Excess cyclovergence in patients with intermittent exotropia

Vision Res. Vol.35, No. 23/24,pp. 3265 3278,1995

Pergamon

0042-6989(95)00027-5

Copyright,,1},1995ElsevierScienceLtd Printed in GreatBritain.All rightsreserved 0042-6989/95$9,50+ 0.00

Excess Cyclovergence in Patients with Intermittent Exotropia A. V. VAN D E N BERG,* L. J. V A N RIJN,* J. TJ. H. N. DE FABERI" Received 17June 1994; in revised form 18 November 1994

Recently, we developed a model of binocular fixation. This model predicts the amount of cyclovergence as a function of target elevation and horizontal target vergence. The prediction derives from the assumption that version and vergence add linearly and that the eye positions are constrained in three respects: (1) the foveae of the two eyes are directed towards the target, (2) the version component follows Listing's law, i.e. cycloversion, and horizontal and vertical version are not independent, (3) the vergence component is restricted to a plane approximately perpendicular to Listing's plane, i.e. horizontal, vertical and torsional vergence are not independent. The version and the vergence components are characterized by a common primary direction for the two eyes. We applied this model to data of patients with intermittent exotropia. In two patients with an amblyopic eye we found that the common primary direction rotates towards the amblyopic eye. In the third patient, not suffering from amblyopia, the common primary direction was practically straight ahead. In all three patients, cyclovergence angles were larger than those found in normal subjects. We found that the increased cyclovergence was compatible with our model for normal subjects if an offset on the horizontal vergence was given. This offset represents the additional convergence effort required in these patients to overcome the exodeviation of the eyes. According to our model the increased horizontal vergence effort results in excess cyciovergence. The relation between horizontal vergence and cyclovergence offers a new method for measuring the angle of exotropia. Eye movements Strabismus Torsion Human

INTRODUCTION

Coil

eye positions have gained widespread attention only recently. It is well known that a disturbance of the binocular In principle, binocular eye positions have six degrees control of eye movements like strabismus, interferes with of freedom that correspond to the horizontal, the vertibinocular single vision. To date, most attention has been cal and the torsional position for each eye. However, given to the description of horizontal and vertical eye normal eye positions are more constrained. Bifoveation position abnormalities, because these movements are of a target reduces the degrees of freedom by one, decisive whether the subject attains bifoveal fixation of because the lines of sight need to intersect at the fixated an object or not. Little attention has been given to the object. Recent studies (Mok et al., 1992; van Rijn & van torsional dimension of the eye position. Obviously, some den Berg, 1993; Minken & Van Gisbergen, 1994) have amount of binocular control o f eye torsion is necessary shown that the torsions of the left and the right eye are to maintain correspondence between the two retinal not independent when fixating targets nearby or in the images. This relative lack of attention is not surprising. distance. It was found that normal eye positions are First, normative data on binocular control of eye torsion approximately constrained to three degrees o f freedom, were unavailable until very recently (Mok, Ro, Cadera, just like the target position in front of the head. This Crawford & Vilis, 1992; van Rijn & van den Berg, 1993). means that the two eyes' positions vary within a narrow Also, the objective measurement of eye torsion in a range about two prescribed orientations that only declinical setting is still quite rare. Finally, the mathematipend on the target position. This is a binocular form of cal tools (Haustein, 1989; Tweed, Cadera & Vilis, 1990) Donders' law. What are these eye positions like during necessary for the evaluation of three-dimensional binocular fixation? For distant targets the lines of sight are parallel and *Department of PhysiologyI, Faculty of Medicine, Erasmus Univer- the two eyes are identically oriented. In this case the eyes sity, Rotterdam, P.O. Box 1738,3000 DR Rotterdam, The Netherobey Listing's law. This law, delivered by von Helmholtz lands [Email [email protected]]. tDepartment of Pediatric Ophthalmology and Strabismus, The (1867), describes eye torsion as a byproduct of the Rotterdam Eye Hospital, The Netherlands. direction of fixation. Eye torsion is a simple function of 3265

3266

A.V. VAN DEN BERG et al.

the horizontal and the vertical eye position. A very concise description of Listing's law states that each eye position can be derived from a special (primary) eye position through rotation about an axis in the Listing plane. The primary direction is perpendicular to Listing's plane. Many studies have confirmed the law's validity for fixation of distant targets or for monocular vision using objective techniques (Ferman, Collewijn & van den Berg, 1987a, b; Tweed & Vilis, 1990; Tweed et al., 1990; Minken, van Opstal & van Gisbergen, 1993; Haslwanter, Straumann, Hepp, Hess & Henn, 1991; Tweed, Fetter, Andreaaki, Koenig & Dichgans, 1992). Also, the shape of Listing's plane is generally maintained when the body is rotated with respect to gravity (Haslwanter, Straumann, Hess & Henn, 1992). Nevertheless, during sleep, after prolonged training or during convergence, violations have been found (Nakayama, 1975, 1983; Balliet & Nakayama, 1978). Also, when large combined head and eye movements are made, the eye positions relative to the head do not obey Listing's law (Glenn & Vilis, 1991). Here, we are mainly concerned with the violations of Listing's law that occur in near vision. These violations have a lawful character and have recently been modeled by two of us (van Rijn & van den Berg, 1993). In correspondence with their model van Rijn and van den Berg found little or no eye torsion for convergence in the plane that contains the primary direction and the interocular axis. While looking upward both eyes intort during convergence. While looking downward the eyes extort when converging. Because the model is essentially binocular and provides a quantitative prediction of eye torsion for near and far vision alike, it offers a useful tool for the interpretation of eye torsion in conditions like strabismus. Thus, we set off with a brief description of the concepts on which our model is based. After a description of the patient data we will propose a modification of our model that enables us to describe the patient data as well. A binocular extension o f Listing's law

To derive a prediction for the eye orientations during binocular fixation we needed to make a number of assumptions on the control of eye movements. We assumed that oculomotor control is essentially binocular and head-centric. Thus, the innervations of each eye's muscles were considered to refer to a common reference frame in the head with an origin, that is located midway between the eyes. In line with Hering's law we proposed that the control signals consist of a version component and a vergence component. The former results in identical changes in eye orientation, the latter results in differences in the orientation of the left and the right eye. The version and the vergence component are additively combined. In normal subjects, the disjunctive component is zero for fixation of targets at optical infinity and the eyes are oriented identically. Because it is known that the eyes obey Listing's law for fixation of distant targets, this means that the conjugate component obeys Listing's law. Also, it means that the conjugate component of the eye positions is located in the same plane

