of behavior could represent an interesting example of functional flexibility of the premotor neural network. However, such a duality of function was not observed in this study of orientingrelated ENNs, even though their locations and projections are similar to RSNs, presumably integrating tectal and vestibular inputs [351. Synaptic responses to tectal and vestibular nerve stimulation indicate that ENNs are dominated by excitatory input from the contralateral SC and have extremely low responsiveness to electrical stimulation of vestibular afferents. This pattern of connections clearly distinguishes them from RSNs mediating the VCR in decerebrate cats. The specificity of ENNs is thus reflected not only in the relationship to a particular type of motor behavior but also in the pattern of their afferent connections. Activation of ENNs during quick phases of vestibular nystagmus, anticompensatory shifts of the beating field, and accompanying EMG activity fits well the interpretation of these motor events as components of orienting behavior. Due to the weakness of local, rhombencephalic input from the vestibular afferents, it must be assumed that recruitment of ENNs depends on facilitatory convergence from higher order structures, probably including the superior colliculus. ACKNOWLEDGEMENTS

The authors thank Stephanie Lemarchand, Michel Ehrette, and Michel Loiron for technical assistance. The study was supported by the Human Frontier Study Program and ESPRIT II Basic Research Action N°6615 (MUCOM). The authors also thank Prof. Carlo Terzuolo, the coordinator of the Human Frontier project, for his advice and encouragement.

REFERENCES 1. Alstermark, B.; Pinter, M. J.; Sasaki S. Descending pathways mediating disynapticexcitationof dorsal neck motoneuronesin the cat: Brain stem relay. Neurosci. Res. 15:42-57; 1992. 2. Bahring, R.; Meier, R. K.; Dieringer, N. Unilateral ablation of the frontal eye field of the rat affects the beating field of ocular nystagmus. Exp. Brain Res. 98:391-400; 1994. 3. Baker, R.; Berthoz, A. Is the prepositus hypoglossi nucleus the source of another vestibulo-ocular pathway? Brain Res. 86:121127; 1975. 4. Berthoz, A.; Anderson,J. Frequency analysis of vestibular influence on extensor motoneurons. II. Relationship between neck and forelimb extensors. Brain Res. 34:376-380; 1971. 5. Bilotto, G.; Goldberg, J.; Peterson, B. W.; Wilson, V. J. Dynamic properties of vestibular reflexes in the decerebrate cat. Exp. Brain Res. 47:343-352; 1982. 6. Bolton, P. S.; Goto, T.; Sehor, R. H.; Wilson, V. J.; Yamagata, Y.; Yates, B. J. Response of pontomedullary reticulospinal neurons to vestibular stimuli in vertical planes. Role in vertical vestibulospinal reflexes of the decerebrate cat. J. Neurophysiol,67:639-647; 1992.

