- Email: [email protected]
Iso-frequency curves of oculomotor neurons in the rhesus monkey
Vision Res. Vol. 25. No. 4. pp. 493-499. 1985 All rights reserved
Pnntcd I” Great Britain.
ISO-FREQUENCY
Copyright C
CURVES OF OCUL~MOTOR IN THE RHESUS MONKEY
0042-6989!85 53.00 + 0.00 1985 Pergamon Press Ltd
NEURONS
K. HEPP’ and V. HENNA ‘Physics Department, ETH, 8093 Ziirich, Switzerland and *Neurology Department, University, 8091 Ziirich, Switzerland (Received 20 July 1984; in
revisedform
7 Nocember
1984)
Abstract-Static tiring frequencies have been determined in extraocular motoneuronal discharge patterns for different eye positions within + 30deg around the primary position. From these data iso-frequency curves were plotted stating all possible eye positions for a given tiring rate. Such curves have been constructed for the lateral, medial, and inferior recti, the superior oblique and for the upward pulling muscles (without distinguishing superior rectus and inferior oblique). Fixation of eye position always involved natural synergistic action of all muscles. The iso-frequency curves of individual motoneurons are a family of almost parallel curves with mainly horizontal or vertical gradients. Especially for the superior oblique, the innervation gradients depend strongly on eye position, Motoneurons subserving the same muscle can have different innervation gradients at the same eye position. Motoneuron
Monkey, rhesus
Kinematics
Innervation
INTRODUaION The kinematics of the eyes was a subject of great interest already to Helmholtz (1866) and Hering (1868) and their pupils. Most of their works were either geometrical calculations, or psychophysical experiments. Hering calculated the direction and amount of eye rotation if single muscles were contracted when the eye was initially in the primary position. Later these calculations were extended to include arbitrary eye positions and considered synergistic action of all extraocular muscles (Westheimer, 1957; Boeder, 1962). Only recently it became possible to perform single cell recordings from motoneurons in the alert monkey (Fuchs and Luschei, 1970; Robinson, 1970; Schiller, 1970; Keller and Robinson, 1972; Henn and Cohen, 1972, 1973; Eckmiller, 1974). Results indicate that for conjugate eye movements the static innervation of every individual eye muscle is a unique function of the direction of gaze, as expected from Listing’s law and from Sherrington’s law of antagonistic innervation (Westheimer, 1981). In the human, the static tension of eye muscles as a function of length and innervation has been measured by Collins et al. (1975). Robinson (1975) calculated the static innervation curves of extraocular muscles for any position in the field of gaze. To reduce the degrees of freedom for these calculations he paired the horizontal recti, the vertical recti and the oblique muscles into three pairs of antagonists. For every muscle he thus obtained a family of curves which were almost straight lines in the range of 30” around the primary position. For larger deviations up to 40” eccentricity some interesting details 493
Dculomotor neurons
emerged which had not been predicted by considering the action of any one muscle in isolation. In view of the importance of the consideration of synergistic muscle action on the eyes for both an understanding of the control of eye movements, as well as a possible application like in strabismus surgery, we have measured the static innervations of individual motoneurons in the alert Rhesus monkey and have analyzed them in two dimensions as a function of horizontal and vertical gaze deviation. Our results (short communications Henn and Hepp, 1981, 1984) show a good correspondence to Robinson’s prediction with some further details.
METHODS Two Rhesus monkeys (Macaca mulatta) were trained to fixate a moving light spot while recordings were made from axons of extraocular motoneurons.
The monkey was positioned at a distance of 50 cm from a tangent screen orthogonal to the monkey’s gaze direction in primary position which was taken to be the optic axis. A small spot of laser light was rear-projected onto the screen and could be deflected horizontally and vertically by two servo-controlled rotating mirrors. T&tli?g
The monkey had to discriminate the dimming of the laser light which occurred at random intervals, and to release a hand-held bar within OSsec for a water reward (Wurtz, 1969). The amount of gained
liquid was controlled e\ery day and supplemented if necessary. The light intensity and dimming step was
set to the minimum at which the animal was able or willing to perform the task. We assume that the monkey foveated and pursued the light spot. The electro-oculogram is linear only over a restricted range of eye positions. Although the eye position traces were very reproducible. during the fixation task, we have no independent measurement of how exact foveation was.
Surger?
After premeditation with 0.5 mg atropine i.m. and pentobarbital (i.p., 30mgjkg body wt) the monkey was intubated and was given a mixture of N,O-0, to breath through a closed system; hatothane was supplemented when necessary. Rectal temperature and ECG were continuously monitored. Over a trephine hole in the skull a plug was stereotaxically implanted which during the experiments supported a micromanipulator. DC silver-silverchloride electrodes were implanted in the bony orbit to monitor horizontal and vertical eye position. In addition, bolts were implanted in dentai cement attached to the skuli in order to hold the head rigid.
