Electroencephalography Findings in Wrong Way Eyes: Unilateral Hemispheric Dysfunction Supporting Smooth Pursuit Asymmetry Hypothesis

Background: Conjugate horizontal eye deviation away from the side of the lesion, termed Wrong Way Eyes (WWE), is a rare manifestation of supratentorial lesions. The proposed etiologic hypotheses include seizure activity, compression of contralateral horizontal gaze pathways from mass effect or midline shift, and asymmetry of hemispheric smooth pursuit mechanisms. We present neurophysiological evidence that favors the asymmetry of hemispheric smooth pursuit hypothesis. Methods: Electroencephalography (EEG) was performed in 2 patients with large left hemispheric supratentorial lesions, capturing fluctuating periods of (a) unresponsiveness with WWE and (b) relative alertness without WWE. One patient had 5 days of continuous EEG, and the other routine EEG. Results: Neither patient had seizures. EEG showed normal right hemispheric activity during both unresponsiveness with WWE and alertness without WWE states. By contrast, more severe left hemispheric dysfunction was evident in the WWE state compared with the non-WWE state in both patients. In one patient, during the relatively alert state, right-beating nystagmus was observed, and drift of the eyes away from the side of the lesion was reliably seen to occur on eyelid closure and after ipsiversive volitional saccades. Conclusions: Seizure activity does not account for WWE. Compression of contralateral horizontal gaze pathways is also unlikely to account for WWE as that hypothetical mechanism should produce EEG abnormalities over the nonlesioned hemisphere, which were not seen. The findings suggest instead that a single dysfunctional hemisphere is sufficient to produce WWE. The repeated rightward drift of the eyes and nystagmus seen in one patient during relative alertness, and the observation of unilateral hemispheric dysfunction on EEG during unresponsiveness with WWE in both patients supports the idea that an imbalance of smooth pursuit mechanisms is most likely to account for this rare phenomenon.

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Electroencephalography Findings in Wrong Way Eyes: Unilateral Hemispheric Dysfunction Supporting Smooth Pursuit Asymmetry Hypothesis Kia Gilani, MD, Felix J. Tyndel, MD, Richard A. Wennberg, MD, PhD Background: Conjugate horizontal eye deviation away from the side of the lesion, termed Wrong Way Eyes (WWE), is a rare manifestation of supratentorial lesions. The proposed etiologic hypotheses include seizure activity, compression of contralateral horizontal gaze pathways from mass effect or midline shift, and asymmetry of hemispheric smooth pursuit mechanisms. We present neurophysiological evidence that favors the asymmetry of hemispheric smooth pursuit hypothesis. Methods: Electroencephalography (EEG) was performed in 2 patients with large left hemispheric supratentorial lesions, capturing fluctuating periods of (a) unresponsiveness with WWE and (b) relative alertness without WWE. One patient had 5 days of continuous EEG, and the other routine EEG. Results: Neither patient had seizures. EEG showed normal right hemispheric activity during both unresponsiveness with WWE and alertness without WWE states. By contrast, more severe left hemispheric dysfunction was evident in the WWE state compared with the non-WWE state in both patients. In one patient, during the relatively alert state, right-beating nystagmus was observed, and drift of the eyes away from the side of the lesion was reliably seen to occur on eyelid closure and after ipsiversive volitional saccades. Conclusions: Seizure activity does not account for WWE. Compression of contralateral horizontal gaze pathways is also unlikely to account for WWE as that hypothetical mechanism should produce EEG abnormalities over the nonlesioned hemisphere, which were not seen. The findings suggest instead that a single dysfunctional hemisphere is sufficient to produce WWE. The repeated rightward drift of the eyes and nystagmus seen in one patient during relative alertness, and the observation of unilateral hemispheric dysfunction on EEG during unresponsiveness with WWE in both patients supports the idea that an imbalance of smooth pursuit mechanisms is most likely to account for this rare phenomenon. Journal of Neuro-Ophthalmology 2023;43:417–422 doi: 10.1097/WNO.0000000000001819 © 2023 by North American Neuro-Ophthalmology Society Division of Neurology, Department of Medicine, University of Toronto, Toronto, Canada. The authors report no conflicts of interest. Address correspondence to Richard Wennberg, Division of Neurology, Clinical Neurophysiology