Overlap of Infantile Nystagmus

and Strabismus

Estimates of the prevalence of strabismus in infantile nystagmus range from 8% to 33%.81,146,157,205,213 Strabismus is essential for latent nystagmus, but incidental to infantile nystagmus.146 Latent nystagmus is much more likely to be accompanied by strabismus than is infantile nystagmus.146 However, the type of the nystagmus that accompanies pediatric strabismus is to a great extent determined by the type of practice. In strabismus clinics, latent nystagmus is by far the most common accompaniment of strabismus. In neurology clinics, however, infantile nystagmus is seen much more commonly than latent nystagmus, so there is empirically a greater chance that the patient with strabismus and nystagmus will have congenital nystagmus.

The presence and nature of an underlying sensory visual disorder seems to influence the likelihood of associated strabismus. In a study of 82 children with infantile nystagmus (diagnosed clinically), Brodsky and Fray81 found the prevalence of strabismus to be 82% in children with optic nerve hypoplasia, 53% in children with albinism, 36% in children with congenital retinal dystrophies, and 17% in children with idiopathic infantile nystagmus. However, clinical assessment of the true incidence of strabismus in the setting of sensory visual disorders is confounded by the fact that children with Leber congenital amaurosis and other congenital visual disorders lack central fixation, making the assessment of strabismus difficult. The finding of esotropia and nystagmus compels the examiner to rule out manifest latent nystagmus that accompanies congenital esotropia and that differs from infantile nystagmus in its visual prognosis. When latent nystagmus and strabismus coexist, treatment of the strabismus often produces resolution of the manifest component of the nystagmus.

Eye Movement Recordings in Infantile

Nystagmus

Immature Infantile Nystagmus Waveforms

Using electro-oculographic recordings, Reinecke et al found a stereotyped waveform evolution in infants with infantile nystagmus.466 When the nystagmus first appears at 2–3 months of age, it takes on a triangular pattern that is occasionally punctuated by small plateaus. At about 7–12 months of age, the nystagmus transforms into a pendular waveform. Between 10 months and 1½ years of age, the pendular waveform gives way to an increasing-velocity jerk waveform characterized by a saccade to the target of

fixation followed by a period of foveation and an increas- ing-velocity slow phase, which again pulls the fovea away from the object of interest. A few residual triangular and pendular cycles continue to be interspersed in the increas- ing-velocity waveform. Using more accurate eye-move- ment measurement techniques, Dell’Osso has found that mature infantile nystagmus waveforms are present and continue to develop during infancy, with an evolution of waveforms from pendular to jerk, (consistent with the notion that jerk waveforms reflect modification of the oscillation by growth and development of the visual sensory system).

Mature Infantile Nystagmus Waveforms

Dell’Osso and Daroff139 have subdivided infantile nystagmus waveforms into 12 distinct categories on the basis of their electro-oculographic characteristics. Although infantile nystagmus waveforms are often subdivided for classification purposes, it is important to recognize that most infantile nystagmus patients display an average of three to five waveforms. These oscillations exist as a continuum of oscillations characterized by a period of foveation followed by an increasing-velocity slip away from the target and, finally, a corrective saccade back toward the target139 (Fig. 8.1). The visual acuity associated with each waveform is related primarily to the length of the foveation period. There is no evidence that any particular waveform is associated with better acuity (such an assessment would require knowledge of which waveform is predominating at the exact instant that visual acuity is being tested). Rather, pure pendular or jerk waveforms without foveation periods are usually associated with poorer vision, whereas waveforms of either type with extended foveation periods indicate better vision. Eye movement recordings raise doubt about the ability to clinically differentiate between pendular and jerk forms of infantile nystagmus as saccades may be seen in a clinically pendular nystagmus, and a pendular waveform without saccades may be seen in a clinically jerk nystagmus.139

Fixation in Infantile Nystagmus

Although infantile nystagmus has been attributed to a faulty fixation mechanism, Dell’Osso et al158 have performed detailed examinations of foveation periods (intrabeat dynamics, accuracy of target foveation, effects of gaze angle, convergence, and base-out prisms on foveation period) and found that idiopathic infantile nystagmus is associated with strong fixation reflexes in that individuals are able to accurately achieve and maintain fixation for long periods. Bedell