Protocol
The animal was kept in its cage except for the of the experimental session--each lasting about 3 hrjday-on average 4 times/week. For the experiment, the animal was seated in a primate chair. Horizontal and vertical eye position was monitored. Units were recorded with varnished tungsten electrodes advanced according to stereotaxic coordinates in the 3rd, 4th and 6th nerve rootlets. Horizontal motoneuronal discharges were recorded while a laser light moved in a vertical direction in a range of $- 30 deg at 0.04 Hz at fixed horizontal eccentricities changed in steps of 5 or 1Odeg. For testing vertical motoneuronal activity the laser light was moved horizontally +30deg at 0.04 Hz at different vertical eccentricities. Single unit activity, horizontal and vertical eye position, light spot position and its brightness, and a digital time code were stored on an FM tape machine. interruption
‘~otunei~ron
iiienfification
By their location in the nerve rootlets and their activity increase with eye movement in a preferred direction, all motoneurons except those with an on-direction could be identified unupward ambiguously. For medial and lateral rectus motoneurons this is obvious. Axons of inferior reetus motoneurons are the only axons in third nerve which convey signals with a downward on-direction and are widely separated from axons in the trochlear rootlets. Trochlear fibers were recorded distal to their crossing
l-3 mm lateral from midline. Ai .!n :nilqw~~~~r!r. check for the stereotaw localization :w ai\+~> -)t,ldt:,i an abducting component in their on-ciireciig:I
All stored data together with instantaneow rno.. toneuron frequency was written out on six-channel rectilinear oscillograph paper. Firing rates were dstermined over 100msec intervais at eye positioni varying in steps of 10 deg. Measurements were averaged over movements in both directions. As the lipht spot had a maximal velocity of 7.5 ,sec across the center, and as it moved orthogonal to the ondirection of the respective motoneuron, any velocity related component was small and was eiiminate~~ together with a possible hysteresis effect (Eckmiller. 1974) by averaging over opposite movement directions. Average firing rates together with eye position were fed into a small computer for statistical calculations and interpolation. Figures and numbers are based on these results. We have taken as coordinates of gaze Cartesian coordinates on the tangent screen, where the horizontal (h) and vertical (c) components are expressed as degrees of eye deviation from primary position along the horizontal or vertical axis. This choice is more symmetrical than the poi;n coordinates (a&) of the gaze direction used by Robinson (1975). However, our horizontal coordinate /z is identical to the horizontal polar angle 0, while the vertical polar angle 4 is related to (h,~.) by $(h,t~) = tan-‘[tan(~)~(i + tan(h)?)“]). One hits # = t‘ along the coordinate axes j/r = 0;. or {r: = Oi, I#J [ < ju 1 with equality up to second order corrections in h or c. For given h, 4(l7,c) differs (in our range of measurements) maximally from 13 for /u 1 = 30”, with 4(lO”, 30’) = 29,6;, 4(20’, 30”) = 28.5’ and &(30’, 30’) = 26.6 . Hence by ;I slight vertical stretching, which increases with lh /, the theoretical curves in Robinson (1975) can be brought into our format. However this small change is within the resolution of Robinson’s Fig. 5 and within our experimental errors. RESULTS
Recordings of activity from more than 150 axons of extraocular motoneurons were sufficiently stable to determine at least one iso-frequency curve. in 60 recordings the fulf range of horizontal and vertical eye deviations of f30deg around primary position could be covered in steps of IO deg. The sample of 60 axonal recordings consisted of 6 lateral rectus, 12 medial rectus, I3 superior oblique, 15 inferior rectus and 14 motoneurons with upward on-directions. The sample is biased towards predominantly tonic neurons with low thresholds (Henn and Cohen, 1972). We shall group our findings in Figs 2-5 according to different eye muscles, which will be related to the right eye only (after a mirror reflection of data from the other eye) as in the paper of Robinson (3975).
Figure I(A) shows the sM2 inncrvslllian surface of a typical tow threshold left medial rkxtus motoTWKOILFor fixed horizontal eye deviation values of* - 10:. 0’ , . . -I-30” the measured average firing rates, f(+r,r), at vertica! eye deviations of -3Q‘, -2o”, . ~_ -i-30” are represented by b?ack dots, whose distance from the base point in the (h,r)-plane is proportionat to _!“(fh,r:)according to the scale at the right hand corner. The static innervation surface of this unit is then obtained by piece-wise tinear connection ofthese points. In rhis perspective drawing one is looking from the left side onto the rising slope, which for &his neuron is slightly steeper in the lower hemifidd. For the same neuron, Fig. f(B) shows the iso-frequency curves (IFC) for average firing ratesS= 50, 100, 150 1
and 200 Hz, computed from the data represented in Fig. I(A) by finear inteqmlation. if OZ-E defines as innervation gradient the vector field orthogonal to the isawfrequency curves and with length proportional to the slope of the innervation surface at every point (h,~), then the innervat& gradient is horizontal around k = 0 and points about 5 deg downward in the lower right hemifield. Clearly, the iso-frequency curves give a compact 2-dimensional representation of the inntmation surface, fn particuIar, the fevel of 50 Hz is related to the threshold of the unit (at about 2S-30 Hz), while/= 150 Hz is representative for the middle range of the activity (see Henn and Hepp 1984, Fig. 2 for another MR unit which remains active over the range of -t-30 deg around primary position).
Fig. 1. (A) innervation surface sf a kft medial. recks motoneuron. Far fixed ~&es of horizontal c3ye deviation 81 - 10, Q + ict, i-20, and +30delg the static frequeencywas me&Sired during verticaf stow pursuit movements between f 30”dq (up) and -30” dq (down). The average frequencies during four runs were plotted. 3y connecting these points one obtains the innervation surface. (B) iso-frequency curves of the stlme motoneuron for average frequencies of 50, 100, 150 and 200 Hz, computed from the data of Fig. 1(A). By lineear iateq&tion between data points w&in &20” on the IFC closest to the primary
position we have determined the average on-directian
of a nerrron.
