Seesaw nystagmus associated with involuntary torsional head oscillations

Seesaw nystagmus (SSN) is a unique, rare binocular disorder characterized by alternating vertical skew deviation and conjugate ocular torsion. While the upper eye intorts, the lower eye extorts. The more common jerk SSN consists of torsional slow phases in one direction (clockwise [cw] or counterclockwise [ccw]) and quick phases in the opposite direction. In contrast, pendular SSN shows slow, smooth eye oscillations in the roll plane.¹ This classification may be of diagnostic vale because both types of SSN probably have different causes and lesion sites, although their pathomechanisms are not yet clear.^2-4 The following report focuses on pendular SSN.

Pendular SSN has frequently been found to be associated with visual disorders^2,5,6 and may also be of congenital origin.^7-9 The visual cues that initiate or stop pendular nystagmus are not known. Moreover, oscillating head movements, which are associated with other forms of pendular (horizontal, vertical) nystagmus,^10,11 have not been described in SSN. In horizontal pendular nystagmus, eye and head movements have been analyzed but their phase relationship is still not clear.^10,11 To determine the functional role and the clinical significance of torsional head oscillations we examined a patient with a pronounced pendular SSN with large torsional amplitudes and clinically visible, synchronous pendular torsional head movements. We addressed two main questions: 1) Are head movements initiated to compensate for SSN and oscillopsia? 2) What is the pathophysiologic mechanisms of pendular SSN?

Patient report. Our patient, a 36-year-old woman, had a history of intermittent vertical and torsional oscillopsia since 1985. Since early childhood the patient had myopia, which was corrected with glasses of -17 diopters (left eye) and -20 diopters (right eye), exotropia with an "A" pattern (seven prism diopters), and a horizontal jerk nystagmus (HN) to the left. After a motorbike accident with a minor head trauma in 1975, she complained of diplopia, and a head tilt to the left was noticed. MRI did not reveal any cerebral damage. A strabismus operation (Kerstenbaum procedure) was performed in 1983 to compensate for the horizontal head tilt. Subsequently, she did not complain of diplopia until 1985, when a pendular SSN was noticed for the first time.

Medical therapy with various drugs (baclofen; pyridostigmine; and tiaprid, an anticholinergic drug used in the therapy of dyskinesias) had no effect on SSN or the horizontal nystagmus. Otherwise the patient's medical history was unremarkable.

She consulted our clinic in 1997 because of intermittent oscillopsia. Neurologic examination revealed HN, which switched into a pendular SSN on attempted near fixation and convergence. SSN changed to HN on attempted fixation of a distant target, during smooth-pursuit eye movements, or during visually evoked saccades. The SSN and the HN never appeared together. There was synchronous torsional head oscillation associated with the SSN. The patient complained of vertical and torsional oscillopsia only during the periods of SSN. The details of eye and head movements are provided later.

No other abnormalities were revealed on neurologic examination, except for corrected vision of 20/100 in the right eye and 20/60 in the left eye, and a bilateral hypacusis. Serial cranial MRI was also normal. In addition, high-resolution MRI (Siemens Vision, Erlangen, Germany, 1.5 T) with a slice thickness of 3 mm of the rostral brainstem failed to detect any lesion in the mesencephalon or optic chiasm. Visually evoked potentials and visual fields on the perimetry (Goldmann) were normal. Laboratory examination (blood count, liver enzymes, kidney function, endocrine function, and spinal fluid) were all within normal ranges.

Methods. Eye and head movements were recorded with the Skalar search-coil system with two orthogonal magnetic fields. Each eye was calibrated separately, and the head was calibrated by a special device to account for nonlinearities in the magnetic field. The calibration procedure included an iteration algorithm that depends only on the fixation of the straight-ahead gaze position and the performance of spontaneous eye movements.¹² Head movements were calibrated using the same procedure as for the eyes. After the patient had given her informed consent, coils (Skalar) specifically designed to record horizontal, vertical, and torsional eye movements were placed under topical anesthesia in each eye and on the forehead. Eye and head movements were recorded with a 12-bit analog-digital converter (DAP 1200) at a sampling rate of 1,000 Hz.

Under one condition the head was fixed with a head-holding device that immobilized the forehead and chin in the center of the magnetic field(head-fixed condition). Under the second condition, head and eye movements were recorded with the head unrestrained (head-free condition).