Laboratory, Toronto Western Hospital, 399 Bathurst Street, Toronto, ON M5T 2S8, Canada; E-mail: r. [e-mail] Gilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 S upratentorial lesions can rarely result in “Wrong Way Eyes” (WWE), which refers to conjugate eye deviation away from the side of the lesion. This unusual phenomenon is at odds with the more commonly occurring pattern of eye deviation, which is toward the side of a supratentorial lesion. Miller Fisher was the first to report a case series of WWE in 3 patients with large thalamic hemorrhages (1). Thereafter, the phenomenon was historically considered to occur in the context of thalamic hemorrhages, although the largest published case series, which identified 12 patients with WWE among 968 consecutive stroke patients, found that 5 of the cases had large ischemic infarcts (2). Three mechanisms have been proposed to date to explain WWE. The first is seizure activity resulting in contralateral eye deviation (3). Until now, however, the reported cases that have included electroencephalography (EEG) recordings have not described any seizure activity associated with WWE (2–6). The second proposed mechanism is the “mass effect hypothesis,” which implicates the interruption of descending horizontal oculomotor pathways in the contralateral hemisphere from the mass effect or midline shift caused by an ipsilateral lesion (3). The site of compression is said to be at the midbrain, as this is the site of a partial decussation of horizontal oculomotor gaze pathways (7). There is neuropathologic support for this hypothesis because the dissection of blood into the midbrain was seen on autopsy of some reported cases of WWE (1,8). The association of vertical eye deviation in some, and reduced level of consciousness in most, patients with WWE is also thought to be suggestive of bilateral deep gray matter or midbrain involvement (2). The final proposed mechanism is the “asymmetry of smooth pursuit” hypothesis, based on findings from a patient with reduced level of consciousness and WWE resulting from a right frontal lobe hemorrhage. After the patient regained consciousness and the WWE resolved, she was found to have impaired smooth pursuit bilaterally, most significantly toward the side of the lesion (5). This asymmetry of smooth pursuit was proposed to be the etiologic factor underlying the WWE phenomenon. The same patient showed spontaneous drift of 417 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution the eyes away from the side of the lesion on eyelid closure, interpreted as additional support for this hypothesis. We report EEG evidence from 2 patients with WWE resulting from supratentorial lesions that further informs discussion of the mechanism underlying the WWE phenomenon. METHODS The first patient is a 55-year-old woman who had a ruptured left distal (M3) middle cerebral artery (MCA) aneurysm resulting in a dorsolateral frontal cortex hemorrhage. The hemorrhage was surgically evacuated, with open clipping of the aneurysm. Three days later, her level of consciousness abruptly declined, and brain computed tomography (CT) showed increased edema around the site of the hemorrhage causing a new 7-mm midline shift (Fig. 1A). At this time, WWE were clinically noted for the first time and also evident on the CT scan (Fig. 1A). The patient was treated with a left decompressive craniectomy, and the next day was noted to have transitory conjugate upward eye deviation, which was treated empirically as seizure activity. Over the following week, her Glasgow coma score (GCS) fluctuated between 3T and 10T, with recurrent sustained episodes of rightward eye deviation associated with declines of the GCS to 3T. The WWE would resolve on recovery of the GCS back to 10T. These fluctuations in state and eye position took place over seconds to minutes. The fluctuations were thought to represent seizure activity, and the patient was treated with benzodiazepines and escalating dosages of antiseizure medications with no improvement. Continuous video-EEG (cEEG) was subsequently acquired for 5 days, capturing many occurrences of WWE. The second patient is a 53-year-old man who had a large left MCA territory infarction from a proximal (M1) segment thrombus, initially presenting with aphasia and right hemiparesis. Brain CT and MRI showed extensive infarction within the left MCA territory (Fig. 1B). Six days after admission, his level of consciousness decreased in association with a new 11mm midline shift, and WWE were noted for the first time. Over the following days, his level of consciousness fluctuated between (a) states of unresponsiveness with WWE and (b) states of relative alertness without WWE. When