K. HEPP and V. HEW
196
A
150 Hr
I
0
-30
adduction
abduction
Fig. 2. Iso-frequency curves of 12medial rectus motoneurons, (A), at 150 Hz, and (Bf. at 50 Hz. The curves are normalized to the right eye with on-direction to the left. In Fig. 2(A) we have represented the iso-frequency 150 Hz for the full sample of 12 medial rectus recordings all drawn as if the on-direction were to the left. One sees that in this middle firing range the innervation gradients are not uniformly horizontal, but that deviations stay within k 5 deg. In Fig. 2(B), the IFC of 50 Hz shows the range of thresholds in this sample which is biased towards low threshold motoneurons. The average on-direction E of a motocurves of
neuron is chosen as the direction of the innervation gradient averaged over the IFC closest to primary position in a vertical range of & 20’. For example, in Fig. 1(B), LXis - 3.2 deg, i.e. slightly downwards relative to the horizontal axis. In our sample of 12 MR axons, cfranges between - 3.2 deg and + 6.8 deg with average Z = 0.8 + 2.9 deg (1 SD). As expected from the neuronal organization of horizontal gaze, the iso-frequency curves of all 6 lateral rectus axons
A
-.20
-3.0
-2.0
-1.0
0
I.0
2.0
h ldwl
3.0
adductioo
Fig. 3. &o-frequency curves of 13 superior oblique motoneurons at 1SOHz (A) and 50 HZ (B), related to the tight eye.
A
20
3,o
h (dwl
Fig. 4. Is&frequency curves af 15 infer&x rctus motoneurons at 150 Hz {A) and 50 l-5~(8X related to :he right ye.
from the Vith nerve (and also of all 5 internuclear fibers recorded in the MU) were found to be indistinguishable from those of the MR axc~~s.
Figure 3 shows the EC at SO and I50 Hz for-our sampie of 13 SO fibers fse+zaiso Menn and Happ, 1984, Fig. 3 for a typical example in the format as in Fig. I). In the medial hemifield the innervation directions are downward, and they are downward and outward in the lateral hemitield. The average
on-directions of truchlear motoneurons strongly fluctuate: 1 varies between t .2 and 21.2 deg downwards lateral, with 5 = 9.1 + 5.3 deg. Some SO units have almost purely downward innervation gradients and IFC which are almost straight lines. Within the same tract through tbe trochlear nerve one can find units with such differences in their on-dirtxtjons, However, &heSO ~o~uiat~on average has an almost downward innervation gradient in the media{ and a downward and uutward direction in the lateral hemifield, even for the two high threshold units in our sample.
A
Fig. 5. Iso-frequency curves of 14 mNmxmns
with upward on-directions at I SOHz (A) and 50 Hz (B), related to the right eye.
In contrast to superior oblique and similar to the horizontal recti the innervation directions of inferior rrctus motonrurons show much smaller fluctuations: in our sample of I5 recordings x varied between 0.7 and 13deg downwards and medial with !i = 4.4 k 2.9 deg. The IFC at 50 and I50 Hz in Fig. 4 show in the lateral hemifield a downward innervation gradient which is slightly medial in the medial hemifield. Although our sample does not contain any high threshold unit. on2 sees clearly that the innervation gradients of this muscle is the same at high and low levels of innervation and that the IFC are straight or convex downwards,
They have not been unambiguously separated into superior rectus and inferior oblique units. In our sample of 14 complete axonal UM recordings (with 6 high threshold units) the IFC in Fig. 5 are straight lines for high innervation values. while for low tiring rates the innervation directions are slightly medial in the medial and lateral in the lateral hemifield. This behavior is different from the downwards-directed motoneurons. However, the sample is obviously rather inhomogeneous. In particular, there is one unit which shows a SO firing pattern inverted at the horizontal axis. The average on-directions r near primary position varied between - 5.7 deg (medial) and 14.6deg (lateral) upwards, with d = 1.7 + 4.6 deg. Again the IFC, except for the strongly curved unit, are parallel. DISCUSSION The static innervation of individual motoneurons in the monkey give a rather consistent picture. The mere existence of iso-frequency curves with small fluctuations over the range of 30 deg around primary position supports the laws of Listing and Sherrington. The innervation gradients for the horizontal
recti are horizontal, for superior oblique downward in the medial and downward lateral in the lateral hemifield, and for inferior rectus they are downward with a small medial component, slightly larger in the medial hemifield. This corresponds to predictions about pulling directions based on the geometry of eye muscles. For the upward motoneurons the innervation gradients are upward and often have a medial component in the medial and a lateral component in the lateral hemifield. Only one unit had an isofrequency curve similar to those of trochlear fibers (reflected at the horizontal axis). For every motoneuron all iso-frequency curves are remarkably parallel for small and large values of innervation, and they are mostly convex from lower to higher values of innervation. We consistently found that iso-frequency curves for high and low innervation are parallel within f 30 deg around primary position. Measuring on2 complete IFC during low activity then allows one to predict the shape of
IFCs at higher ~it;vtty. This I\ ;L p~~tc’nt~,rlai! .,!I:[~:;:’ for slectromyographic recording !n h.~~~n.~ :\h<:tb for technical reasons the rssolution 10r ~nple un~[> 1. good at IOU- innervat:Jn levels clnly. 0~ restlit; suggest that such recordings are sufficisnt to pred:ct unit behavior at highrr innervation levels. Within rhe populations of horizontal recti and of the infrr~or rectus moton2urons. the arsrage on-direction at primary position varies little. and about half as much as within the population of superior oblique neurons, That there are differences in on-directions of individual neurons of the same muscle pool could be statistical fluctuations. changing alertness of the animal. or. as we believe. reflect a different orientation of the respective muscle fiber between its tendon insertions As muscle insertions on the globe are broad, one would expect different innervations for individual fibres to obtain the same tension over the cro’;ssection of a muscle. Our experimental tindings in the Rhesus monkey basically agree with the theoretical predictions by Robinson (1975) for man. However. there art‘ also points of disagreement. In Robinson’s calculation medial rectus motoneurons have not the same innervation as lateral rectus neurons and the convexity of the iso-frequency curves is opposite to our data. The measured iso-frequency curves for trochlear motoneurons remain more parallel over the range 01 + 30 deg than predicted by Robinson. These discrepancies could be a species difference, could be due to the theoretical assumption of pairing the vertical recti and obliques. or could be a bias of the smallness of our sample. Lesion studies indicate that horizontal and vertical saccadic eye movements are organized in iwo Cartesian directions (Bender and Shanzer, 1964: reviewed in Henn et al., 1984). We propose that a graded, eye-position dependent innervation to all I :! eye muscles in accordance to Listing’s law is accomThe function of this plished by an “integrator”. integrator is to convert the velocity related activity in premotor units to the sustained activity found in motoneurons. We have to admit that details of its neuronal realization and localization has remained an enigma. We have not resolved the question of how to pair the vertical muscles in antagonists or in yoked partners. The anatomy and electrophysiology of the disynaptic vestibulo-ocular reflex suggest an antagonistic pairing of the superior rectus (superior oblique) with the inferior rectus (inferior oblique) (Precht. 1978). This is the basic scheme which Robinson (1975) has applied for his calculations. In the cat (Uchino, 1980) and therefore probably also in the monkey, the disynaptic connection from one canal involves more than only one antagonistic pair (reviewed in Ito. 1984). The anterior canal. e.g. excites both superior rectus muscles and the inferior oblique of the ipsilateral eye. and inhibition is similarly conveyed to both the inferior rectus and superior oblique.