During the head-fixed condition, eye movements were recorded using the following paradigms: 1) spontaneous eye movements in light and dark, 2) monocular and binocular fixation of a laser spot (0.1-deg diameter) on a white tangent screen 120 or 30 cm distant from the patient's eyes, and 3) vertical and horizontal smooth-pursuit eye movements following the laser target with a horizontal and vertical amplitude of ±20 deg (0.3 Hz). Vertical and horizontal optokinetic nystagmus was elicited with a moving black-and-white stripe pattern (30 deg/sec) projected on the tangent screen. For near fixation the tangent screen was moved to the patient.

During the head-free condition, the task was 1) to fixate the laser target without moving the head, 2) to move the head actively, or 3) to fixate the laser target while the investigator moved the patient's head. For vestibular evaluation the head was moved passively and sinusoidally (0.3 Hz, ±40 deg) in the pitch, roll, and yaw planes in the dark. All measurements, recordings, and calibrations were performed while the patient was wearing her glasses to ensure accurate fixation.

Eye positions were measured and calculated as quaternions¹³ with respect to the straight-ahead gaze position. For greater clarity, the quaternions were expressed in degrees, and coordinates were transformed into the following coordinate system (eye position components: x = torsion [+, -], y = vertical [up, +; down, -], z = horizontal [right, +; left, -]). Torsion was defined as positive with cw rotation of the right eye (excyclorotation) and as negative with ccw rotation of the same eye (incyclotorsion, from the point of view of the subject). In the head-free condition, the eye-in-head quaternion (e) was calculated from eye-in-space (gaze, p) and the inverse head-in-space quaternion (q^-1) with the quaternion product (e = pq^-1).¹³

Eye and head velocities were calculated as angular velocities from the positional data (ω = 2*q′q^-1) where q is the quaternion, q′ is a derivative of q, q^-1 is an inverse quaternion of q, and ω is the angular velocity.¹⁴ We measured actual eye position in the head independent of visual fixation.¹² Therefore, no correction was necessary for the high refraction of her glasses.

Amplitudes of the pendular eye and head movements were calculated from peak to peak. Frequency was analyzed using fast Fournier transformation and least squares fit with the equation y = k1 + k2⁸sin(k3*× + k4), where k1 through k4 are constants of eye and head velocity. Vestibulo-ocular reflex (VOR) gain was calculated from the maximal velocities of head-in-space and eye-in-head at the point of maximal gaze velocity. Positive phase values indicate a phase lag of eye-to-head velocity, and negative values indicate a phase lead. The x, y, and z coordinates of eye velocity were used as the rotational axis of the eye during SSN.

Conventional electro-oculography was used for caloric irrigation and horizontal constant-velocity vestibular stimulation.¹⁵ Fundus photography was performed to evaluate the tonic effects on torsional eye position.¹⁶

Results. Eye movement recordings during the head-fixed condition. Oculomotor performance. Eye movement recordings at intervals of HN showed a normal oculomotor range, and accurate vertical and horizontal saccades with normal velocities. Response to caloric irrigation and horizontal constant-velocity vestibular stimulation revealed normal vestibular function. Horizontal but not vertical optokinetic stimulation(OKN) showed an inversion of the optokinetic nystagmus.¹⁷ Vertical and horizontal sinusoidal smooth pursuit was superimposed by HN but was normal in gain (0.9). During SSN only a few horizontal quick phases were noticed. No other eye movements were possible during SSN.

Positional maneuvers (head movements up, down, to the left and right, and head tilt with the right ear down or the left ear down) and head shaking with maintained fixation did not change the frequency and amplitude of HN, but SSN stopped immediately.

Search coil recordings showed a normal vertical VOR gain (0.7) during passive head rotations at 0.3 Hz (±40 deg). In contrast, torsional VOR during passive sinusoidal head movements showed a decreased gain (0.24). VOR could not be tested during SSN, which disappeared during active or passive head movements or during an oculomotor task (OKN, smooth-pursuit eye movements, VOR, and visually guided or spontaneous saccades). Fundus photography did not show any tonic torsional deviation of the eyes during HN.