irrespective of the fixation distance. Hence, the common primary position (which is the normal direction to this plane) acts in our model as a true "constant", a characteristic of oculomotor fixation. In contrast, no such constancy applies to the primary position of each eye separately. This holds, because the primary position of each eye can be shown to rotate laterally during convergence (Mok et al., 1992; van Rijn & van den Berg, 1993). Our formulation of Listing's law essentially states that the conjugate torsion of the eyes depends on the conjugate horizontal and vertical components of the eye position. We proposed a similar law for the vergence component, which means that cyclovergence, horizontal vergence and vertical vergence are not independently controlled. Finally, the oculomotor system aims to direct both foveae towards the target (both lines of sight in the plane of regard), Together these three constraints reduce the number of degrees of freedom for the two eyes from six to three. Modern descriptions of eye positions (including torsion) often use "rotation vectors" (Haustein, 1989). The rotation vector (r) emphasizes the direction of the rotation axis rather than the direction of gaze. Its three components are related to the turn angles about three head-fixed axes that are perpendicular to one another. Together they specify the direction of the axis through the globe about which the eye is turned from the reference orientation (determined by the measurement system) towards the current eye position. The length of r specifies the tangent of half the turn angle about this axis. The sign of the rotation vector indicates the direction of the turn about the axis according to the right-hand rule. (To lift the valve of a water tap the hand must turn leftward. Similarly, an upward rotation vector denotes a leftward gaze shift.) This is quite different from the ordinary eye position coordinates in terms of Fick or Helmholtz angles, in which only one axis has a fixed orientation with respect to the head [Fick coordinates, the vertical axis; Helmholtz coordinates, the horizontal, interocular axis (cf. Howard, 1982)] and in which the other two axes move with respect to the head. The rotation vector scheme has gained some popularity because the three eye muscle pairs turn the eye about head-fixed axes and also because the rotation vector scheme allows for a very concise description of Listing's law: all rotation vectors of the eye are located in a plane. We used a mixture of rotation vectors and Helmholtz angles (Table 1 specifies the meaning of the components and their directions) to investigate the effect of the constraints mentioned above (s~ = 0, g2 = 0, dO = 0) on the predicted eye orientations. In our scheme the version component (s) is found by averaging the rotation vectors of the left (r0 and the right (r,) eye (s = (rj + rr)/2); the vergence component (g) is found by taking half the difference of the rotation vectors of the left and the right eye (g = (r~ - r,)/2). In normal subjects we found that s is located in an approximately fronto-parallel plane irrespective of the fixation distance. Correspondingly, the torsional component of s (s~) was about zero. This meant that the common primary direction of the eyes

EXCESS C Y C L O V E R G E N C E

3267

T A B L E l(a). Monocular Hemholtz angles and their relation to rotation vectors Gaze c o m p o n e n t

Helmholtz angles

Rotation vector

Horizontal Vertical Torsional

ct, (left positive) 0~ (down positive) 0~ (CW positive)

r3i = ctd2 - 0~0d4 (up positive) r2~ = 0i/2 + ~0~/4 (left positive) rti = 0d2 + 0~ctg/4 (forward positive)

Whereas Helmholtz coordinates specify angles of rotation about one head fixed axis and two axes that rotate with respect to the head, the components of the rotation vectors refer to angles of rotation about three head-fixed axes (x, y, z in Fig. 1). Both representations use the same reference direction of the eye. It corresponds to the viewing direction perpendicular to the revolving magnetic field of the Skalar system. The symbols carry a subscript 'i ', which can take the values T or 'r' for the left or the right eye respectively. The expressions for the rotation vectors are not exact, but approximations with deviations less than 2.5% for the range of eye positions that we measured. The rotation vector is directed according to the rotation axis required to turn the eye from the reference position (ct = 0 = 0 = 0) to the current eye position (ct, 0, 0). To illustrate, a downward rotation (0 > 0 , a = 0 = 0) corresponds to a leftward rotation vector (r 2 > 0, r t = r 3 = 0). Clockwise torsion (CW) translates the upper pole of the left eye towards the nose.

was directed perpendicular to the interocular axis. Similarly, g was found to be located in the mid-sagittal plane (i.e. g2 ~ 0). Figure 1 shows the orientation of the planes relative to the head and the s and g components corresponding to left upward fixation. These findings support the notion of a "Listing law" for both version and vergence in normal subjects. In addition, the three constraints mentioned above resulted in a prediction for the cyclovergence component of the rotation vectors

of the visual stimulus, the fixation distance and the convergence effort of the subject. The questions we aimed to answer were: (1) Do the normal relations between eye torsion and horizontal/vertical eye positions apply in these patients? (2) Is there a relation between our model parameters and the clinical characteristics (angle of strabismus, type of exotropia) in these patients. (3) Can we give an indication for the site of the 'deviation' (in one or in both eyes; either centrally or peripherally)?

(g,): gl ---- 2s2g3 ~

Or~4

(1)

i.e. cyclovergence is proportional to the product of horizontal vergence (v) and elevation (0), in Helmholtz coordinates. This linear relation was also well supported by the data in five normal subjects, although we found usually about 10-20% smaller cyclovergence than predicted. One major restriction of our model is that binocular fixation of the targets is required. As yet, the model is not suited for manifest strabismus. For this reason we tested in the present experiment patients with intermittent exotropia. This is a conditions in which the eyes in absence of visual stimuli are rotated outward with respect to each other. In the presence of visual stimuli binocular vision may exist, depending on the nature

To anticipate, we found two major differences compared to the data in normals, as described in van Rijn and van den Berg (1993): (a) the amount of cyclovergence concomitant with horizontal vergence was much larger; (b) in two patients with an amblyopic eye the common primary direction was rotated towards this eye. Some of these findings have been published in abstract form (van Rijn, van den Berg & De Faber, 1993).

METHODS Subjects

Five patients with intermittent

T A B L E l(b). Binocular Hemholtz angles and their relation to vergence and version components of the rotation vectors Gaze component

Helmholtz angles

Rotation vector

Horizontal version Vertical version Torsional version Horizontal vergence Vertical vergence

ct (left positive) 0 (down positive) 0 (CW positive) v (divergence positive) dO (left eye down right eye up positive) d o (intorsion positive)

s 3 = ct/2 - 00/4 s? = 0/2 + ~ 0 / 4 s I 0 / 2 + 0a/4 g3 = v /4 -- d00/8 - d 0 0 / 8 g2 = d0/4 + d0ct/8 + Or~8

Torsional vergence

=

gl = d 0 / 4 + d0ct/8 + Ov/8.

Notice that the Helmholtz version angles specify the viewing direction towards the target as seen from the point midway between the eyes. Similarly, the components of s specify the rotation vector required to shift the gaze of an imaginary eye at the ego-centre. Positive directions for the components of s and g are identical to those from the rotation vectors of each eye [see Table l(a)].

exotropia

participated

A.V. VAN DEN BERG et al.