Fig. 8.1  Eye movement recordings showing three common waveforms in infantile nystagmus. (Upward deflections denote rightward eye movements; downward deflections denote leftward eye movements). (a) Eye position (POS) and velocity (VEL) record of pure jerk nystagmus. Target foveation occurs briefly at termination of each rightward saccade. Velocity spikes clearly identify rightward-jerk direction. (b) Eye position (POS) and velocity (VEL) record of jerk nystagmus with extended foveation. Note that target is foveated for longer period following each saccade than with pure jerk nystagmus. The velocity wave-

form readily demonstrates leftward direction of saccades, which is difficult to discern from position tracing alone. (c) Eye position (POS) and velocity (VEL) record of pseudo-cycloid form of jerk nystagmus. In waveform, leftward saccades are corrective in nature but are of insufficient amplitude to fully refoveate target. Each saccade is followed by smooth eye movement that refoveates target. This waveform is often misidentified clinically as pendular nystagmus. Velocity waveform is particularly useful in identifying saccadic component of each cycle.

Adapted, with permission, from Dell’Osso LF et al139

et al58 found greater standard deviations in foveation periods of two albinos than in patients with idiopathic infantile nystagmus and suggested that the effects of macular hypoplasia on the fixation mechanism may have a secondary effect on vision in albinism.

Visual acuity in infantile nystagmus has been found to correlate with fixation parameters such as the accuracy of target foveation, the duration of target foveation, and the repeatability of foveation from cycle to cycle.58,158 According to Dell’Osso, the fixation subsystem is only able to prolong foveation and maintain temporary fixation when the target image is on the fovea and moving with a velocity (or acceleration) that falls below a critical value (estimated at 4 degrees per second).155,156 This may explain why foveation periods are part of the infantile nystagmus but not the acquired nystagmus waveform because the initial slow-phase velocities in acquired nystagmus are usually too high for the fixation subsystem to extend foveation and improve visual acuity.161

While the fixation mechanism appears to be robust in infantile nystagmus, the observation that infantile nystagmus increases during attempted fixation and ceases during nonvisual tasks such as daydreaming or sleep158 suggests that the presence of an abnormal circuitry between the fixation system and the remaining ocular stabilization systems that allows the effort associated with fixation to influence the oscillation.158 Alternatively, the effort to see appears to be one of many psychological inputs (e.g., excitement, fear, anxiety) that raise the gain of the circuitry controlling the inherent oscillatory nature of the smooth pursuit system.

Smooth Pursuit System in Infantile Nystagmus

The smooth pursuit waveform in the infantile nystagmus patient bears little resemblance to that of the normal individual.159 This lack of correspondence has, in the past, been misconstrued as a possible smooth pursuit deficit in infantile nystagmus.341,441,591 Dell’Osso has stressed that the fundamental error of equating the summation of smooth pursuit movements plus the superimposed infantile nystagmus waveform with the pursuit movement alone inevitably leads to the erroneous conclusion that there is an inherent defect in the pursuit system. He has further demonstrated that during pursuit of a visual target, the slow phases of infantile nystagmus consist of normal pursuit movements plus the nystagmus itself, but that the eye position and velocity consistently matches the target position during foveation periods.148,159 If one examines the upper tracing in Fig. 8.2 (in which eye position has been superimposed on target position) and confines this examination to only the foveation periods, it becomes evident that the eye position accurately matches the target position during most of the foveation periods.148,159 Such findings cast serious doubt on the hypothesis that defective pursuit is either the cause of, or the necessary result of infantile nystagmus.159

The notion of “inverted pursuit movements” and “inverted optokinetic responses” has created further confusion regarding the role of smooth pursuit in infantile nystagmus. It is widely recognized that patients with infantile nystagmus often show an apparent reversal of their optokinetic responses

Fig. 8.2  Eye movement recording from patient with infantile nystagmus demonstrating smooth pursuit of moving, constant-velocity target. Upper tracing shows target position with right eye (RE) position superimposed. Lower tracing shows left eye position. (POS position; VEL velocity.)

Note that congenital nystagmus waveform is punctuated by brief foveation periods in which eye position precisely matches target position. Adapted, with permission, from Dell’Osso LF.148 Published with permission from the journal Neuro-Ophthalmology. Copyright by Aeolus Press)

Vestibulo-ocular Reflex in Infantile Nystagmus

Many attempts to evaluate the vestibulo-ocular reflex (VOR) in subjects with infantile nystagmus have failed to successfully separate the slow-phase velocity associated with the underlying nystagmus from that due to the VOR itself.160 Because of the superimposition of an ever-present and changing infantile nystagmus waveform on the eye movements resulting from the normal VOR, the measured responses do not resemble normal ones. Dell’Osso et al160 have stressed that calculation of the VOR gain in infantile nystagmus must be limited to foveation periods (Fig. 8.3). At any other point in the infantile nystagmus cycle (when there is neither target foveation nor clear vision due to the obligate retinal slip), the calculation of VOR gain is meaningless, both in the mathe-