Three-dimensional recordings of nystagmus. Horizontal jerk nystagmus. Search coil recordings in the light showed a binocular, conjugated HN to the left(figure 1) with exponentially increasing slow phase velocities, a frequency of 4 to 5 Hz, and a horizontal amplitude of 6.1± 1.5 deg on average. The amplitude was slightly larger in the left eye (0.8 deg mean difference). The vertical and torsional components of the HN were smaller, with an average torsional amplitude of 3.0 ± 0.5 deg and an average vertical amplitude of 2.0 ± 0.4 deg. HN increased on left gaze and persisted in darkness with the eyes open. There was a slight change in its characteristics but there still existed an exponentially increasing slow phase. The nystagmus now appeared to be more regular with a more linear slow phase. Due to the exponentially increasing slow phase, the change in nystagmus characteristics in the dark, and the reversal of the horizontal OKN, this HN was classified as congenital HN.¹⁸

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Seesaw nystagmus (SSN) in the head-fixed condition. Binocular search coil recording of torsional (x), vertical (y), and horizontal (z) eye positions in the light. The right eye(black line) and the left eye (gray line) show conjugate torsional eye movements but disconjugate horizontal and vertical eye movements. The large vertical dotted line serves as a reference for the simultaneous comparison of eye positions. Arrows indicate the beginning (black arrowhead) and the end(open arrowhead) of pendular SSN. Before and after the period of SSN there is a left-beating horizontal jerk nystagmus, with a few small jerks during SSN.

Seesaw nystagmus. Horizontal nystagmus always switched to SSN on attempted near vision or convergence (seefigure 1, black arrow). The conversion of SSN back to HN was seen in the head-fixed condition during different oculomotor tasks, such as OKN or spontaneous or visually guided saccades. The transposition from HN to SSN was prompt, but there was still a minor increase in torsional amplitude over a few seconds.

While the left eye (see figure 1, gray line) was elevated, the right eye (see figure 1, black line) was depressed, but both eye movements had the same torsional direction (cw; see figure 1, dotted line, for reference). The frequency of SSN on binocular fixation, which was slightly higher at the beginning of an SSN period, was 0,6 Hz, with a mean torsional amplitude (maximum to minimum) of 21 deg (right eye, 18.9 ± 7.2 deg), an average vertical amplitude of 6.3 deg (right eye, 6.8 ± 2.3 deg; left eye, 6.0± 1.9 deg), and an average horizontal amplitude of 2.1 deg (right eye, 3.5 ± 1.6 deg; left eye, 0.8 ± 0.7 deg). Although the vertical component was smaller for the left eye, its torsional component was larger than that for the right eye (see figure 1).

Near vision decreased the exotropia from about 8.0 ± 0.3 deg in the straight-ahead gaze position during fixation of the target at a 120-cm distance to 4.0 ± 0.4 deg at a 30-cm distance. During SSN there were a few horizontal quick phases (see figure 1).

The conversion from HN (see figure 1, black arrow) to SSN was accompanied by a mean change in vertical (3 deg, right eye down, left eye up) and torsional (10-deg, negative direction) eye position. Horizontal eye position shifted immediately with the beginning of SSN to the left side, in particular in the left eye, thus increasing the disparity between both eyes. Pendular SSN was also dependent on eye position. This was visible in a horizontal position only up to 6 deg to the left from the straight-ahead gaze position, but in various vertical positions (-20 to 20 deg).

To investigate the visual influence on SSN we performed a monocular eye cover test. On monocular left-eye fixation there was no change in SNN from that of binocular vision (figure 2, A and B). In contrast, on monocular right-eye fixation (figure 2C) the torsional amplitude was diminished from 18.9 ± 6.0 deg to 12.0± 2.2 deg for the right eye and from 22.1 ± 7.2 deg to 14.3± 2.9 deg for the left eye. The SSN frequency of 0.57 Hz on binocular fixation increased to 1.10 Hz on right-eye fixation (seefigure 2C). The vertical amplitude of the right eye was diminished strongly, from 6.8 ± deg to 1.4 ± 0.4 deg, during right-eye fixation. In addition, the horizontal eye position shifted to the left (see figure 2C) by 10 deg.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Seesaw nystagmus (SSN) changes with monocular and binocular fixations. Torsional (x), vertical (y), and horizontal (z) eye positions, displayed as eye rotation vectors in degrees, for the right eye (black line) and left eye (gray line) are shown. Monocular fixation of the left eye (A) does not change the amplitude and frequency of SNN compared with binocular fixation (B), whereas monocular fixation of the right eye (C) decreases the vertical amplitude asymmetrically and increases the frequency. Also, torsional amplitudes are decreased and the frequency is increased. There is a horizontal positional shift to the left for both eyes in panel C.