alert he could obey simple commands. The recurrent episodes of WWE were empirically treated as seizures and a routine 30-minute videoEEG recording was obtained, which captured several fluctuations between periods during which the patient was unresponsive, with WWE, and periods during which he was relatively alert, without WWE. The 23-channel cEEG and routine EEG recordings were performed using standard clinical methods and systems (Natus Xltek, Oakville, Canada; Compumedics Neuroscan, Charlotte, NC), with international 10-20 system electrode placement plus simultaneous 2-channel electrooculography (EOG). In Patient 1, EEG electrodes positioned within the left frontotemporoparietal craniectomy site recorded from the scalp directly overlying cortex, in the absence of intervening bone. 418 RESULTS Patient 1 had many instances of obtundation associated with WWE during the 5 days of cEEG monitoring, with no evidence of seizure activity (Fig. 2). Patient 2 fluctuated between states of obtundation and increased alertness, with and without WWE, respectively, during routine EEG, and no seizure activity was noted (Fig. 3). Antiseizure medications were discontinued in Patient 2 after the EEG study. No interictal epileptiform discharges were seen in either patient over the duration of both recordings. The background EEG activity over the right hemisphere was within normal limits in both patients during WWE and non-WWE states alike, and there was no apparent difference in right hemispheric activity during transitions between these 2 states. By contrast, electrographic activity over the left hemisphere differed slightly between the WWE and non-WWE states. In the unresponsive with WWE state, the left hemispheric activity was more disorganized, with a greater abundance and more widespread distribution of polymorphic delta activity. As alertness increased and WWE resolved, the left hemispheric activity became better organized, with a greater abundance of faster (theta/alpha) frequencies and decreased delta power, representing a slight improvement in left-sided cortical and subcortical function (Figs. 2, 3). In Patient 1, during periods of increased alertness without WWE, right-beating nystagmus and spontaneous rightward drift of the eyes were at times noted clinically. EOG analysis of Patient 1’s eye movements while relatively alert, without WWE, and being tested by examiners, revealed spontaneous drifts of the eyes toward the right, often with superimposed right-beating jerk nystagmus, and rightward drifts immediately after volitional leftward (ipsiversive) saccades (Fig. 4). The velocity of the postsaccadic drift toward the right was increased by eye blink, and eyelid closure during relative alertness without WWE reliably initiated conjugate rightward eye deviation (Fig. 5). No nystagmus was noted in Patient 2’s EOG channels, and this patient’s saccadic eye movements were not assessed during the routine EEG study. One week after commencement of Patient 1’s cEEG, her GCS had improved in a stable fashion and WWE did not recur. She was discharged from hospital 28 days after admission with a normal level of consciousness but with residual aphasia and right hemiparesis. Patient 2’s cerebral edema and midline shift gradually resolved over 3 days following the EEG recording and the patient’s GCS stabilized, with no recurrence of WWE. DISCUSSION The clinical findings in these patients are similar to those previously described. Both developed WWE after a delay, simultaneous with the development of new midline shift and depressed level of consciousness (2). Both patients had Gilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 1. A. Patient 1. Brain CT 3 days after surgical evacuation of left frontal hemorrhage and MCA aneurysm repair; WWE first noted this day. Decompressive left hemicraniectomy performed one day later. B. Patient 2. Brain CT 6 days after large left MCA territory infarct; WWE first noted this day. Diffusion-weighted MRI (right panel) showing extent of infarction. CT, computed tomography; MCA, middle cerebral artery; WWE, wrong way eyes. fluctuating levels of consciousness, with WWE only manifesting when they were unresponsive. Only the lesioned hemisphere showed worsening on EEG when unconscious. The association of WWE with a decline in level of consciousness makes the seizure hypothesis appealing. This resulted in treatment with antiseizure medications in both patients, with Patient 1 receiving multiple agents for presumed seizures before cEEG was started. Both patients underwent state changes during the EEG recordings, affording many opportunities to capture seizure activity, which was not seen. The data from Patient 1 most strongly