Patterns of ocular oscillation in oculopalatal tremor

Palatal tremor (PT) consists of continuous rhythmic movements of the soft palate, sometimes accompanied by movements of adjacent structures derived from the branchial arch.^1–3 When PT is associated with pendular nystagmus (PN), it is called oculopalatal tremor (OPT, also designated oculopalatal myoclonus).^4,5 Vertical PN is typical of OPT, but the oscillations can be asymmetric between the eyes.⁶ PT with or without PN is typically a manifestation of focal brainstem or cerebellar damage, but can also be a feature of brain degeneration in the syndrome of progressive ataxia and PT (PAPT).^5,7

OPT is a delayed complication of damage to the dentatorubro-olivary pathway (the Guillain-Mollaret triangle) and subsequent hypertrophic olivary degeneration.^3,8 Damage to the dentatorubral circuit is associated with hypertrophy of the contralateral inferior olivary nucleus (ION), whereas involvement of the central tegmental tract between the red nucleus and the medulla is associated with hypertrophy of the ipsilateral ION.^8–10 OPT has been proposed to exhibit either lateral or midline forms.¹¹ The term lateral form was applied to unilateral (asymmetric) PT and was thought to be associated with disparate vertical amplitudes of PN between each eye and with conjugate torsional oscillation; it was called lateral because it was postulated to signify unilateral brainstem damage. In the lateral form, PT was reported to occur on the side of the ocular oscillation with greater vertical amplitude.¹¹ The proposed midline form¹¹ was thought to consist of symmetric PT and vertical PN and was postulated to signify bilateral brainstem damage. However, neither imaging of the inferior olives nor neuropathologic evidence of these lateral and midline categories was provided.¹¹

Clinicoanatomic correlation, or recording of eye motion in OPT, has been sparse.^6,9–11 To determine patterns of nystagmus in OPT and whether asymmetry of PN and PT correlates with asymmetry of inferior olivary changes, we investigated PT, eye motion, and MRI findings in 22 patients with OPT.

Methods.

Twenty-two patients with OPT were recruited from the Neuro-Ophthalmologic Clinics at the University Health Network, Toronto, and Seoul National University Bundang Hospital. Twenty patients developed OPT from focal brainstem lesions. Some had cerebellar damage, but none had isolated cerebellar lesions. Two had the PAPT syndrome (table). Focal brainstem lesions included pontine or pontomesencephalic hemorrhage (n = 15), brainstem without or with cerebellar infarction (n = 3), brainstem and cerebellar arteriovenous malformation (n = 1), and pontomesencephalic abscess (n = 1). All the patients received full neurologic and neuro-ophthalmologic evaluation by two of the investigators (J.S.K., J.A.S.). Initial clinical evaluation or recording of eye movements was done 1 day to 15 years after the onset of symptoms. Duration of follow-up from the symptom onset ranged from 1 to 20 years (mean, 8 years; median, 7 years). In 15 patients, eye motion and PT were videotaped for later assessment of asymmetry of PT and synchronicity between PN and PT as previously described in one of our patients.⁴ The person judging the symmetry of PT was blinded to the MRI findings. We have previously reported features other than the ocular oscillations and their correlation with imaging in three of our 22 patients (Patients 2⁴, 18¹², and Patient 21⁵).

Results.

MRI demonstrated increased signal intensity or enlargement of one or both ION on proton density or T2-weighted images in all 22 patients (figure 1; figures E-1 and E-2 on the Neurology Web site at www.neurology.org).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. MRIs of four representative patients with unilateral (A and B) or bilateral (C and D) pseudohypertrophic degeneration of the inferior olivary nucleus, showing signal changes in the left column and the causative lesion in the right column. MRI of Patient 1 (A) ( table ), MRI of Patient 7 (B), MRI of Patient 15 (C), and MRI of Patient 16 (D). Patient 1 had a diffuse pontomesencephalic abscess (not shown in A, right column). MRIs of other patients are shown in the supplementary material on the Neurology Web site.

Pendular nystagmus.