Motoneuron
iso-frequency curves
Our data show clearly that the inferior rectus of one eye and the superior oblique muscle of the other eye do not receive the same innervation during conjugate downward gaze. Therefore these two muscles are not yoked together in the sense of Tiering. Our data also indicated that the upward motoneurons do not fall into two populations with innervations. which are the mirror images of the inferior rectus and superior oblique motoneurons. Therefore we favor the hypothesis that there is no rigid pairing and yoking of vertical motoneurons which predicts conjugate eye movements for tertiary eye positions. In our experimental setup we were unprepared to encounter such complexities. Future experiments should therefore aim at a more precise measuring of motoneuron activity in relation to eye positions and cyclorotation, during slow and rapid eye movements, and consider different inputs from the visual and vestibular systems. ~e~no~~~e~ge~~ffr~-We are grateful to Dr H. Reisinc for helpful comments and to MS V. Isoviita for technical assistance. K. H. would like to thank Professor G. Baurngartner for his kind hospitality.
REFERENCES Bender IM. B. and Shanzer S. (1964) GcuIomotor pathways defined by electric st~muiation and lesions in the brainstem of the monkey. In 7% Ocutomoror System (Edited by Bender M. B.) pp. 81-140. Harper % Row, New York. Boeder P. (1962) Co-operative action of extra-ocular muscles. Br. J. Ophthaf. 46, 397-403. Collins C. C.. G’Meara D. and Scott A. B. (1975) Muscle tension during unrestrained human eye movements. J. Physiot. 245, 354-369.
Eckmiller R. (1974) Hysteresis in the static characteristics of eye position coded neurons in the alert monkey. PfIiigers Arch. ges. Ph_ysiof. 350, 249-258.
Fuchs A. F. and Luschei E. S. (1970) Firing patterns of abducens neurons of alert monkeys in relationship to horizontal eye movement. J. Seurophysiot. 33, 382-392.
499
Helrnholtz H. v. ( 1856-i 866) Handbuch der physioiogischen Optik. Voss, Leipzig. Henn V. and Cohen B. (1972) Eye muscle motor neurons with different functional characteristics. Brain Rex 45, 56 i-568.
Henn V. Cohen B. (1973) Quantitative analysis of activity in eye muscle motoneurons during saccadic eye movements and fixation. J. Keurophysiol. 36, 115-126. Henn V. and Hepp K. (198I ) Two-dimensional analysis of eye movements in oculomotor and premotor structures. In Progress in Qculomotor Research (Edited by Fuchs A. F. and Becker W.). pp. 81-88. Elsevieri’North-Holland. Henn V. and Hepp K. (1984) Motoneuron activity around the primary position and in extreme positions of gaze. In Trans. Ftyth htt. ~rr~lopt~r Congr. (Edited by Ravault A. P. and Lenk M.), pp. 233-238. Lips, Lyon. Henn V., Lang W.. Hepp K. and Reisine H. ( 1984)Experimental gaze palsies in monkeys and their relation to human pathology. Bruin 107, 619-636. Hering E. (1868) Die Lehre vorn binocularen Sehen. Eng&nann, Leipzig. English translation: The Theory of Binocular Vision. Plenum Press. New York (1977). Ito M. (1984) The Cerebellum and Neural Control. Raven Press, New York. Keller E. L. and Robinson D. A. (1972) Abducens unit behavior in the monkey during vergence movements. Vision Res.
12, 369-382.
Precht W. (1978) Neuronat Operations in the Yestibular System. Studies in Brain Function. Vol 2. Springer, Berlin. Robinson D. A. (1970) Oculomotor unit behavior in the monkey. J. Neurophysiol. 33, 393-104. Robinson D. A. (1975) A auantitative analvsis of extraocular muscle cooperation and squint. invest. Ophthal. sisual Sci. 14, 801-825.
Schiiler P. H. (I 970) The discharge characteristics of single units in the oculomotor and abducens nuclei of the unanesthetized monkey. Expf Bruin Res. IO, 347-362. Uchino Y., Suzuki S. and Watanabe S. (1980) Vertical semicircular canal inputs to cat extraocular motoneurons. Expt Brain Res. 41, 45-53.
Westheimer G. (1957) Kinematics of the eye. J. opr. Sot. Am. 47, 967-974.
Westheimer G. (1981) Donders’. Listing’s, and Hering’s Laws and Their Implications. In Models of Ocuiomotor Behavior and Control (Edited by Zuber B. L.), pp. 149-159. CRC Press, Boca Raton, Florida. Wurtz R. H. (1969) Visual receptive fields of striate cortex neurons in awake rnonkevs. 32. 727-742. , J. NeuroohwioL Li
Pnntcd I” Great Britain.