Axis of rotation vectors during periods of SSN are shown in the torsional/vertical (figure 3, A and C) and torsional/horizontal (figure 3, B and D) projections of the eye velocity components of each eye. In addition, the rotational vectors for the straight-ahead gaze of each vertical and torsional eye muscle(superior rectus [SR], superior oblique [SO], inferior rectus [IR], inferior oblique [IO]), calculated from orbital geometry,¹⁹ are displayed in relation to the sum of the rotational vectors of paired vertical and oblique eye muscles (SR + SO and IR + IO). The rotational axis of the vertical/torsional eye movement components aligned best with the vectors of the rotational axis of the intorters (SR + SO) and extorters (IR + IO; see figure 3), provided there is a coactivation of the rectus and oblique eye muscles by the same amount. The partial misalignment of muscle rotational vectors in the z component (figure 3, B and D) is discussed later.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Eye rotation and muscle rotation vectors. Torsional (VELX), vertical (VELY), and horizontal (VELZ) eye velocity of seesaw nystagmus (SSN) is displayed for the left (A,B) and right(C, D) eyes to show rotational vectors of eye movements during SSN. The single loops (some marked with stars) show quick phases during SSN. The muscle rotation vectors in the straight-ahead gaze-calculated from orbital geometry¹⁹ (dotted lines) for the superior rectus (SR), superior oblique (SO), inferior rectus (IR), and inferior oblique (IO) muscles-as well as the sum (black lines) of SR + SO and IO + IR are plotted. The rotational axis of SSN does not coincide with any direction of a single vertical or oblique muscle. It aligns best with the direction of the coactivated SR + SO and IR + IO rotational axis for both eyes (A, C). Horizontal eye rotation vectors of SSN (B, D) deviate slightly from the predicted sum muscle rotational vectors, especially with negative torsion. The sign of the rotational vector of the inferior oblique eye muscle¹⁹ was corrected by multiplying the z component with -1 to achieve the expected abducting component, as described for human eye muscles.²⁰ cw = clockwise; cw = counterclockwise.

Eye and head movement recordings during the head-free condition. At the onset of SSN under the head-free condition, the patient developed slow, pendular torsional head movements in phase with the torsional eye movements(figure 4A). In contrast, there were no involuntary head movements during HN (figure 4B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Search coil recording of eye and head movements. The left eye-in-head position (gray) and the head-in-space position (black) are displayed as torsional (x), vertical (y), and horizontal(z) positional components during (A) seesaw nystagmus (SSN) and (B) horizontal jerk nystagmus (HN). Eye and head movements during SSN are plotted using different scales. The torsional head position is in phase with the torsional eye position. There is a 40-deg phase advance of the eye compared with the head position. The vertical line serves as a reference for simultaneous comparison of eye and head position. In contrast, there is no head oscillation during HN (B).

With respect to the head-in-space and the eye-in-head movements, the torsional head movement component was 2.0 ± 0.7 deg on average (see figure 4A). Vertical (0.6 ± 0.4 deg) and horizontal (0.5 ± 0.3 deg) head movement components were small. Thus, torsional head movements were 10 times smaller than the torsional eye movement components. The phase relation of the torsional head oscillation to the torsional components of the SSN was stable. Eye velocity (position) led head velocity (position) by an average of -40 ± 3 deg(see figure 4A). In some cycles of pendular head oscillation, the head movement decreased its velocity, while the eye reached its maximum or minimum in position (see figure 4). However, this was not consistent in all cycles (e.g., for the left eye, the vertical eye and head velocities [positions] were also in phase).