3268

Z

x

FIGURE 1. The model for binocular fixation in normal subjects. The figure shows the orientation of the plane of s vectors and the plane of g vectors relative to the head. In this idealized situation the x, y and z axes of the measurement system are aligned with the main axes of the s and the g planes. When fixating a target in the reference direction (the central target on the screen) the rotation vector is zero. For left-upward fixation the arrow in the s plane represents the conjugate part of the rotation vectors. Its horizontal component (s2), indicating rotation about the y-axis, is rightward directed because rotation about this axis according to the right-hand rule shifts the gaze upward. Its vertical component (s3), indicating rotation about the z-axis is directed upward. Its x-component (st) is zero, because Listing's law applies to s. Rotation about the z-axis according to the right-hand rule shifts the gaze leftward. If the screen is nearby the eyes need to converge and the g component is not zero. The arrow in the g plane indicates the vergence part of the rotation vectors required to fixate the left-upward target on the screen. The vergence component is anti-symmetric. It is added to the s component to obtain the rotation vector for the left eye. It is subtracted from the s component to find the rotation vector for the right eye. The indicated g vector causes rightward (g3 negative) and clockwise (g~ positive) rotation for the left eye and leftward and anti-clockwise rotation for the right eye (n.b. - g is added to s in the right eye). Thus, both eyes intort.

in these experiments. W e r e p o r t the d a t a o f three patients only, because the o t h e r two patients did n o t achieve b i n o c u l a r fixation o f the target in m a n y presentations. Patients were recruited f r o m the o u t p a t i e n t d e p a r t m e n t o f the R o t t e r d a m Eye H o s p i t a l . T h e y were c o n s i d e r e d for s t r a b i s m u s surgery. N o n e o f t h e m h a d a n y strabism u s surgery in the past. E x o d e v i a t i o n s exist in different types a n d we first s u m m a r i z e these briefly before c o n t i n u i n g with a detailed d e s c r i p t i o n o f o u r subjects. D u a n e (1896) a n d others (Von N o o r d e n , 1990; Burke, 1985) distinguish f o u r types o f (intermittent) e x o t r o p i a , d e p e n d i n g o n its b e h a v i o u r when fixating a n e a r b y o r d i s t a n t target: (a) divergence excess, the d e v i a t i o n is g r e a t e r for a d i s t a n t target; (b) basic exodeviation, the d e v i a t i o n is i n d e p e n d e n t o f the distance; (c) convergence insufficiency, the d e v i a t i o n is greater for n e a r vision; (d) simulated divergence excess pattern, a basic

e x o d e v i a t i o n type in which the n e a r d e v i a t i o n is o b s c u r e d by excessive convergence. O f the patients, one h a d a basic exodeviation; one h a d a basic e x o d e v i a t i o n c o m b i n e d with a small convergence insufficiency a n d one h a d a s i m u l a t e d divergence excess pattern. P a t i e n t F R was a 32 yr old female who h a d episodes o f e x o t r o p i a since early c h i l d h o o d . U p to a b o u t her 20th y e a r o f age the e x o t r o p i a was well c o n t r o l l e d b y vergence training, thereafter c o n t r o l got progressively m o r e difficult. She experienced d o u b l e images especially when l o o k i n g left- a n d r i g h t w a r d . O p h t h a l m o l o g i c e x a m i n a t i o n s : Cycloplegic refraction was S + 0.75 for the right eye ( O D ) a n d S + 3.75 C-0.25 axis 90 deg for the left eye (OS). Visual acuity was 1 . 0 0 D a n d 0.25 OS with o p t i m a l correction. Stereopsis t h r e s h o l d was 80 sec o f arc (Titmus test; T i t m u s O p t i c a l Co., Virginia). There was an e x o p h o r i a o f 12A (6.8 deg) d u r i n g n e a r fixation ( m e a s u r e d with the s y n p o t o p h o r at 33 cm distance) a n d 30A (16.7 deg) for fixation at distance ( m e a s u r e d with a M a d d o x cross at 2.5 m distance) ( A - - p r i s m diopters).

EXCESS CYCLOVERGENCE During near fixation there was excessive pupilloconstriction. These findings are compatible with a simulated divergence excess pattern. There was evidence of a mild IVth nerve palsy and latent nystagmus. Patient VL, a 33 yr old male, had since his 12th year episodes of exotropia associated with fatigue. These episodes increased in frequency in recent years. He never experienced double images. Ophthalmologic examinations revealed a visual acuity of 1 . 6 0 D S and no refractive errors. The stereopsis threshold was 40 sec of arc. The angle of exophoria was 30A (16.7 deg). The angle was identical for fixation nearby and at distance (basic exodeviation). Ocular motility was normal. Patient AM was a 56yr old male who since his childhood experienced double images during near fixation. In addition, he had increasing complaints of strain in his left eye. Ophthalmoiogic examinations revealed a cycloplegic refraction of S + 2 . 5 0 D and S + 4 . 5 C-2.0 axis 30deg OS. For reading he used an addition of +2.75 ODS. The visual acuity with optimal correction was 2 . 0 0 D and 1 . 0 0 S . Stereopsis was present (Fly positive: Titmus test) The angle of exophoria was 18A (10.2 deg) for near fixation and 12A (6.8 deg) for fixation at distance (basic exodeviation with small convergence insufficiency). Ocular motility was normal.

Recording techniques The positions of both eyes were measured simultaneously and in three dimensions (horizontal, vertical and torsional), using scleral coils of the combination type (Collewijn, van der Steen, Ferman & Jansen, 1985; supplied by Skalar, Delft, The Netherlands). Coil signals were amplified, low-pass filtered (cut-off frequency 125 Hz) and sampled at 250 Hz. Samples were stored on a minicomputer (DEC PDP 11/73) for off-line analysis. Prior to each experiment the gains of the amplifiers were calibrated and offsets were nulled, with the coils mounted on a gimbals device. Protocol The experimental protocol was almost identical to that described in van Rijn and van den Berg (1993). Its total duration was restricted to about 30min to minimize the discomfort for the patient. Subjects were seated with their torso in the magnetic field cubical, their heads kept in position by a bite board. Point targets for fixation were presented in various gaze directions at one of two distances. Single point targets were presented in an otherwise dark room to preclude visual cues that could provide a reference for eye torsion (see, e.g. van Rijn, van der Steen & Collewijn, 1992). The distant targets were rear-projected on a translucent screen 143 cm in front of the subject's eyes (Fig. 1). To present nearby targets, a frame was placed at 35 cm distance from the eyes. This frame contained LEDs in the same set of gaze directions with respect to the straight ahead as for the distant targets. To reduce stray light the frame was covered with featureless material identical to that of the translucent screen.

3269

An experiment consisted of 80 measurements of 4 sec. During a measurement one target was presented, in the straight ahead direction or in one of eight equally spaced viewing directions at a constant eccentricity of either 10 or 15deg (Fig. 1). The subjects fixated the target binocularly. Recording of the eye positions was started by the subject when binocular fusion of the target was achieved. Each target position was presented twice in the course of an experiment. Two calibration measurements (one for each eye) were intercalated between every seven or eight normal measurements. In calibration measurements the target was presented straight ahead (at 143 cm) with respect to one eye while the other eye was covered. Thus, two calibration targets were used with a lateral displacement equal to the inter-ocular distance. The averages of the direction signals and the torsion signal of the viewing eye during the calibration quantified the coil misalignment. The repeated calibration measurements enabled us to detect changes in the coil misalignment during the experiment, which indicate coil slippage. Differences between two successive calibration measurements were no more than 2 deg and usually less than 1 deg.