ISO-FREQUENCY
Copyright C
CURVES OF OCUL~MOTOR IN THE RHESUS MONKEY
0042-6989!85 53.00 + 0.00 1985 Pergamon Press Ltd
NEURONS
K. HEPP’ and V. HENNA ‘Physics Department, ETH, 8093 Ziirich, Switzerland and *Neurology Department, University, 8091 Ziirich, Switzerland (Received 20 July 1984; in
revisedform
7 Nocember
1984)
Abstract-Static tiring frequencies have been determined in extraocular motoneuronal discharge patterns for different eye positions within + 30deg around the primary position. From these data iso-frequency curves were plotted stating all possible eye positions for a given tiring rate. Such curves have been constructed for the lateral, medial, and inferior recti, the superior oblique and for the upward pulling muscles (without distinguishing superior rectus and inferior oblique). Fixation of eye position always involved natural synergistic action of all muscles. The iso-frequency curves of individual motoneurons are a family of almost parallel curves with mainly horizontal or vertical gradients. Especially for the superior oblique, the innervation gradients depend strongly on eye position, Motoneurons subserving the same muscle can have different innervation gradients at the same eye position. Motoneuron
Monkey, rhesus
Kinematics
Innervation
INTRODUaION The kinematics of the eyes was a subject of great interest already to Helmholtz (1866) and Hering (1868) and their pupils. Most of their works were either geometrical calculations, or psychophysical experiments. Hering calculated the direction and amount of eye rotation if single muscles were contracted when the eye was initially in the primary position. Later these calculations were extended to include arbitrary eye positions and considered synergistic action of all extraocular muscles (Westheimer, 1957; Boeder, 1962). Only recently it became possible to perform single cell recordings from motoneurons in the alert monkey (Fuchs and Luschei, 1970; Robinson, 1970; Schiller, 1970; Keller and Robinson, 1972; Henn and Cohen, 1972, 1973; Eckmiller, 1974). Results indicate that for conjugate eye movements the static innervation of every individual eye muscle is a unique function of the direction of gaze, as expected from Listing’s law and from Sherrington’s law of antagonistic innervation (Westheimer, 1981). In the human, the static tension of eye muscles as a function of length and innervation has been measured by Collins et al. (1975). Robinson (1975) calculated the static innervation curves of extraocular muscles for any position in the field of gaze. To reduce the degrees of freedom for these calculations he paired the horizontal recti, the vertical recti and the oblique muscles into three pairs of antagonists. For every muscle he thus obtained a family of curves which were almost straight lines in the range of 30” around the primary position. For larger deviations up to 40” eccentricity some interesting details 493
Dculomotor neurons
emerged which had not been predicted by considering the action of any one muscle in isolation. In view of the importance of the consideration of synergistic muscle action on the eyes for both an understanding of the control of eye movements, as well as a possible application like in strabismus surgery, we have measured the static innervations of individual motoneurons in the alert Rhesus monkey and have analyzed them in two dimensions as a function of horizontal and vertical gaze deviation. Our results (short communications Henn and Hepp, 1981, 1984) show a good correspondence to Robinson’s prediction with some further details.
METHODS Two Rhesus monkeys (Macaca mulatta) were trained to fixate a moving light spot while recordings were made from axons of extraocular motoneurons.
The monkey was positioned at a distance of 50 cm from a tangent screen orthogonal to the monkey’s gaze direction in primary position which was taken to be the optic axis. A small spot of laser light was rear-projected onto the screen and could be deflected horizontally and vertically by two servo-controlled rotating mirrors. T&tli?g
The monkey had to discriminate the dimming of the laser light which occurred at random intervals, and to release a hand-held bar within OSsec for a water reward (Wurtz, 1969). The amount of gained
liquid was controlled e\ery day and supplemented if necessary. The light intensity and dimming step was
set to the minimum at which the animal was able or willing to perform the task. We assume that the monkey foveated and pursued the light spot. The electro-oculogram is linear only over a restricted range of eye positions. Although the eye position traces were very reproducible. during the fixation task, we have no independent measurement of how exact foveation was.
Surger?
After premeditation with 0.5 mg atropine i.m. and pentobarbital (i.p., 30mgjkg body wt) the monkey was intubated and was given a mixture of N,O-0, to breath through a closed system; hatothane was supplemented when necessary. Rectal temperature and ECG were continuously monitored. Over a trephine hole in the skull a plug was stereotaxically implanted which during the experiments supported a micromanipulator. DC silver-silverchloride electrodes were implanted in the bony orbit to monitor horizontal and vertical eye position. In addition, bolts were implanted in dentai cement attached to the skuli in order to hold the head rigid.