Discussion. In this first reported study using three-dimensional recordings to examine binocular eye and head movements in pendular SSN, we found that torsional components of eye and head oscillations moved in phase with each other. Thus, because head movements are not compensatory for SSN, common oscillators are most probable. Several lines of evidence point to a congenital origin for SSN: 1) the patient's history of HN since childhood with patterns of congenital nystagmus (exponential increase of slow-phase velocity, inversion of optokinetic nystagmus), 2) the conversion from HN to SSN and vice versa, and 3) the absence of structural lesions on cranial high-resolution MRI. The conversion of HN to SSN, which has not been reported previously, might also suggest a common origin. However, the fact that torsional head movements occurred only during SSN but not during HN indicates a specific disorder of torsional eye-head coupling.

Abnormalities of eye-head coordination. Involuntary torsional head movements associated with SSN have not been reported previously. Horizontal or vertical oscillatory head movements were seen in acquired horizontal or vertical pendular nystagmus in children with impaired vision, but head movements were not related specifically to eye movements, in particular not with respect to their phase relationship and thus to their function.¹¹ If they were compensatory for the eye movements, head movements would be expected to occur in the same specific, vestibular rotational plane. Due to the synchronous torsional eye and head movements in the same direction and the low torsional VOR gain, the patient's gaze could not be stabilized to achieve better fixation. SSN ceased during active and passive head movements, although this may have been due to changes in the vergence angle, which we were not able to monitor during these tests.

Possible etiologies of SSN. Jerk and pendular SSN must be distinguished clearly because they may be caused by different lesion sites. Whereas bilateral mesodiencephalic lesions are thought to elicit pendular SSN,^21,22 unilateral upper brainstem pathology can cause jerk SSN.^4,23-25

Possible brainstem lesion sites for jerk SSN include the medial longitudinal fasciculus (MLF)³ and the interstitial nucleus of Cajal (iC).^24,25 Recently, a unilateral lesion of iC with sparing of the torsional fast-phase generator, the adjacent rostral MLF (riMLF), was proposed to elicit jerk SSN.⁴ However, neither jerk nor pendular SSN was found in experimental unilateral iC lesions in the monkey, which were confined to the borders of iC.²⁶

Pendular SSN appears to require an intact iC to develop. This is consistent with the observation that electric stimulation of iC exacerbated pendular SSN in a patient, whereas destruction of the nucleus abolished it.²⁷ Furthermore, using high-resolution MRI we did not find a structural lesion in the brainstem, in particular in the midbrain of our patient. Her oculomotor performance did not show signs of iC or riMLF lesions.^24,28,29 Therefore, an intact iC, having afferent and efferent pathways to the vestibular nuclei,^28,30 is probably essential for the development and maintenance of pendular SSN.

In contrast to jerk SSN, only pendular SSN may be associated with visual disorders.^5,9,31 Pendular SSN was found to be associated with tumors in the region of the optic chiasm or thalamus,⁶ pretectum,³² and aberrant crossing fibers in the chiasm.^2,33 Recently, pendular SSN has been found to develop in later life due to progressive vision loss in the absence of any rostral brainstem lesion.⁵ A delay in the visual feedback, which normally helps to stabilize gaze, is capable of inducing pendular nystagmus up to 2 Hz, even in normal subjects.³⁴ Our patient's vision was poor, and there is some evidence that her SSN (average frequencies < 1 Hz) was influenced by visual signals. The frequency of the SSN increased, and the torsional and vertical amplitude decreased on fixation with the nondominant right eye. Prolonged visual feedback may have contributed to the generation of the low-frequency oscillating signal causing SSN.

HN changed to SSN in this patient only when she viewed a near target or on attempted convergence, and switched back to HN during active or passive eye and head movements. SSN was also dependent or horizontal eye position because it was elicited only in specific horizontal eye positions, independent of eye elevation. Near vision increased the probability of inducing SSN, which suggests an input from the vergence system.