Data analysis Data analysis started with the correction for coilmisalignment. Correction was done using the pair of calibrations, presented most closely in time before or after the measurement. The three coil signals of the measurement were transformed using a procedure described in Ferman, Collewijn, Jansen and van den Berg (1987). This transformation is the mathematical equivalent of shifting the coil's position on the eye so as to align its surface normal with the visual axis and to align the plane of the torsion coil with the vertical plane. Corrected data specified the eye position in Fick coordinates relative to the (precalibrated) reference direction. Subsequently, the data were transformed into Helmholtz coordinates (azimuth ~, elevation 0, torsion ~O). For each measurement the means of ~, 0 and q; for each eye were calculated and entered into a spreadsheet for further calculations. These included: (a) a transformation of the three Helmholtz angles for each eye to obtain its rotation vector [r~ or r,, see Table l(a)], (b) computation of the version (s = (rl + rr)/2) and vergence (g = (rl - r,)/2) parts of the pair of rotation vectors, (c) computation of the binocular primary direction from the orientation of the plane of s vectors (see van Rijn & van den Berg, 1993). In addition we performed a regression analysis to evaluate the relation between cyclo-vergence and horizontal vergence.

RESULTS

Vergence and version components of the rotation vectors The version and the vergence components of the rotation vectors were found to be located in planes, just as in normals (van Rijn & van den Berg, 1993). The plane of s roughly coincides with the frontal plane,

3270

A . V . V A N D E N B E R G et al.

although there is in all subjects some backward tilt. In two of the subjects (AM and FR) there is also some lateral rotation of the plane of the rotation vectors, towards the amblyopic eye. The direction of the backward tilt corresponds to our findings in normal subjects. Also, the magnitudes of the backward tilt were within the range we previously described for normals. We argued before that it may be a consequence of the absence of any objective zero for pitch head position. The lateral rotation of the plane of s vectors was not observed in normals. The plane of g vectors was aligned with the sagittal plane in subject VL. In the other two subjects it deviated from the sagittal plane towards the amblyopic eye. The orientation of the plane of s vectors was determined with a regression analysis fitting s~ as a function of s2 and s 3 taking the pooled data of both distances. From this analysis we found the common primary direction (Ps) of the eyes. Results are listed in Table 2. The numbers in this table reflect the magnitudes of the lateral and the upward shift of the primary direction with respect to the reference direction (the centre fixation point) in our setup. In subjects AM and FR, who suffered from amblyopia of the left eye, the plane of s-vectors (the version part of the rotation vectors) was rotated to the left by about 5 deg. Consequently, the common primary direction was rotated to the left by twice this amount (Tweed et al., 1990), i.e. about 10 deg. Interestingly, in these patients we found a leftward rotation of the plane of rotation vectors (displacement plane) by about 10 deg for the left eye but no such rotation for the right eye. The planes of g vectors were rotated also towards the amblyopic eye in subjects FR and AM. The leftward rotation was about 3 deg in AM and 5 deg in FR. The next step in the analysis involved a change of the frame of reference. All horizontal and vertical Helmholtz angles were offset with the common primary direction of 'Ps' and the rotation vectors and s and g components were recalculated on the basis of these corrected angles. As a result, all rotations were now defined with respect to the common primary direction 'Ps' instead of the precalibrated reference direction. Figure 2 shows for each subject the planar arrangement of the s and g

T A B L E 2. Azimuth and elevation angles of the c o m m o n primary direction (Ps), i,e. the primary direction based on the rotation vectors of both eyes (s-plane) Patient AM FR VL

Azimuth (deg)

Elevation (deg)

Amblyopic eye

+ 10.4 +8.1 - 1.6

-4.9 - 14.7 - 5.7

L (mild) L none

Leftward and downward rotations are positive. In subjects A M and F R there was a lateral rotation of Ps (non-zero azimuth) that was more prominent than in normal subjects. Rotation of the s-plane was in the direction of the amblyopic eye. In all patients, Ps also pointed upward, corresponding to the findings in normals.

vectors relative to 'Ps'. Calculations of thicknesses of planes was done using these corrected values. One may observe in subjects AM and FR that a leftward rotation of the plane of g vectors remained despite the coordinate transformation.

Testing of model assumptions Our scheme which was confirmed in five normal subjects produces quantitative predictions of binocular eye positions because binocular control of the eyes is restricted in three ways. In this section we evaluate whether these constraints are valid for the patients with intermittent exotropia as well. Our hypothesis that version and vergence comply with a form of Listing's law predicts that s and g are restricted to planes, irrespective of the fixation distance. The validity of these assumptions can be appreciated from Fig. 2: the planes for the two fixation distances (represented by open and closed symbols) do not show major differences. We tested these assumptions formally by calculating the SD in s~ and g2 values, separated according to fixation distance. Results are listed in Table 3. The s planes were about 25% thicker than in normals, but as in normal subjects there was no fixation distance dependence. The SD for the g planes was not different from the normals and also did not vary with the fixation distance. Bifoveal fixation of the target, the third constraint in our model, implies that the two lines of sight must be located in one plane, the plane of regard. Stated otherwise: the vertical (Helmholtz) vergence dO equals zero. There was a small but significant difference from zero: dO's averaged across all measurements ranged from -0.25 to 0.17 deg among our subjects with, SDs of about 0.3 deg within a subject. We reported similar values for normal subjects (van Rijn & van den Berg, 1993). We conclude that the model constraints were well supported by the data, but that the variation in sl was somewhat larger than in normal subjects.

Cyclovergence An important implication of our scheme is the prediction that the cyclovergence component of the rotation vector merely depends on the horizontal vergence and the elevation angles. The fact that the vergence component is located in a plane in our subjects just as for normal subjects does not necessarily mean that the relation between horizontal vergence and cyclovergence is also normal. This follows because the g component may be shifted within the g plane, resulting possibly in a different ratio between horizontal vergence and g~. We found in all our patients much larger values of cyclovergence than in normal subjects. Figure 3 shows for three subjects the predicted (2szg3) and the measured values of g~ (cyclovergence), separated according to the fixation distance. In normal subjects we found a single line with a slope slightly less than 1 (van Rijn & van den Berg, 1993). The results in the patients showed a similar linear relation between the predicted and the measured g~. However, the slopes of these regression lines were

EXCESS C Y C L O V E R G E N C E

3271

(a) D

2

.3

.15

AM

d

.I

b

~5

cf t

.2

P

15

.05

~1.1

P

i

-.05

)5

~o

-.1

0 • 0.35 rn o 1.43 m

-.15 U-.2

D aM fl

~..1 -.15

-.1

-.05

0

.05

.1

.15

.2

-,2

-.15

-.1

0

-.05

Cycloversion

$1

L+R+

.05

.1

.15

.2

sl

Di

.04

.08

.03

.06

.02

.04

,01

.02 .