TECTO-VESTIBULAR C O N V E R G E N C E ON RETICULAR NEURONS

7. Carleton, S. C.; Carpenter, M. B. Afferent and efferent connections of the medial, inferior and lateral vestibular nuclei in the cat and monkey. Brain Res. 278:29-51; 1983. 8. Chun, K.-S.; Robinson, D. A. A model of quick phase generation in the vestibuloocular reflex. Biol. Cybernetics. 28:209-221; 1978. 9. Donaghy, M. The cat's vestibulo-ocular reflex. J. Physiol, (London) 300:337-351; 1980. 10. Eccles, J. C.; Nicoll, R. A.; Schwarz, D. W. F.; Tfibor~ovfi, H.; Willey, T. J. Topographic studies on medial reticular nucleus. J. Neurophysiol. 39:109-118; 1976. 11. Ezure, K.; Sasaki, S. Frequency-response analysis of vestibular-induced neck reflex in cat. 1. Characteristics of neural transmission from horizontal semicircular canal to neck motoneurons. J. Neurophysiol. 41:445-471; 1978. 12. Flandrin, J. M.; Jeannerod, M. Effects of unilateral superior colliculus ablation on oculomotor and vestibulo-ocular responses in the cat. Exp. Brain Res. 42:73-80; 1981. 13. Grantyn, A.; Berthoz, A. Reticulo-spinal neurons participating in the control of synergic eye and head movements during orienting in the cat. I. Behavioral properties. Exp. Brain Res. 66:339-354; 1987. 14. Grantyn, A.; Hardy, O.; Olivier, E.; Gourdon, A. Relationship between task-related discharge patterns and axonal morphology of brainstem projection neurons involved in orienting eye and head movements. In: Shimazu, H.; Shinoda, Y., eds. Vestibular and brain stem control of eye, head and body movements. Tokyo: Japan Sci Soc Press; Basel: S. Karger; 1992:255-273. 15. Grantyn, A.; Olivier, E.; Kitama, T. Tracing premotor brain stem networks of orienting movements. Current Opinion Neurobiol. 3:973-981 ; 1993. 16. Grantyn, A.; Ong-Meang Jacques, V.; Berthoz A. Reticulo-spinal neurons participating in the control of synergic eye and head movements during orienting in the cat. 1I. Morphological properties as revealed by intra-axonal injections of horseradish peroxidase. Exp. Brain Res. 66:355-377; 1987. 17. Hardy, O.; Corvisier, J. Firing behavior of anticompensatory neurones in the prepositus hypoglossi nucleus of alert cat. Neurosci. Lett. 139:178-182; 1992. 18. Hepp, K.; Henn, V.; Vilis, T.; Cohen, B. Brainstem regions related to saccade generation. In: Wurtz, R. H.; Goldberg, M. E., eds., The neurobiology of saccadic eye movements. Reviews of oculomotor research, Vol. 3. Amsterdam, New York, Oxford: Elsevier Biomedical Press; 1989:105-212. 19. lwamoto, Y.; Sasaki, S.; Suzuki, I. Input-output organization of reticulospinal neurones, with special reference to connexions with dorsal neck motoneurones in the cat. Exp. Brain Res. 80:260-276; 1990. 20. Kitama, T.; Grantyn, A.; Berthoz, A. Synaptic and functional convergence of tectal and vestibular inputs on "eye-neck" reticular formation neurons. European J. Neurosci. Suppl 6:272; 1993. 21. Kitama, T.; Ohki, Y.; Shimazu, H.; Yoshida, K. Physiological and morphological properties of brain stem neurons related to vestibular and saccadic eye movements in the cat. In: Hwang, J. C.; Daunton, N. G.; Wilson, V., eds. Basic and applied aspects of vestibular function. Hong Kong: Hong Kong University Press; 1988:45-53. 22. Kitama, T.; Shimazu, H.; Tanaka, M.; Yoshida, K. Vestibular and visual interaction in generation of rapid eye movements. Ann. N.Y. Acad. Sci. 656:396-407; 1992. 23. Lambertz, M.; Schulz, G.; Langhorst P. Reticular formation of the lower brainstem. A common system for cardio-respiratory and somatomotor functions. Considerations aided by computer simulations. J. Autonom. Nerv. Syst. 12:63-75; 1985. 24. Manzoni, D.; Pompeiano, O.; Stampaccia, G.; Srivastava, U. C. Responses of medullary reticulospinal neurons to sinusoidal stimulation of labyrinth receptors in the decerebrate cat. J. Neurophysiol. 50:1059-1080; 1983.