SSN and ocular tilt reaction (OTR). SSN shares certain characteristic oculomotor features with OTR. OTR is a well-known disorder of torsional eye-head coordination consisting of the synkinesis of vertical skew deviation, ocular torsion, and a noncompensatory torsional head tilt in the same direction.^1,35 Thus, both disorders (OTR and SSN) show vertical misalignment of the visual axes and a pathologic conjugate ocular torsion in the direction of the tonic head tilt or torsional head movements. OTR may be tonic or paroxysmal,^4,36,37 or may exhibit a slowly alternating skew deviation.^32,38,39 Jerk SSN and OTR may occur concomitantly in patients with midbrain lesions involving the iC,⁴ a center of eye-head coordination.⁴⁰ Therefore, a common pathomechanism of jerk SSN and OTR has been proposed. However, these combined disorders of torsional eye-head coordination are not specific. Electric stimulation^41,42 not only in the region of iC elicits ipsilesional (or, in the case of lesions, contralesional⁴³) OTR, but it also causes similar phenomena when applied to the contralateral utricular nerve.⁴⁴ Therefore, analogous to tonic OTR, which is thought to be caused by an imbalance in the otolithic-ocular reflex¹ or the vertical semicircular canal inputs,³⁵ SSN may reflect a sinusoidal oscillation involving central vestibular (otolithic) pathways.¹ These assumptions would lead us to expect torsional head oscillations in SSN. Accordingly, head and eye movements in our patient were not compensatory, but moved in the same direction, as in the case of OTR. Therefore, torsional head oscillations may be generated by the same mechanism-a common oscillating signal.

A common oscillator for eye and head movements-a possible mechanism. Eye and head oscillations appear to be elicited by a common oscillator. Evidence for this hypothesis rests on several features. First, both rotations had the same frequency. Second, the phase relation (40 deg) between eye and head rotations was constant throughout the recording. Third, head oscillation did not produce a compensatory VOR to stabilize gaze during SSN. Fourth, torsional head oscillations were associated specifically with SSN, because the head was stationary during the periods of HN.

Unanswered questions that remain are, for example, how does a common oscillator account for the vertical disparity in SSN? In other words, how does dynamic vertical divergence of both eyes arise from one oscillating signal? The dynamic vertical disparity partially resembles tonic skew deviation as an essential part of OTR in midbrain lesions.⁴² However, dynamic vertical divergence has not been reported in SSN with midbrain lesions in which SSN was associated with tonic or paroxysmal OTR.⁴

Skew deviation and SSN. Skew deviation may reflect a phylogenetically older but physiologic response to lateral head tilt in lateral-eyed animals that was needed to stabilize gaze via vestibular pathways called "primary vestibular pathways."^45,46 The existence of additional "secondary" vestibular pathways corrects VOR in frontal-eyed animals and allows compensatory eye movements in the canal planes.⁴⁴ Primary vestibular pathways for vertical and torsional VOR are connected directly to the motoneurons of the vertical and oblique eye muscles. Lateral (torsional) head tilt usually leads to an equal activation of both intorters (SR and SO) on the one side and of both extorters (IR and IO) on the other.⁴⁵ If SSN is due to a sinusoidal oscillation in these primary vestibular afferents, then the rotational axis of the eyes in SSN should be similar to the main rotational axis of the equally coactivated eye muscles. This hypothesis is consistent with the findings in our patient. The mean axis of rotational vectors of SSN showed a good correlation with the axis of the sum of the rotational vectors¹⁹ of the coactivated eye muscles (SR/SO and IR/IO) for both eyes (seefigure 3). Thus, SSN in this patient is probably caused by an alternating equal activation of both extorting (IR/IO) and intorting(SR/SO) muscle pairs on either side, which is consistent with the hypothesis of Zee.⁴⁵

The small disparity in the z component between the sum of the eye rotational vector and SSN could be explained in two ways. First, the muscle rotational vectors are calculated, based on orbital geometry, for straight-ahead gaze and are fixed to the orbit during eye movements.¹⁹ Our SSN is not exactly in the straight-ahead gaze position. Thus, a position-dependent change of the rotational axis is not addressed, which could produce an additional horizontal component (seefigure 3). Second, it is possible that there is an additional activation of horizontal muscles during SSN. Therefore, a sinusoidal oscillating signal activating the primary vestibular pathways for vertical VOR could produce not only pendular torsional components of SSN but also alternating vertical disparity. How this oscillator affects head oscillation remains unclear, but the lesional effects of iC inactivation on torsional head and eye position suggest an involvement of an intact iC. According to our hypothesis, pendular SSN should always be associated with some torsional head oscillation. It was the unusually large amplitude (>20 deg) of torsional eye movements during SSN in this patient that made head movements clinically visible, although it was only 10% of the torsional eye movement amplitude. This may also explain why head movements in SSN with eye movement amplitudes <10 deg⁴ are not reported.

Acknowledgment

The authors thank Mrs. J. Benson for reading the manuscript.