0

•

,I,

^ =~ala

~

I

,.o.-tr

°

~s

o%~

,~Wo%oo

-.02

-.01 -.02

5dl • •

oe~

gSo ~,

•

-.03 -.08

-.04

L~R~ -.04

-.03

-.02

-.01

0 gl

.01

.02

.03

.04

Con

-.025

-.02

Cyclovergence

-.015

-.01

-.005

0

.005

.01

.015

gl

F I G U R E 2(a). Caption on p. 3273.

much larger than one, indicating excess cyclovergence in these patients compared to normals. This was particularly the case for the largest fixation distance (143 cm). To explain this observation one may note that for our range of fixation directions the horizontal vergence is practically constant (SD of horizontal vergence for one target plane ranged from 0.28 to 0.58 deg in our subjects). This implies that for each fixation distance the predicted gl is a linear function of the elevation (=2s2 ,~ 0) with the horizontal vergence ( = g 3 ~ v/4) as the parameter, determining the slope of the regression line. The range of elevations was identical for the two fixation distances. Following this interpretation the increased slope reveals an apparent increase of the horizontal vergence in these patients. Because the measured horizontal vergence was adequate for binocular fixation of the target, this may mean that larger horizontal vergence signals were required in these patients to fixate the target. We assumed that the relation between the internal

representation of the target position and the innervations of the eyes is just like for normal subjects. However, peripheral of the neural network that implements this relation, there is an offset on the position of one (asymmetric exodeviation) or both eyes (symmetric exodeviation). The central command signals thus operate in a rotated reference frame. To achieve binocular fixation despite the abnormal resting position(s) of the eye(s), the relation between horizontal target vergence and its neural representation is altered. For a symmetric exodeviation it would seem that the neural representation of horizontal target vergence would only need to increase in strength, to overcome the angle of exodeviation. Consequently, the cyclovergence (which is linearly related to the horizontal vergence in normal subjects) would also increase in these patients. Thus, we expect that if the horizontal exodeviation is symmetric (corresponding to a g-vector with only a z-component, see Fig. 1), the orientations of the s-plane and the g-plane are as in normal subjects, however, the relation

3272

A.V. VAN DEN BERG et al.

(b) D

.a

o

L

•

.3

FR

.25

.25

aD

.2

.2

'~llp

.15

qr.

Z

• 0.35 m o 1.43 rn

.05 0

0

-.05

U -.1

.1

q -.05

-.05

0

.05

.1

.15

.2

.25

sl

.3

t-.1

-.1

% -.05

0

.05

L~R~ .04J

.1

.15

.2

.25

.~1

Cycloversion

06

.03 04

.02 02

.01

~o

0

-.01

~oo

02

-.02i 04

-.03

r

•

• KS So

L~R~ -.04 -.04

-.03

-.02

-.01

0

gl

.01

.02

.03

.04

-.04

Cyclovergence

-.02

0

.02

.04

.06

gl

FIGURE 2(b). Caption on facing page.

between the measured cyclovergence (gl) and the horizontal vergence (v) of the eyes is changed to: g] = O(v - v~)/4

(2)

where '0' is the vertical component of the version required to fixate the target. 're' denotes the horizontal vergence deviation (positive ve corresponds to an angle of exodeviation). When the exodeviation is asymmetric, we find that the s-plane is rotated about the vertical axis towards the left (amblyopic) eye. For targets at 1.43 m, we found a leftward rotation of the plane of rotation vectors of the left eye by twice the amount of rotation for the s-plane, in the absence of such rotation for the non-amblyopic right eye. Our model for normal subjects (van Rijn & van den Berg, 1993) predicts that for an angle of horizontal convergence v the plane of rotation vectors of each eye turns laterally by v / 2 (and the primary direction of each eye by v). Thus, an increase of the convergence signal (as in the patient with the symmetric exodeviation) would in

the patients with an amblyopic eye turn the plane of rotation vectors of the left eye more to the left (by an angle of re~2) compared to the s-plane and would turn the displacement plane of the right eye to the right compared to the s-plane. Because we find that the plane of rotation vectors is fronto-parallel for the right eye, we conclude that the rotation of the s-plane equals one-half of the angle of exodeviation. The relation between the measured cyclovergence (gl) and the horizontal vergence of the eyes is then given by: g, = O(v -

0.5Ve)/4.

(3)

From the slopes of the observed gj vs the predicted gl ( = v O / 4 ~ 2s2g3) we computed estimates for ve for each fixation distance and patient. Results are presented in Table 4. The offset angle vo was positive in all patients, which corresponds to an offset angle of divergence [see Table l(b)]. In subject VL (with a symmetric exodeviation) clinical estimates of the phoria angle and the offset angle as estimated from the cyclovergence corresponded

EXCESS C Y C L O V E R G E N C E

3273

(c) D

.2 Oq

VIL

.2

)0

dP aD

.15

15

.1

.1

.05

E)5

~o

=oc

".OS

O

~h

35

#

• 0.35 m O 1.43 m

-.1

-.15

¢,0 kD

'.1

q

U -.2

.2

o=o

-.15

-. 1

-.05

0

.05

.1

sl

.15

.2

-.2

-.15

-.1

-.05

0

.05

.1

.15

.2

sl

Cycloversion

L~R~ .08

.04 .02

~3) On

•

..-x_~

~O

O 0

-.02

ee ~ o o

-,04

"" . ~ V ~ j . ,

-.06

-,08

-

"

-.02

.02

.o,

.o8

-.06

gl

-.64

-.ha

0 gl

.6a

.64

.6e

.08

Cyclovergence F I G U R E 2(c). F I G U R E 2. The version Is, upper panels of (a), (b) and (c)] and vergence [g, lower panels of (a), (b) and (c)] parts of the rotation vectors of the left and right eyes for three subjects (AM, F R and VL). The left panels show top views (y and x components, cf. Fig. 1) o f s and g, respectively. The right panels show side views o f s and g (z and x components). © indicate the components of the rotation vector when fixating in the distant plane (1.43 m), • symbols denote fixation in the near plane (0.35 m). The insets in each panel show the orientation of the head. Note that the points in each figure indicate the end point of a rotation vector. A vertical rotation vector for version (s 3 :/: 0 and s~ = s 2 = 0) corresponds to a horizontal gaze direction. Similarly, a horizontal rotation vector (s 2 # 0 and s t = s 3 = 0) corresponds to a vertical gaze direction. Labels along the axes indicate these gaze directions. Positive s~ denotes clockwise torsion [see Table l(a)]. Similarly, for the vergence component (g), labels along the axes indicate the type of gaze shift corresponding to the rotation vectors in the figure (Di, divergence; Con, convergence; L + R 1", left eye down, right eye up; L T R +, left eye up, right eye down). Positive cyclovergence denotes intorsion of both eyes. Values correspond to tan(~b/2) with ~ the angle of rotation about any one of the three head fixed rotation axes (x, y or z, cf. Fig. 1). Multiplication by 100 gives approximate values in degrees. It is clear that s is restricted to an approximately frontal plane, because s t is close to zero. Similarly, g is restricted to the sagittal plane since g2 is close to zero. These restrictions on the positions of s and g correspond to the findings in normal subjects that were reported earlier (van Rijn & van den Berg, 1993). Variability in s t in the present data is slightly larger. For clarity, the scales on the axes are different in the panel of g~ vs g3 in subject AM.

closely. In the patients with an amblyopic eye (AM and FR), it was assumed that the exodeviation was fully asymmetric. Thus, an estimate of ve based on the horizontal rotation of the s-plane was also made in the latter subjects (Table 4, third column). In subjects AM and FR there was variation in the estimates of the exodeviation,

and the angle of exodeviation was in some cases larger than the corresponding clinical estimates. Note that cyclovergence in normal subjects was about 10-20% lower than predicted. If the offset angle is computed using the observed gt in normals one would find 10 20% higher estimates for ve.