Fig. 8.3  Vestibulo-ocular reflex in infantile nystagmus. Note that during head movement, nystagmus continues to be punctuated by foveation periods (middle tracing) during which position of gaze remains steady. Adapted, with permission, from Dell’Osso LF et al160

matical sense and as an indication of the performance of the VOR. Failure to recognize this interrelationship has led some to suggest that the VOR itself is deficient.98,173,210 Others have recognized that the infantile nystagmus confounds the calculations of VOR gain and have concluded that the VOR was not deficient.148,237,240,254,353 Symptomatically, it is noteworthy that patients with infantile nystagmus rarely complain of oscillopsia or exhibit symptoms that normally accompany deficits in the VOR during ambulation.

Saccadic System in Infantile Nystagmus

Although visual feedback provides a means of sampling and assessing the accuracy of foveation periods in infantile nystagmus, a number of observations suggest that fast phases are not produced in response to a retinal displacement error signal between the fovea and the target image.589 Worfolk and Abadi589 have offered the following evidence to support this supposition:

1. Jerk infantile nystagmus can continue with the eyes closed. 2. Infantile nystagmus continues and its parameters remain unchanged as individuals track paracentral afterimages, suggesting that the timing and direction of the fast phases

are not dependent on retinal feedback.342

In pendular infantile nystagmus with foveating saccades (Fig. 8.1b), the retinal displacement error signal is opposite in sign to the forthcoming fast phase until about 70 ms before the saccade, allowing insufficient time to program the quick phases using visual information. The foregoing evidence suggests that the fast phases in infantile nystagmus are likely to be initiated on a predictive basis or in response to effer- ence-copy information.589

The peak fast-phase velocity in infantile nystagmus is reduced by approximately 10% with respect to normals.8 This finding is consistent with the slightly reduced saccadic velocity in normals who are making saccades on the basis of nonvisual information rather than visually-guided saccades,8 and further suggests that factors responsible for the fast phase in infantile nystagmus may include nonvisual elements.8,342 Saccades and gaze holding are normal in infantile nystagmus, and the saccades contained within the nystagmus waveforms are always corrective and not the initiating movement responsible for the nystagmus.8,168

Suppression of Oscillopsia in Infantile Nystagmus

Several mechanisms have been proposed to account for the stability of the perceived world in the face of nearly constant motion across the retinas in individuals with infantile nystagmus.10 These include the notion of visual information sampling only during foveation periods with suppression at other times,12,14,155,156,170 use of an extraretinal signal to cancel out the visual effects of eye motion, central elevation of motion detection threshold,3,134,219,341 and adaptation to retinal image motion.17,56 Such extraretinal signals include efference copy of the relative image motion,56,57,135,147 and proprioception.10 The suggestion that individuals with infantile nystagmus periodically sample their visual environment only during foveation periods with total suppression at all other times14 (i.e., “stroboscopic” vision) was a simplistic inference drawn from the observation that clear

and stable vision was possible only during foveation periods, and has been dispelled.15

Temporal modulation studies demonstrate that individuals with infantile nystagmus process retinal information continuouslyratherthanselectivelyduringfoveationperiods.314,568 It is not surprising that infantile nystagmus patients have elevated motion detection thresholds when compared to normal patients with still eyes.179 The fact that these individuals are unable to see their nystagmus in a mirror presumably results from the simultaneous movement of the mirror images with the eyes (retinal image stabilization) because these same individuals can recognize their nystagmus on a videotape. The observation that the vision is clearest during foveation periods when the eyes are relatively still and degraded during ocular movement is a normal physiological finding that should not be misconstrued as an a priori elevation in motion detection thresholds.

Bedell59 found no evidence of decreased sensitivity to oscillatory target motion in patients with infantile nystagmus compared with control patients viewing a target with sinusoidal or ramp motion to simulate the retinal image motion that occurs with retinal eye movements. Based on his experimental results, an abnormally low sensitivity to oscillatory target motion cannot be invoked to explain the absence of oscillopsia in individuals with infantile nystagmus.