The nystagmus was either symmetric between the eyes or dissociated with each eye moving in different amplitudes or directions. Symmetric nystagmus was observed in 14 (figure 2; see video E-1) and dissociated nystagmus occurred in eight of our 22 patients (figure 3, see video E-2). Of the 14 patients with symmetric PN, eight had bilateral and six showed unilateral ION signal changes on MRI. All eight patients with dissociated PN exhibited asymmetric pseudohypertrophy of ION, and six of them showed more prominent PN on the side contralateral to the imaged olivary changes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Magnetic search coil recording of pendular nystagmus in Patient 1 (left inferior olivary hypertrophy in figure 1A) shows symmetric vertical-torsional nystagmus. In the horizontal plane, the oscillation is convergent-divergent pendular nystagmus synchronous with the torsional and vertical component at about 3 Hz. In this and figures 3 and 4, upward trace deflection is rightward for horizontal motion, up for vertical motion, and clockwise for torsional motion from patient's reference (upper pole of eye rotates toward patient's right shoulder). Downward trace deflection is leftward for horizontal, down for vertical, and counterclockwise for torsional motion (counterclockwise is from patient's reference, i.e., upper poles of eye rotates toward left shoulder). LH = horizontal position of the left eye; RH = horizontal position of the right eye; LV = vertical position of the left eye; RV = vertical position of the right eye; LT = torsional position of the left eye; RT = torsional position of the right eye in this and following figures.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Recording of pendular nystagmus in Patient 7 (left inferior olivary hypertrophy in figure 1B) shows mixed vertical-torsional nystagmus in the right eye and pure torsional nystagmus in the left eye. Upward displacement of the right eye was accompanied by intorsion of the right eye and extorsion of the left eye.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Recording of eye motion in Patient 12 (right inferior olivary hypertrophy) shows conjugate torsional and asymmetric vertical oscillation. Dissociated horizontal (vergence) oscillations, greater in the left eye, are associated with the vertical-torsional nystagmus.

Of the 14 patients with symmetric PN (SI <10%), nine had symmetric and two showed asymmetric PT. PT was not observed in three of them. Of the eight patients with dissociated PN, five exhibited symmetric and two showed asymmetric PT. One patient did not show PT. PT was synchronous with PN in 12 patients. However, PT was not synchronous with PN in six patients (video E-3), as previously described in one patient.⁴

Most patients showed mainly vertical oscillations with combinations of torsional and horizontal components, which resulted in varied waveforms of PN. The frequency was 1.5 to 3.0 Hz and was mostly irregular, varying during sequential cycles. The amplitude varied beat to beat from 2 to 10 degrees. The vertical component of oscillations was conjugate (same amplitude and direction in both eyes) and greater than the torsional or horizontal component in all these patients except one (Patient 6), who showed mainly horizontal torsional nystagmus with minimal vertical component (video E-4). The horizontal and torsional oscillations were usually conjugate. However, two patients (Patients 1 and 2) showed disjunctive horizontal (vergence) oscillation (figure 2, video E-1), and one with PAPT syndrome (Patient 22) had a disjunctive torsional component (video E-5). The vergence oscillations were synchronous with the vertical and torsional oscillations.

In five of the eight patients with dissociated PN (Patients 7 to 11), upward displacement of one eye was accompanied by intorsion, whereas the other eye had relatively pure torsional component with no or minimal vertical excursion. The torsional oscillations were in the same direction in both eyes (figure 3, video E-2). This pattern of nystagmus resembled the alleged lateral form of OPT as designated by Nakada and Kwee.¹¹ One of them (Patient 11) showed dissociated horizontal oscillations (greater in the left eye) accompanied by mixed vertical-torsional nystagmus.

Of the other three patients with dissociated PN, one patient (Patient 12) with inferior olivary pseudohypertrophy on the right side showed conjugate torsional and asymmetric vertical oscillation (figure 4, video E-6) and dissociated horizontal (vergence) oscillations, greater in the left eye. In the remaining two patients with asymmetric nystagmus, PN was mainly vertical with disparate amplitudes between the eyes (Patient 13) or mainly vertical-horizontal (Patient 14).

Discussion.