3274

A . V . VAN D E N BERG et al.

DISCUSSION Previously, we reported that eye torsion of normal subjects during binocular fixation of point targets in the dark is fully determined by the target's direction and depth. We presented and confirmed in five normal subjects a model that described the torsion of the two eyes as a function of the horizontal vergence and the vertical version required to fixate the target (van Rijn & van den Berg, 1993). Here, we show for three patients with intermittent exotropia, that basically the same relations hold. However, we find much larger cyclovergence than in normal subjects. In addition, we found in two patients with an amblyopic eye a rotation of Listing's plane towards the affected eye.

The constraints are conserved in these patients Our model for binocular fixation describes the positions of two eyes as a simple function of the direction in which the target is seen from a point midway between the eyes and the horizontal vergence required to fixate the target (Fig. 9 in van Rijn & van den Berg, 1993). This prediction for binocular eye position during fixation was based on the assumption that version and vergence add linearly, and that the eye positions are constrained in three ways: (1) The foveae of the two eyes are directed towards the target; (2) The version component follows Listing's law; (3) The vergence component is restricted similarly, i.e. horizontal, vertical and torsional vergence are not independent. Using rotation vectors to describe eye positions, the latter two constraints can be reformulated; the version and vergence components of the rotation vectors lie in two (perpendicular) planes. Our results show that these constraints also apply to fixation by patients with intermittent exotropia. The vertical vergence in Helmholtz coordinates, which directly evaluates the validity of the first constraint, was in these patients practically identical to that in normal subjects. Also, the version and vergence components of the eye positions were constrained to planes just as in normals although the thickness of the planes was slightly larger in the patients.

Peripheral offset on the eye position compensated by increased vergence effort Because the three constraints are supported by the data in both the normal subjects and the patients and because our model directly derives predictions for cyclo-

vergence from these constraints, it would seem that the cyclovergence responses should also be the same. What then causes the excess cyclovergence in the patients? We showed that the increased cyclovergence was compatible with an offset on the horizontal vergence. We stress, however, that the measured horizontal vergence of our patients was about equal to the target vergence and showed no offset. Thus, the cyclovergence in the patients corresponded to that for normal subjects if these had converged an additional amount equal to the offset. Apparently, the patients had to make a larger effort than normal subjects to fixate the same nearby target. We propose that in the patients the 'convergence effort' has to overcome the proximity of the target as well as the angle of exotropia, caused by a peripheral offset on the position of one or both eyes. It then follows, that the cyclovergence is increased even when the relation between horizontal vergence effort and cyclovergence effort is the same as in normals. Whereas the angle of exotropia balances the extra innervation for horizontal convergence, no such compensation occurs in the torsional direction, because the patients had no cyclotropia. Thus, we believe that the increased cyclovergence does not reflect changes in the input-output relations of the control structures involved in eye torsion. Rather it is the consequence of the increased vergence effort that these patients have to make to binocularly fixate nearby targets.

Why not peripheral compensation for the peripheral offset? If the cause for the departures from normal eye positions is indeed peripheral, one may ask why the oculomotor system chooses to compensate for its effects by changing the convergence effort, that is, a central command signal? One might think for instance, that the system would be better off if at the most peripheral level (say the motor neurons) an offset signal were generated, that causes an inward turn of the deviated eye, and keeping identical control signals at more central levels. One reason may be that the integrator is the cause of the deviation. Because there is no alternative route than through the oculomotor integrator to obtain innervation signals that cause a maintained eccentricity of the eye, no such compensation could be obtained. Another reason may have to do with the non-commutative nature of rotations. Simple addition of a peripheral offset signal may succeed to align the ocular axes for distant fixation straight ahead. However, to attain a tertiary position with aligned visual axes one cannot simply add the normal control signal for such a gaze shift and the offset

TABLE 3. Variation of s~ and g2 (rad), separated according to the fixation distance Fixation distance = 143 cm Patient AM FR VL

Fixation distance --- 35 cm

sI

gz

sl

g2

0.005 0.0069 0.0064

0.0016 0.0022 0.001

0.0053 0.0079 0.007

0.0014 0.0028 0.0012

Values represent SDs for each subject.

EXCESS CYCLOVERGENCE y = 1.341x - 3.456E-4, r 2 = .764

y = 4.434x + .001, r2 ffi .66 .015

.015

i

AM .01

0 1.43m

.005

.-/

.oi

AM • 0-35m

e /

0.i

.005

C

N

3275

0

IP

o

r e

S

-.00~ -.01 -.01

..015

-.015

•

-.02 -.025

I

-.02

-.0~ -.015 -.01 -.005

.005

.01

-.025 -.025

.016

. s . Q . -.02 -.015 -.01

.03

.

.02

FR

.

.

.

[ /

e D /

"

• 0.35 m

e

o01

O

0

.(X)5

.01

.015

y = 1.931x + .02, r 2 = ..572

y = 4 . 5 6 7 x + .008, r 2 = .45 .02

o,

-.005

•

.

.

.

/

e~///

-/.

-.oi

• -.01

/ e : ~

•

-.02

•.02

O~~eQeO

•

-.03

-.~

-.04

-.03

-.02

-.01

0

.01

-.04

.02

-.03

-.02

-.01

VL •

VL 0

II

1.43 rn

¢

-.01

-.02

-.03

-.04

-,03

-.02

-.01

0

.01

.02

.03

.04

-.06 .04

(

/../

0.35 m

.01

OI

.0~

y = 2.538x + .002, r 2 = .822

y = 11.189x + .01, r2 = .782 .0~

.02

.01

-.03

-.02

.01

-.01

0v/4

.02

.03

.04

0v/4

Predicted Cyclovergence FIGURE 3. The measured cyclovergence(g~) as a function of the predicted cyclovergence(2s2g3 which equals Ov/4when third order terms are neglected). Data of three subjects and two fixation distances are shown in separate panels. The best fitting linear relation is shown in each graph. Its parameters are indicated above each panel.

signal in the deviated eye, because of the n o n - c o m m u t a tive n a t u r e o f eye rotations. Thus, simple a d d i t i o n o f peripheral offset a n d n o r m a l c o n t r o l signals does n o t c o m p l y with bifoveal fixation.

Which eye is affected? I n the two a m b l y o p i c subjects we f o u n d a shift o f the c o m m o n p r i m a r y direction towards the a m b l y o p i c eye.