Protocol
The animal was kept in its cage except for the of the experimental session--each lasting about 3 hrjday-on average 4 times/week. For the experiment, the animal was seated in a primate chair. Horizontal and vertical eye position was monitored. Units were recorded with varnished tungsten electrodes advanced according to stereotaxic coordinates in the 3rd, 4th and 6th nerve rootlets. Horizontal motoneuronal discharges were recorded while a laser light moved in a vertical direction in a range of $- 30 deg at 0.04 Hz at fixed horizontal eccentricities changed in steps of 5 or 1Odeg. For testing vertical motoneuronal activity the laser light was moved horizontally +30deg at 0.04 Hz at different vertical eccentricities. Single unit activity, horizontal and vertical eye position, light spot position and its brightness, and a digital time code were stored on an FM tape machine. interruption
‘~otunei~ron
iiienfification
By their location in the nerve rootlets and their activity increase with eye movement in a preferred direction, all motoneurons except those with an on-direction could be identified unupward ambiguously. For medial and lateral rectus motoneurons this is obvious. Axons of inferior reetus motoneurons are the only axons in third nerve which convey signals with a downward on-direction and are widely separated from axons in the trochlear rootlets. Trochlear fibers were recorded distal to their crossing
l-3 mm lateral from midline. Ai .!n :nilqw~~~~r!r. check for the stereotaw localization :w ai\+~> -)t,ldt:,i an abducting component in their on-ciireciig:I
All stored data together with instantaneow rno.. toneuron frequency was written out on six-channel rectilinear oscillograph paper. Firing rates were dstermined over 100msec intervais at eye positioni varying in steps of 10 deg. Measurements were averaged over movements in both directions. As the lipht spot had a maximal velocity of 7.5 ,sec across the center, and as it moved orthogonal to the ondirection of the respective motoneuron, any velocity related component was small and was eiiminate~~ together with a possible hysteresis effect (Eckmiller. 1974) by averaging over opposite movement directions. Average firing rates together with eye position were fed into a small computer for statistical calculations and interpolation. Figures and numbers are based on these results. We have taken as coordinates of gaze Cartesian coordinates on the tangent screen, where the horizontal (h) and vertical (c) components are expressed as degrees of eye deviation from primary position along the horizontal or vertical axis. This choice is more symmetrical than the poi;n coordinates (a&) of the gaze direction used by Robinson (1975). However, our horizontal coordinate /z is identical to the horizontal polar angle 0, while the vertical polar angle 4 is related to (h,~.) by $(h,t~) = tan-‘[tan(~)~(i + tan(h)?)“]). One hits # = t‘ along the coordinate axes j/r = 0;. or {r: = Oi, I#J [ < ju 1 with equality up to second order corrections in h or c. For given h, 4(l7,c) differs (in our range of measurements) maximally from 13 for /u 1 = 30”, with 4(lO”, 30’) = 29,6;, 4(20’, 30”) = 28.5’ and &(30’, 30’) = 26.6 . Hence by ;I slight vertical stretching, which increases with lh /, the theoretical curves in Robinson (1975) can be brought into our format. However this small change is within the resolution of Robinson’s Fig. 5 and within our experimental errors. RESULTS
Recordings of activity from more than 150 axons of extraocular motoneurons were sufficiently stable to determine at least one iso-frequency curve. in 60 recordings the fulf range of horizontal and vertical eye deviations of f30deg around primary position could be covered in steps of IO deg. The sample of 60 axonal recordings consisted of 6 lateral rectus, 12 medial rectus, I3 superior oblique, 15 inferior rectus and 14 motoneurons with upward on-directions. The sample is biased towards predominantly tonic neurons with low thresholds (Henn and Cohen, 1972). We shall group our findings in Figs 2-5 according to different eye muscles, which will be related to the right eye only (after a mirror reflection of data from the other eye) as in the paper of Robinson (3975).
Figure I(A) shows the sM2 inncrvslllian surface of a typical tow threshold left medial rkxtus motoTWKOILFor fixed horizontal eye deviation values of* - 10:. 0’ , . . -I-30” the measured average firing rates, f(+r,r), at vertica! eye deviations of -3Q‘, -2o”, . ~_ -i-30” are represented by b?ack dots, whose distance from the base point in the (h,r)-plane is proportionat to _!“(fh,r:)according to the scale at the right hand corner. The static innervation surface of this unit is then obtained by piece-wise tinear connection ofthese points. In rhis perspective drawing one is looking from the left side onto the rising slope, which for &his neuron is slightly steeper in the lower hemifidd. For the same neuron, Fig. f(B) shows the iso-frequency curves (IFC) for average firing ratesS= 50, 100, 150 1
and 200 Hz, computed from the data represented in Fig. I(A) by finear inteqmlation. if OZ-E defines as innervation gradient the vector field orthogonal to the isawfrequency curves and with length proportional to the slope of the innervation surface at every point (h,~), then the innervat& gradient is horizontal around k = 0 and points about 5 deg downward in the lower right hemifield. Clearly, the iso-frequency curves give a compact 2-dimensional representation of the inntmation surface, fn particuIar, the fevel of 50 Hz is related to the threshold of the unit (at about 2S-30 Hz), while/= 150 Hz is representative for the middle range of the activity (see Henn and Hepp 1984, Fig. 2 for another MR unit which remains active over the range of -t-30 deg around primary position).
Fig. 1. (A) innervation surface sf a kft medial. recks motoneuron. Far fixed ~&es of horizontal c3ye deviation 81 - 10, Q + ict, i-20, and +30delg the static frequeencywas me&Sired during verticaf stow pursuit movements between f 30”dq (up) and -30” dq (down). The average frequencies during four runs were plotted. 3y connecting these points one obtains the innervation surface. (B) iso-frequency curves of the stlme motoneuron for average frequencies of 50, 100, 150 and 200 Hz, computed from the data of Fig. 1(A). By lineear iateq&tion between data points w&in &20” on the IFC closest to the primary
position we have determined the average on-directian
of a nerrron.
K. HEPP and V. HEW
196
A
150 Hr
I
0
-30
adduction
abduction
Fig. 2. Iso-frequency curves of 12medial rectus motoneurons, (A), at 150 Hz, and (Bf. at 50 Hz. The curves are normalized to the right eye with on-direction to the left. In Fig. 2(A) we have represented the iso-frequency 150 Hz for the full sample of 12 medial rectus recordings all drawn as if the on-direction were to the left. One sees that in this middle firing range the innervation gradients are not uniformly horizontal, but that deviations stay within k 5 deg. In Fig. 2(B), the IFC of 50 Hz shows the range of thresholds in this sample which is biased towards low threshold motoneurons. The average on-direction E of a motocurves of
neuron is chosen as the direction of the innervation gradient averaged over the IFC closest to primary position in a vertical range of & 20’. For example, in Fig. 1(B), LXis - 3.2 deg, i.e. slightly downwards relative to the horizontal axis. In our sample of 12 MR axons, cfranges between - 3.2 deg and + 6.8 deg with average Z = 0.8 + 2.9 deg (1 SD). As expected from the neuronal organization of horizontal gaze, the iso-frequency curves of all 6 lateral rectus axons
A
-.20
-3.0
-2.0
-1.0
0
I.0
2.0
h ldwl
3.0
adductioo
Fig. 3. &o-frequency curves of 13 superior oblique motoneurons at 1SOHz (A) and 50 HZ (B), related to the tight eye.