Inferior olivary pseudohypertrophy has generally been considered to be crucial for the genesis of OPT.^8,11 Inferior olivary neuron membranes generate rhythmic excitation and inhibition.²³ Loss of gamma-aminobutyric acid modulation on gap junctions may lead to hypersynchronous firing of inferior olivary cells and OPT.^24,25 MRI detects inferior olivary change that is well correlated with pathologic findings.^16–18 Hyperintense signal change in the ION can appear on T2-weighted or proton density MRIs as early as 3 weeks after a causal event¹⁴ and may persist indefinitely.^17,26 Hyperintense signal change is attributed to increased water content and gliosis.^16,19 Hypertrophy of the olive corresponds to enlargement of both neurons and astrocytes.¹⁸ Neurons become vacuolated. Later after many months or a few years, when neurons undergo death, the term pseudohypertrophy is appropriate. Ultimately, after many years, olivary atrophy appears.¹⁸

PT with or without PN typically develops several months to years after a causal event,²⁷ but the nystagmus may be evident within 24 hours and precede PT.²⁸ We identified PN with no PT over years of observation in four of our 22 patients with inferior olivary signal change. Furthermore, PN and PT ceased in one of our patients, and another had nearly complete resolution of PN, after many years, although repeated MRIs in both patients showed persistent signal change in the ION.

PT is generated by rhythmic contraction of the levator veli palatini, innervated from the nucleus ambiguus by the facial or glossopharyngeal nerve²⁹ and may be synchronous³⁰ or asynchronous³¹ with ocular oscillation. In 13 of our 18 patients with PT and PN, PT appeared to be synchronous with PN. Unilateral PT had been reported with mixed vertical-torsional nystagmus, giving rise to the term lateral form of OPT, whereas symmetric bilateral PT was reported with vertical PN and termed the midline form.¹¹

Although palatal contraction has been recorded on the side contralateral to inferior olivary hypertrophy seen on MRI,⁹ previous reports have conflicted on the lateralizing value of PT.^29,32–34 Only four of our 14 patients with unilateral olivary changes on MRI showed asymmetric PT on visual inspection; PT was more prominent on the opposite side of olivary change. Stimulation of the ION or adjacent reticular formation generates ipsilateral or bilateral PT, and stimulation of the central tegmental tract causes varied directions of contractions.²⁹ Due to a common insertion of both levator muscles at the medial aponeurosis, small unilateral contraction of the levator muscle generate bilateral movement of the soft palate.⁹

According to one hypothesis,¹¹ bilateral lesions generate symmetric vertical PN and unilateral lesions give rise to dissociated vertical-torsional PN,¹¹ but imaging of the ION was not available to support the hypothesis. Among our 22 patients with OPT (14 with symmetric PN and eight with dissociated PN), symmetric PN occurred in six patients with unilateral and in eight patients with bilateral inferior olivary hypertrophy, whereas dissociated PN occurred in all eight patients with abnormality of one ION. The various patterns of PN in our patients with OPT did not correlate with laterality of inferior olivary changes on MRI and do not support the contention¹¹ that symmetric vertical nystagmus indicates bilateral lesions. Dissociated PN predicted asymmetric (unilateral) ION change on MRI with 100% accuracy, but symmetric PN was associated with either unilateral or bilateral ION signal changes. However, without confirmation at necropsy, it cannot be excluded that the patients with symmetric PN and unilateral ION signal change may have had microscopic pathologic changes in the ION that appeared to be normal on MRI. Nonetheless, MRI is believed to be sensitive in detecting increased water contents and gliosis of degenerating ION.^16–18 The frequent bilaterality of the causative lesions in our patients with unilateral olivary changes and symmetric PN is also suggestive of possible ION pathology undetected by MRI. Higher resolution of MR signal changes may become available in future to improve the sensitivity of imaging.

All 20 of our patients with focal lesions had damage in the brainstem tegmentum. Although two of them also had focal damage to the cerebellum (table and MR images on the Neurology Web site), none had isolated cerebellar lesions Two additional patients had sporadic degenerative PAPT⁵ without focal lesions. We are not aware of any report of PN either isolated or associated with PT, with either pathologic or modern imaging correlation, that documents PN in a patient with a structural lesion confined to the cerebellum. Our results suggest that the PN of OPT specifies intrinsic brainstem damage to the central tegmental tract or neighboring projections to the ION.

ION neurons project climbing fibers to Purkinje cells in the contralateral cerebellum through the inferior cerebellar peduncle. The ION participates in the learning and timing of movements.³⁴ Functional inactivation of the inferior olive in animals abolishes adaptation in the vestibulo-ocular reflex (VOR); inactivation prevents increase or decrease in the gain of the VOR.³⁵ The ION may regulate VOR adaptation by modifying synaptic transmission of parallel fibers onto the floccular Purkinje cells.³⁶ Previously, PN in OPT was postulated to result from an impaired adaptation of the VOR due to degeneration of the ION.¹¹ Although ocular oscillation in OPT might be accompanied by impaired adaptation of the VOR mediated by the ION, this may not be relevant to the genesis of the oscillation. Indeed, visual modulation of VOR gain during near viewing does not correlate with a change in nystagmus amplitude in OPT.³⁷ A disturbance of central VOR circuits does not readily explain the PN associated with PT.