3276

A . V . V A N D E N BERG et al. T A B L E 4. Offset angles (deg) v~ on the horizontal vergence (v) that could account for the excess of cyclovergence (g~)

Patient AM FR VL

Exo. for fixation distance = 143 cm

Exo. for fixation distance = 34 cm

6.8 15.9 20.4

12.1 17.9 15.4

Clinical exodeviation Exo. from s-plane dist. = 250 cm dist. = 33 cm 10.4 8.1 --

7 17 17

10 7 17

ve was calculated from the slope of the measured gt vs predicted g~ in Fig. 3 [v, = - a , ( s l o p e - 1 , 0 ) * v(d)] for each distance (d) separately. For asymmetric exotropes a = 2, for symmetric exotropes a = 1. The right most columns indicate the angles of exophoria that were found clinically. These measures have been rounded to the nearest integer, to comply with the accuracy of the measurement. The third column indicates the estimate of the exodeviation from the rotation of the s-plane about the vertical axis.

From this common primary direction and the additional vergence effort [re in equation (2)] we can derive estimates for the primary direction of each eye. According to our model binocular fixation of targets in an isovergence surface results in rotation vectors that for each eye are located in a plane. These planes are parallel if vergence is absent; for convergence (v < 0), the primary direction of each eye rotates laterally by the same amount as the angle of vergence. Thus the effect of an additional convergence effort to counter the angle o f exotropia ve is to rotate the primary direction o f each eye temporally relative to the common primary direction (Ps) by ve deg. As shown in Table 5 in the amblyopes the primary direction of the amblyopic (left) eye is more deviating from the straight ahead position than the eye with normal vision. This is in correspondence with the finding that surgically induced esotropia in young monkeys causes amblyopia in the displaced eye, not in the normal eye (Kiorpes & Boothe, 1980). These authors also noted that right after the operation the monkey preferentially fixated with the unoperated eye causing amblyopia in the non-fixating eye. Retinal correspondence model and estimates of angle o f exotropia

Mok et al. (1992) presented a different model and a different set of normative data for the dependence of cyclovergence on horizontal vergence. These authors report cyclovergence that is about one-third to one-half as large as the normative data obtained from van Rijn

and van den Berg (1993). The difference in the data by Mok et al. and our data for normal subjects is not understood. Differences in stimulus presentation and data analysis could explain the different outcomes. What consequence would the use of this different normative data set have for the interpretation of our patient data? We showed before (van Rijn & van den Berg, 1993) that Mok's model corresponds to a different set of constraints for binocular control: (1) The foveae of the two eyes are directed towards the target; (2) The version component of control to maintain an eccentric eye position follows Listing's law, i.e. cycloversion, and horizontal and vertical version are not independent; (3) Cyclovergence optimizes the correspondence between the retinal images. It follows from these constraints that gj is related to elevation and horizontal vergence as: gl = Ov/8.

Using this different set of constraints we find a predicted g~ that is halved compared to equation (1) and in line with the data by Mok et al. (1992). Consequently, we would obtain different estimates for the angle of the offset vergence (CMok). Using equation (4) for predicted gj we would find regression coefficients that are twice as large as shown in Fig. 3. Thus, the offset angle would be related to the slope reported in Fig. 3 as: CMok = --a * (2 * slope -- 1) • v(d)

T A B L E 5. Azimuth components of the primary directions for the left and right eyes separately, calculated on the basis of the azimuth of the c o m m o n primary direction (Table 2, I st column) and the offset ve (Table 4, average of 1st and 2nd column) Patient AM FR VL

Left eye (deg)

Right eye (deg)

Amblyopia

4-19.9 +25 4- 16.3

4- l -8.8 - 19.5

L L None

Positive numbers indicate leftward rotations. These computed primary directions were found to deviate less than 4 d e g from the estimates of the primary direction based on the rotation vectors o f each eye separately.

(4)

(5)

where v(d) denotes the horizontal vergence at fixation distance ' d ' (negative for convergence), and a = 1 for symmetric exodeviation and a = 2 for asymmetric exodeviation. Table 6 compares the estimates for the offset angle based on the data by Mok et al. (1992) with the clinical angles of exotropia. Clearly, the offset angles are now much larger and in some cases the estimate is unrealistically large compared to the clinical data. Apparently, for the interpretation of the patient data our model constraints result in estimates of the angle of exotropia that are more consistent with the clinical data. We cannot exclude the possibility, though, that using the different procedures of Mok et al. our patients might

EXCESS CYCLOVERGENCE

3277

TABLE 6. Offset angles (deg) CMok on the horizontal vergence (v) that could account for the excess of cyclovergence (gl) Clinical phoria Patient AM FR VL

CMok for fixation

CMok for fixation

distance = 143 cm

distance = 34 cm

distance = 250 cm

distance = 33 cm

27.4 41.2 42.6

33.7 50.8 40.6

7 17 17

10 7 17

CMokwas calculated from the slope of the measured gl VSpredicted gl ( : OV/8) as would apply to Mok et al.'s data. The right most columns indicate the angles of exophoria that were found clinically. have shown smaller angles o f cyclovergence t h e r e b y lowering the estimates for CMok. A n y w a y , we can c o n c l u d e t h a t the cyclovergence in o u r p a t i e n t s is m u c h larger t h a n w o u l d be o p t i m a l for retinal c o r r e s p o n d e n c e . The i m p o r t a n c e o f t o r s i o n a l c o r r e s p o n d e n c e was recently a d v o c a t e d by van Rijn, van der Steen a n d Collewijn (1994). These a u t h o r s d e m o n s t r a t e d in n o r m a l subjects a m u c h smaller s p o n t a n e o u s v a r i a t i o n o f cyclovergence t h a n o f cycloversion. T h e difference was m o r e p r o n o u n c e d in the presence o f a large s t r u c t u r e d b a c k g r o u n d pattern. The r e d u c e d torsional c o r r e s p o n d e n c e in o u r p a t i e n t s c o u l d be an extra f a c t o r c o m p l i c a t i n g b i n o c u l a r single vision. Therapeutical implications

W e derived an estimate o f the angle o f e x o t r o p i a that is i n d e p e n d e n t f r o m the clinical measures. O u r m e t h o d extends the clinical estimates in t h a t it p r o v i d e s a w a y to distinguish between s y m m e t r i c a n d a s y m m e t r i c deviations. M o r e o v e r , it is based on the p a t t e r n o f d e v i a t i o n s in the eye t o r s i o n d u r i n g b i n o c u l a r fixation s t r a i g h t a h e a d a n d d u r i n g eccentric fixation. Thus, one need n o t a s s u m e as in the clinical m e t h o d s that the covered eye has a t t a i n e d its resting position. In a d d i t i o n , because the estimate is b a s e d on several different fixation p o s i t i o n s it p r o v i d e s an angle o f e x o d e v i a t i o n t h a t is characteristic o f a large p a r t o f the o c u l o - m o t o r range. O u r e x p l a n a t i o n for the increased cyclovergence in the p a t i e n t s with i n t e r m i t t e n t e x o t r o p i a p r o p o s e s t h a t the d e v i a t i o n o f one o r b o t h eyes has a p e r i p h e r a l origin (for e x a m p l e , a bias in the o c u l o m o t o r i n t e g r a t o r in the b r a i n s t e m o r d e v i a t i o n s in the muscle insertions on the globe), which is c o m p e n s a t e d b y increased vergence effort. This gives a r a t i o n a l e for surgical t h e r a p y , which c h a n g e s this p e r i p h e r a l offset. W e f o u n d in two p a t i e n t s that the i n t e r m i t t e n t e x o t r o p i a was related to a d e v i a t i o n o f one eye only. In case o f these a m b l y o p i c p a t i e n t s the p r i m a r y p o s i t i o n o f the a m b l y o p i c eye was e x o r o t a t e d most. This s u p p o r t s m o n o c u l a r surgery on the a m b l y opic eye. W e f o u n d t h a t the vergence-offset m a y differ s o m e w h a t f r o m that d e t e r m i n e d in the clinical setting. O u r t e c h n i q u e m a y s u p p l y a m o r e objective m e a n s for m e a s u r i n g the squint angle t h a t one needs to c o m p e n s a t e surgically. N o t i c e , t h a t the h o r i z o n t a l vergence offset t h a t we d e t e r m i n e f r o m the excess cyclovergence corres p o n d s to a r o t a t i o n o f the p r i m a r y p o s i t i o n o f each eye o f the same a m o u n t . Thus, in case o f s y m m e t r i c exo t r o p i a one s h o u l d correct a d e v i a t i o n that, in t o t a l is