347

25. McCrea, R. A.; Baker, R. Anatomical connections of the nucleus prepositus in the cat. J. Comp. Neurol. 237:377-407; 1985. 26. McCrea, R. A.; Strassman, A.; May, E.; Highstein, S. M. Anatomical and physiological characteristics of vestibular neurons mediating the horizontal vestibulo-ocular reflex of the squirrel monkey. J. Comp. Neurol. 264:547-570; 1987. 27. Meier, R. K.; Dieringer, N. The role of compensatory eye and head movements in the rat for image stabilization and gaze orientation. Exp. Brain Res. 96:54-64; 1993. 28. Munoz, D, P.; Guitton, D. Tectospinal neurons in the cat have discharges coding gaze position error. Brain Res. 341 : 184-188; 1985. 29. Ohgaki, T.; Curthoys, I. S.; Markham, C. H. Morphology of physiologically identified second-order vestibular neurons in cat, with intracellularly injected HRP. J. Comp. Neurol. 276:387-411; 1988. 30. Ohki, Y.; Shimazu, H.; Suzuki, I. Functional organization of the horizontal vestibulo-oculomotor system in the cat brain stem. In: Hwang, J. C.; Daunton, N. G.; Wilson, V. J., eds. Basic and applied aspects of vestibular function. Hong Kong: Hong Kong University Press; 1988:35-43. 31. Ohki, Y.; Shimazu, H.; Suzuki, 1. Excitatory input to burst neurons from the labyrinth and its mediating pathway in the cat: Location and functional characteristics of burster-driving neurons. Exp. Brain Res. 72:457-472; 1988. 32. Peterson, B. W.; Abzug, C. Properties of projections from vestibular nuclei to medial reticular formation in the cat. J. Neurophysiol. 38:1421-1435; 1975. 33. Peterson, B. W.; Felpel, L. P. Excitation and inhibition of reticulospinal neurons by vestibular, cortical and cutaneous stimulation. Brain Res. 27:373-376; 1971. 34. Peterson, B. W.; Filion, M.; Felpel, L. P.; Abzug, C. Responses of medial reticular neurons to stimulation of the vestibular nerve. Exp. Brain Res. 22:335-350; 1975. 35. Peterson, B. W.; Fukushima, K. The reticulo-spinal system and its role in generating vestibular and visuomotor reflexes. In: Sj61und, B.; Bj6rklund, A., eds. Brain stem control of spinal mechanisms. Amsterdam, New York, Oxford: Elsevier Biomedical Press; 1982:225 - 25 I. 36. Peterson, B. W.; Fukushima, K.; Hirai, N.; Schor, R. H.; Wilson, V. J. Responses of vestibulospinal and reticulospinal neurons to sinusoidal vestibular stimulation. J. Neurophysiol. 43:1236-1250; 1980. 37. Peterson, B. W.; Maunz, R. A.; Pitts, N. G.; Mackel, R. G. Patterns of projection and branching of reticulospinal neurons. Exp. Brain Res. 23:333-351; 1975. 38. Sasaki, S.; Shimazu, H. Reticulovestibular organization participating in generation of horizontal fast eye movement. In: Cohen, B., ed. Vestibular and oculomotor physiology. New York: The New York Academy of Sciences; 1981 : 130-143. 39. Scheibel, M. E.; Scbeibel, A. B. Structural substrates for integrative patterns in the brain stem reticular core. In: Jasper, H. H.; Proctor, L. D.; Knighton, R. S.; Noshay, R. T.; Costello, R. T., eds. Reticular tbrmation of the brain. Boston, Toronto: Little, Brown and Co.; 1958:31-55. 40. Scbeibel, M. E.; Scbeibel, A. B.; Mollica, A.; Moruzzi, G. Convergence and interaction of afferent impulses on single units of reticular formation. J. Neurophysiol. 18:309-331; 1955. 41. Siegel, J. M. Behavioral functions of the reticular formation. Brain Res. Rev. 1:69-105; 1979. 42. Vidal, P. P.; Corvisier, J.; Berthoz, A. Eye and neck motor signals in periabducens reticular neurons of the alert cat. Exp. Brain Res. 53:16-28; 1983. 43. Vidal, P. P.; Roucoux, A.; Bertboz, A. Horizontal eye position-related activity in neck muscles of the alert cat. Exp. Brain Res. 46:448-453; 1982. 44. Wurtz, R. H.; Goldberg, M. E. Activity of superior colliculus in behaving monkey. III. Cells discharging before eye movements. J. Neurophysiol. 35:575-586; 1972.