Although inferior olivary pseudohypertrophy has been considered a pathologic substrate of OPT,^8,11,29 the pendular ocular oscillations and PT in OPT often do not occur together, as illustrated by eight of our patients with ION signal change. Distinct but neighboring anatomic lesions may be responsible for each. Structures adjacent to the ION, such as the dorsal cap of the ION, the dorsolateral reticular formation, or the paramedian tracts cell groups, may be alternative generators of the PN component of OPT. Instability in feedback control of the eye velocity-to-position integrator has been invoked as a mechanism of acquired PN.¹⁴ The nucleus prepositus hypoglossi, the medial vestibular nucleus, and the cerebellar flocculus are elements of the velocity-to-position neural integrator, which converts horizontal eye velocity signals to position signals, and thereby serves maintenance of steady gaze.³⁸ The interstitial nucleus of Cajal, vestibular nuclei, and paramedian tracts are elements of the velocity-to-position integrator for vertical eye motion.^39–41 Neurons of the paramedian tracts in the midline of the pontine and medullary tegmentum discharge as burst-tonic units with activity related to vertical eye position.⁴¹ Their efferents project along the midline, then pass laterally to follow the ventral external arcuate fibers around the surface of the medulla, overlying the inferior olives, and then ascend into the restiform body⁴⁰ to terminate in the flocculus and ventral paraflocculus. In our patients, the causative lesions, which were mainly located bilaterally in the pontomedullary tegmentum, may have damaged these neural integrators on both sides, giving rise to symmetric PN even in patients with unilateral olivary changes. It is also possible that pseudohypertrophy of ION distorts and damages the overlying arcuate fibers,⁴⁰ giving rise to instability in the integrator function of the paramedian tracts and thereby to the PN component of OPT.

Simulation using a network model of the neural integrator¹⁵ predicts phase shifts of PN by saccades that would reset the neural integrator; a phase lag would follow larger saccades and a small phase lead would follow smaller saccades.¹⁴ Saccades cause such phase shifts of acquired PN in patients with multiple sclerosis (MS).¹⁴ The phase shift is greater for larger than for smaller saccades. Those findings support a hypothesis that an unstable neural integrator generates PN. In contrast, we identified a phase lead of PN after saccades in a patient with OPT. The phase lead was independent of saccadic amplitudes. Because this result of saccade-induced phase shifts in PN of OPT is based on a single patient, the application of this observation to other OPT patients remains uncertain. However, our findings suggest a mechanism of pendular oscillations in OPT distinct from that postulated for acquired PN in MS.¹⁴ The PN of OPT has lower and more irregular frequency, more predominant vertical oscillation, and usually greater amplitude. A disturbance extrinsic to the velocity-to-position integration related to PN of MS may give rise to the PN of OPT.

The ventrolateral outgrowth and dorsal cap of the ION relay visual signals of the accessory optic system to the flocculus.^42,43 The rostral part of the dorsal cap of the rabbit receives its input from the contralateral eye through the terminal nuclei of the accessory optic tract and ventromedial tegmental area and participates in vertical smooth ocular tracking.⁴³ The caudal dorsal cap receives its sensory input predominantly from the nucleus of the optic tract⁴⁴ and is primarily involved in horizontal smooth eye movements.^43,45 The dorsal cap also receives eye motion signals from the nucleus prepositus hypoglossi and transmits these motor signals as complex spikes on climbing fibers to the flocculus.⁴⁶ The motor signals are required to distinguish slippage of retinal images caused by internally generated eye motion from image slip caused by movement of the visual environment.⁴⁶ If eye motion signals can predict the occurrence of retinal image slip, the visual signals are thought to be gated and not relayed to the flocculus. Dysfunction of the dorsal cap may prevent stabilization of the retinal image slip generated during nystagmus and preclude repair of the abnormal eye motion. In our patients, the dorsal cap could not be resolved from the ION by imaging. Brainstem damage that denervates the dorsal cap is a possible foundation for the development of PN associated with PT.