twice as large as the e s t i m a t e d angle o f exodeviation. Also, in a s y m m e t r i c e x o t r o p i a one s h o u l d correct one eye b y twice the a m o u n t o f m e a s u r e d exodeviation. O u r e x p l a n a t i o n for the excess cyclovergence carries the p r o m i s e that surgical c o m p e n s a t i o n o f the deviation(s) o f the p r i m a r y position(s) m a y at the same time i m p r o v e the t o r s i o n a l c o r r e s p o n d e n c e , by r e d u c t i o n o f the ' c o n v e r g e n c e effort'.

REFERENCES

Balliet, R. & Nakayama, K. (1978). Training of voluntary torsion. Investigative ophthalmology & Visual Science, 17, 303-314. Burke, M. J. (1985). Intermittent exotropia. In Nelson, L. B. & Wagner, R. S. (Eds), Strabismus surgery. International ophthalmology clinics (Vol. 25, pp. 53-68). Boston, Mass.: Little, Brown & Co. Collewijn, H., van der Steen, J., Ferman, L. & Jansen, T. C. (1985). Human ocular counterroll: Assessment of static and dynamic properties from electromagnetic scleral coil, recordings. Experimental Brain Research, 59, 185-196. Duane, A. (1896). A new classification of the motor-anomalies of the eyes based upon physiological principles, together with their symptoms, diagnosis and treatment. American Journal of Ophthalmology and Otolaryngology, 5, 969. Ferman, L., Collewijn, H. & van den Berg, A. V. (1987a). A direct test of Listing's law--I. Human ocular torsion measured in static tertiairy positions. Vision Research, 27, 929-938. Ferman, L., Collewijn, H. & van den Berg, A. V. (1987b). A direct test of Listing's law--lI. Human ocular torsion measured under dynamic conditions. Vision Research, 27, 939 951. Ferman, L., Collewijn, H., Jansen T. C. & van den Berg, A. V. (1987). Human gaze stability in the horizontal, vertical and torsional direction during voluntary head movements, evaluated with a three-dimensional scleral induction coil technique. Vision Research, 27, 8ll-828. Glenn, B. & Vilis, T. (1991). Violations of Listing's law after large eye and head gaze shifts. Journal of Neurophysiology, 68, 309-318. Haslwanter, Th., Straumann, D., Hess, B. J. M. & Henn, V. (1992). Static roll and pitch in the monkey: Shift and rotation of Listing's plane. Vision Research, 32, 1341-1348. Haslwanter, Th., Straumann, D., Hepp, K., Hess, B. J. M. & Henn, V. (1991). Smooth pursuit eye movements obey Listing's law in the monkey. Experimental Brain Research, 87, 470-472. Haustein, W. (1989). Considerations on Listing's law and the primary position by means of a matrix description of eye position control. Biological Cybernetics, 60, 411-420. von Helmholtz, H. (1867). In Handbuch der physiologischen optik (lst edn, Vol. 3). Translated into English as Treatise of physiological optics by Southall, J. P. C. (1962). New York: Dover. Howard, I. P. (1982). Human visual orientation. Chichester: Wiley. Kiorpes, L. & Boothe, R. G. (1980). The time course for the development of strabismic amblyopia in infant monkeys (Macaca nemestrina). Investigative Ophthalmology & Visual Science, 19, 841 845.

3278

A.V. VAN DEN BERG et al.

Minken, A. W. H. & Van Gisbergen, J. A. M. (1994) A threedimensional analysis of vergence movements at various levels of elevation. Experimental Brain Research 101, 331-345. Minken, A. W. H., van Opstal, A. J. & van Gisbergen, J. A. M. (1993). Three-dimensional analysis of strongly curved saccades elicited by double-step stimuli. Experimental Brain Research, 93, 521-533. Mok, D., Ro, A., Cadera, W., Crawford, J. D. & Vilis, T. (1992). Rotataion of Listing's plane during vergence. Vision Research, 32, 2055 2064. Nakayama, K. (1975) Photographic determination of the rotational state of the eye using matrices. American Journal of Optometry and Physiological Optics, 51, 736 742. Nakayama, K. (1983). Diagnosis and treatment of strabismus. In Schor, C. M. & Ciuffreda, K. J. (Eds), Vergence eye mot,ements: Basic and clinical aspects (pp. 541-579). Boston, Mass.: Butterworths. van Rijn, L. J., van der Steen, J. & Collewijn, H. (1994). Instability of ocular torsion during fixation: Cyclovergence is more stable than cycloversion. Vision Research, 34, 1077-1087. van Rijn, L. J. & van den Berg, A. V. (1993) Binocular eye orientationduring fixations: Listing's law extended to include eye vergence. Vision Research, 33, 691-708.

van Rijn, L. J., van den Berg, A. V. & De Faber, J-Tj. H. N. (1993) Intermittent exotropia can be interpreted as peripheral offset on eye position and unchanged central control. Investigative Ophthalmology and Visual Science (Suppl.), 34, 1309. van Rijn, L. J., van der Steen, J. & Collewijn, H. (1992). Visually induced cycloversion and cyclovergence. Vision Research, 32, 1875-1883. Tweed, D. & Vilis, T. (1990). Geometric relations of eye position and velocity vectors during saccades. Vision Research, 30, 111-127. Tweed, D., Cadera, W. & Vilis, T. (1990). Computing three-dimensional eye position quaternions and eye velocity from search coil signals. Vision Research, 30, 97 110. Tweed, D., Fetter, M., Andreadaki, S., Koenig, E. & Dichgans, J. (1992). Three-dimensional properties of human pursuit eye movements. Vision Research, 32, 1225-1238. Von Noorden, G. K. (1990). Burian-Von Noorden's binocular vision and ocular motility. Theory and management o[ strabismus (pp. 323-339). St Louis: Mosby.

Acknowledgements--We gratefully acknowledge the helpful comments of Drs T Vilis and D. Tweed on a draft of this paper.