A
20
3,o
h (dwl
Fig. 4. Is&frequency curves af 15 infer&x rctus motoneurons at 150 Hz {A) and 50 l-5~(8X related to :he right ye.
from the Vith nerve (and also of all 5 internuclear fibers recorded in the MU) were found to be indistinguishable from those of the MR axc~~s.
Figure 3 shows the EC at SO and I50 Hz for-our sampie of 13 SO fibers fse+zaiso Menn and Happ, 1984, Fig. 3 for a typical example in the format as in Fig. I). In the medial hemifield the innervation directions are downward, and they are downward and outward in the lateral hemitield. The average
on-directions of truchlear motoneurons strongly fluctuate: 1 varies between t .2 and 21.2 deg downwards lateral, with 5 = 9.1 + 5.3 deg. Some SO units have almost purely downward innervation gradients and IFC which are almost straight lines. Within the same tract through tbe trochlear nerve one can find units with such differences in their on-dirtxtjons, However, &heSO ~o~uiat~on average has an almost downward innervation gradient in the media{ and a downward and uutward direction in the lateral hemifield, even for the two high threshold units in our sample.
A
Fig. 5. Iso-frequency curves of 14 mNmxmns
with upward on-directions at I SOHz (A) and 50 Hz (B), related to the right eye.
In contrast to superior oblique and similar to the horizontal recti the innervation directions of inferior rrctus motonrurons show much smaller fluctuations: in our sample of I5 recordings x varied between 0.7 and 13deg downwards and medial with !i = 4.4 k 2.9 deg. The IFC at 50 and I50 Hz in Fig. 4 show in the lateral hemifield a downward innervation gradient which is slightly medial in the medial hemifield. Although our sample does not contain any high threshold unit. on2 sees clearly that the innervation gradients of this muscle is the same at high and low levels of innervation and that the IFC are straight or convex downwards,
They have not been unambiguously separated into superior rectus and inferior oblique units. In our sample of 14 complete axonal UM recordings (with 6 high threshold units) the IFC in Fig. 5 are straight lines for high innervation values. while for low tiring rates the innervation directions are slightly medial in the medial and lateral in the lateral hemifield. This behavior is different from the downwards-directed motoneurons. However, the sample is obviously rather inhomogeneous. In particular, there is one unit which shows a SO firing pattern inverted at the horizontal axis. The average on-directions r near primary position varied between - 5.7 deg (medial) and 14.6deg (lateral) upwards, with d = 1.7 + 4.6 deg. Again the IFC, except for the strongly curved unit, are parallel. DISCUSSION The static innervation of individual motoneurons in the monkey give a rather consistent picture. The mere existence of iso-frequency curves with small fluctuations over the range of 30 deg around primary position supports the laws of Listing and Sherrington. The innervation gradients for the horizontal
recti are horizontal, for superior oblique downward in the medial and downward lateral in the lateral hemifield, and for inferior rectus they are downward with a small medial component, slightly larger in the medial hemifield. This corresponds to predictions about pulling directions based on the geometry of eye muscles. For the upward motoneurons the innervation gradients are upward and often have a medial component in the medial and a lateral component in the lateral hemifield. Only one unit had an isofrequency curve similar to those of trochlear fibers (reflected at the horizontal axis). For every motoneuron all iso-frequency curves are remarkably parallel for small and large values of innervation, and they are mostly convex from lower to higher values of innervation. We consistently found that iso-frequency curves for high and low innervation are parallel within f 30 deg around primary position. Measuring on2 complete IFC during low activity then allows one to predict the shape of
IFCs at higher ~it;vtty. This I\ ;L p~~tc’nt~,rlai! .,!I:[~:;:’ for slectromyographic recording !n h.~~~n.~ :\h<:tb for technical reasons the rssolution 10r ~nple un~[> 1. good at IOU- innervat:Jn levels clnly. 0~ restlit; suggest that such recordings are sufficisnt to pred:ct unit behavior at highrr innervation levels. Within rhe populations of horizontal recti and of the infrr~or rectus moton2urons. the arsrage on-direction at primary position varies little. and about half as much as within the population of superior oblique neurons, That there are differences in on-directions of individual neurons of the same muscle pool could be statistical fluctuations. changing alertness of the animal. or. as we believe. reflect a different orientation of the respective muscle fiber between its tendon insertions As muscle insertions on the globe are broad, one would expect different innervations for individual fibres to obtain the same tension over the cro’;ssection of a muscle. Our experimental tindings in the Rhesus monkey basically agree with the theoretical predictions by Robinson (1975) for man. However. there art‘ also points of disagreement. In Robinson’s calculation medial rectus motoneurons have not the same innervation as lateral rectus neurons and the convexity of the iso-frequency curves is opposite to our data. The measured iso-frequency curves for trochlear motoneurons remain more parallel over the range 01 + 30 deg than predicted by Robinson. These discrepancies could be a species difference, could be due to the theoretical assumption of pairing the vertical recti and obliques. or could be a bias of the smallness of our sample. Lesion studies indicate that horizontal and vertical saccadic eye movements are organized in iwo Cartesian directions (Bender and Shanzer, 1964: reviewed in Henn et al., 1984). We propose that a graded, eye-position dependent innervation to all I :! eye muscles in accordance to Listing’s law is accomThe function of this plished by an “integrator”. integrator is to convert the velocity related activity in premotor units to the sustained activity found in motoneurons. We have to admit that details of its neuronal realization and localization has remained an enigma. We have not resolved the question of how to pair the vertical muscles in antagonists or in yoked partners. The anatomy and electrophysiology of the disynaptic vestibulo-ocular reflex suggest an antagonistic pairing of the superior rectus (superior oblique) with the inferior rectus (inferior oblique) (Precht. 1978). This is the basic scheme which Robinson (1975) has applied for his calculations. In the cat (Uchino, 1980) and therefore probably also in the monkey, the disynaptic connection from one canal involves more than only one antagonistic pair (reviewed in Ito. 1984). The anterior canal. e.g. excites both superior rectus muscles and the inferior oblique of the ipsilateral eye. and inhibition is similarly conveyed to both the inferior rectus and superior oblique.