The fact that retinal image stabilization produces oscillopsia in individuals with infantile nystagmus suggests that an extraretinal signal (efference copy) is used by the brain to cancel out the infantile nystagmus waveform.10,342,363 Dell’Osso and colleagues154,159 have demonstrated that individuals with infantile nystagmus also require well-defined, repeatable foveation periods from one cycle to the next to perceive a nonmoving visual world (Fig. 8.4).154,159 Perturbations in the infantile nystagmus cycle related to external or internal factors (e.g., head trauma, medications) can result in oscillopsia.14 In one patient, oscillopsia was present only when the waveform failed to enter the foveation window;155 in another, when the foveation period fell below a minimal duration.14

Summary of Ocular Stabilization Systems

in Infantile Nystagmus

In examining how the ocular stabilization systems function in the setting of infantile nystagmus, one must confine the analysis to the foveation periods. It is during this portion of the infantile nystagmus waveform that the oscillation has subsided, vision is clear, and some degree of ocular stabilization is possible. Eye-movement recordings and phase-plane portraits in infantile nystagmus demonstrate the following:79

1. The oscillations of infantile nystagmus supersede the ocular stabilization systems but do not extinguish them.

Fig. 8.4  Phase plane portrait demonstrating multiple consecutive cycles in patient with infantile nystagmus. Figure does not depict the trajectory of eyes. Its purpose is to simultaneously display position and velocity of eye at any point in nystagmus cycle. By touching line at any point with pencil, examiner can simultaneously assess position and velocity of eyes at that point in time. Phase plane portraits are useful in understanding visual acuity and suppression of oscillopsia in congenital nystagmus. For good visual acuity, eye position must simultaneously fall within ½ degree of fovea (bracketed by vertical lines) and have velocity of less than about 4 degrees per second (bracketed by vertical lines). Time function is not linear along each tracing; relatively less time is spent in positions of high velocity, and more time is spent in positions of low velocity. Note stereotyped appearance of each repetitive cycle, which appears to be prerequisite for suppression of oscillopsia. Adapted, with permission, from Dell’Osso LF et al158

2. Amidst the ongoing oscillations of infantile nystagmus, these systems exert their primary influence on vision during foveation periods.

3. Defects in ocular stabilization are neither the cause nor the necessary result of infantile nystagmus.

Contrast Sensitivity and Pattern Detection

Thresholds in Infantile Nystagmus

The threshold for acuity, contrast sensitivity, motion detection, and stereoacuity are typically elevated in patients with infantile nystagmus.57,60a A reduction in contrast sensitivity for medium to high spatial frequency vision and increased pattern detection thresholds in infantile nystagmus impairs the detection of vertically oriented stationary and moving grating patterns more so than horizontal ones. The increased contrast sensitivity and pattern detection thresholds are secondary to the oscillation itself and

Theories of Causation

Early theories regarding the cause of infantile nystagmus focused on the notion that the oscillation must result from an inherent abnormality in one of the ocular stabilization systems (i.e., smooth pursuit system, optokinetic system, VOR, or fixation system). Over the last two decades, however, the accumulated clinical and eye movement evidence has refuted these hypotheses. Attempts to attribute the oscillation to a neuronal misdirection as seen in albinism or achiasma298 provide insight into the specific mechanism by which chiasmal misrouting may precipitate infantile nystagmus, but do not explain the wide variety of other visual disorders associated with infantile nystagmus.

Harris and Berry261 have resurrected the century-old theory of Swanzy522 that infantile nystagmus results from a failure of sensorimotor integration in infancy and beautifully elaborated it in modern neurobiological terms. The first few months of life are a period of rapid visual development in which motor development can be influenced by postnatal visual experience.261 This plasticity may be under active genetic control that can itself be influenced by visual experience.384 Abnormal postnatal visual experience may induce an adaptive oculomotor response that leads to nystagmus during a critical period of heightened plasticity.261,262

Contrast sensitivity to low spatial frequencies is enhanced by motion of the image across the retina. Harris and Berry have proposed that the best compromise between moving the image and maintaining the image near the fovea (or its remnant) is to oscillate the eyes with jerk nystagmus with increasing velocity waveforms, as seen empirically.261,262 The result may be a developmental “funnel,” in which loss of high spatial frequency information (whether caused by foveal, optic nerve, or optical aberrations) could lead to oscillatory strategies to maximize low-frequency information.261,262 VEP testing in children with idiopathic infantile nystagmus shows decreased responses relative to normals, suggesting that a delay in the development of high spatial frequency contrast could indeed precipitate infantile nystagmus.571a This perspective views infantile nystagmus not as a defect but as a maladaptation in which the developing visual system has the potential to develop many different adaptive control systems. According to Harris and Berry, “evolution would need to tread a fine line by programming the development of ocular motor control in tandem with foveal maturation to maximize visual contrast without causing nystagmus.”212 This developmental theory is supported by the finding that patients with idiopathic infantile nystagmus do not have any known