Motoneuron
iso-frequency curves
Our data show clearly that the inferior rectus of one eye and the superior oblique muscle of the other eye do not receive the same innervation during conjugate downward gaze. Therefore these two muscles are not yoked together in the sense of Tiering. Our data also indicated that the upward motoneurons do not fall into two populations with innervations. which are the mirror images of the inferior rectus and superior oblique motoneurons. Therefore we favor the hypothesis that there is no rigid pairing and yoking of vertical motoneurons which predicts conjugate eye movements for tertiary eye positions. In our experimental setup we were unprepared to encounter such complexities. Future experiments should therefore aim at a more precise measuring of motoneuron activity in relation to eye positions and cyclorotation, during slow and rapid eye movements, and consider different inputs from the visual and vestibular systems. ~e~no~~~e~ge~~ffr~-We are grateful to Dr H. Reisinc for helpful comments and to MS V. Isoviita for technical assistance. K. H. would like to thank Professor G. Baurngartner for his kind hospitality.
REFERENCES Bender IM. B. and Shanzer S. (1964) GcuIomotor pathways defined by electric st~muiation and lesions in the brainstem of the monkey. In 7% Ocutomoror System (Edited by Bender M. B.) pp. 81-140. Harper % Row, New York. Boeder P. (1962) Co-operative action of extra-ocular muscles. Br. J. Ophthaf. 46, 397-403. Collins C. C.. G’Meara D. and Scott A. B. (1975) Muscle tension during unrestrained human eye movements. J. Physiot. 245, 354-369.
Eckmiller R. (1974) Hysteresis in the static characteristics of eye position coded neurons in the alert monkey. PfIiigers Arch. ges. Ph_ysiof. 350, 249-258.
Fuchs A. F. and Luschei E. S. (1970) Firing patterns of abducens neurons of alert monkeys in relationship to horizontal eye movement. J. Seurophysiot. 33, 382-392.
499
Helrnholtz H. v. ( 1856-i 866) Handbuch der physioiogischen Optik. Voss, Leipzig. Henn V. and Cohen B. (1972) Eye muscle motor neurons with different functional characteristics. Brain Rex 45, 56 i-568.
Henn V. Cohen B. (1973) Quantitative analysis of activity in eye muscle motoneurons during saccadic eye movements and fixation. J. Keurophysiol. 36, 115-126. Henn V. and Hepp K. (198I ) Two-dimensional analysis of eye movements in oculomotor and premotor structures. In Progress in Qculomotor Research (Edited by Fuchs A. F. and Becker W.). pp. 81-88. Elsevieri’North-Holland. Henn V. and Hepp K. (1984) Motoneuron activity around the primary position and in extreme positions of gaze. In Trans. Ftyth htt. ~rr~lopt~r Congr. (Edited by Ravault A. P. and Lenk M.), pp. 233-238. Lips, Lyon. Henn V., Lang W.. Hepp K. and Reisine H. ( 1984)Experimental gaze palsies in monkeys and their relation to human pathology. Bruin 107, 619-636. Hering E. (1868) Die Lehre vorn binocularen Sehen. Eng&nann, Leipzig. English translation: The Theory of Binocular Vision. Plenum Press. New York (1977). Ito M. (1984) The Cerebellum and Neural Control. Raven Press, New York. Keller E. L. and Robinson D. A. (1972) Abducens unit behavior in the monkey during vergence movements. Vision Res.
12, 369-382.
Precht W. (1978) Neuronat Operations in the Yestibular System. Studies in Brain Function. Vol 2. Springer, Berlin. Robinson D. A. (1970) Oculomotor unit behavior in the monkey. J. Neurophysiol. 33, 393-104. Robinson D. A. (1975) A auantitative analvsis of extraocular muscle cooperation and squint. invest. Ophthal. sisual Sci. 14, 801-825.
Schiiler P. H. (I 970) The discharge characteristics of single units in the oculomotor and abducens nuclei of the unanesthetized monkey. Expf Bruin Res. IO, 347-362. Uchino Y., Suzuki S. and Watanabe S. (1980) Vertical semicircular canal inputs to cat extraocular motoneurons. Expt Brain Res. 41, 45-53.
Westheimer G. (1957) Kinematics of the eye. J. opr. Sot. Am. 47, 967-974.
Westheimer G. (1981) Donders’. Listing’s, and Hering’s Laws and Their Implications. In Models of Ocuiomotor Behavior and Control (Edited by Zuber B. L.), pp. 149-159. CRC Press, Boca Raton, Florida. Wurtz R. H. (1969) Visual receptive fields of striate cortex neurons in awake rnonkevs. 32. 727-742. , J. NeuroohwioL Li