refute seizures as the cause of WWE, as she had 5 days of cEEG after a craniectomy, increasing the sensitivity of EEG for detecting abnormalities that might otherwise be missed (as the absent left cranial bone rendered the recording similar to intracranial EEG). Prior reports of EEG findings in patients with WWE have not described the duration of recording or whether patients had WWE at the time of recording (2,3,5,6). To our knowledge, this is the first report of a patient with WWE who had both cEEG and a craniectomy overlying the lesioned hemisphere, adding to the body of evidence that seizure activity does not account for WWE. The “mass effect hypothesis” is favored in more recent reports because it accounts for many of the clinical features seen in patients with WWE (2,3). In the largest ever case series of 12 WWE patients, a delay before the onset of WWE was noted in all cases. By the time patients developed WWE, all had midline shift with mass effect on the midbrain (2). Vertical eye deviation has been noted in many patients with WWE (2,4), including Patient 1 in our report, suggesting involveGilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 ment of the midbrain (2). Despite these suggestive features, the EEG recordings from our patients did not reveal abnormalities in the nonlesioned hemisphere associated with the WWE state, abnormalities that would be expected if there was significant contralateral midline dysfunction of deep gray matter structures (e.g., thalamus, basal ganglia) as a result of mass effect (9,10). In prior reports of WWE resulting from a unilateral lesion where EEG abnormalities are described, diffuse slow wave activity has been reported (2–4), which could indicate either bilateral midline dysfunction of deep gray matter, if synchronous and/or rhythmic, or subcortical white matter dysfunction, if intermittent, polymorphic and focal or multifocal in distribution (9). The absence of EEG changes over the nonlesioned hemisphere during WWE states in our patients suggests that these previously reported EEG findings reflecting bilateral subcortical dysfunction caused by the lesion and associated mass effect are not a necessary electrographic correlate of the WWE phenomenon. Furthermore, although a majority of cases are associated with midline shift, a punctate thalamic infarct has been reported to cause WWE in a patient in the absence of any midline shift (6). The fluctuations of consciousness without EEG changes in the nonlesioned hemisphere in our patients suggests that transient increases in unilateral edema may be just enough to produce unconsciousness and WWE, without affecting the function of the contralateral hemisphere. Although these findings do not indicate whether the reduced level of consciousness is causative or merely correlated with WWE, they do provide evidence that a single dysfunctional hemisphere is all that is required to produce WWE, making the “mass effect hypothesis” less likely. 419 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 2. Patient 1. A. EEG during period of unresponsiveness with WWE. B. EEG 20 minutes later during a state of relative alertness (able to make volitional and purposeful movements) without WWE. See text for details. Anterior–posterior bipolar longitudinal montage; HFF, high-frequency filter; LFF, low-frequency filter, this and following figures. EEG, electroencephalography; WWE, wrong way eyes. The “smooth pursuit asymmetry” hypothesis proposed by Sharpe et al (5) argues that an asymmetry of hemispheric smooth pursuit mechanisms may be the underlying cause of WWE. Because the asymmetrical smooth pursuit hypothesis requires that a single hemisphere be dysfunctional, the unilateral EEG disturbances seen in our patients during WWE states lends most support to this hypothesis. Asymmetry of smooth pursuit from supratentorial lesions, when present, typically affects eye movements toward the side of the lesion. In a study of 72 patients with cerebral lesions of various etiologies, 33 were noted to have asymmetrical smooth pursuit, 30 of whom had ipsilateral pursuit affected more than contralateral pursuit (11). The lesioned areas associated with smooth pursuit asymmetry involved a long strip extending from the temporo-occipital V5 area through the internal capsule to the dorsomedial frontal lobe; the most severe ipsilateral smooth pursuit asymmetries were seen in patients with V5 lesions (11). Although the presence or absence of WWE was not mentioned in these patients, given the rarity of the phenomenon, it is probable that most of these patients did not have WWE. Although the smooth pursuit asymmetry noted in the patient reported by Sharpe et al (5) may not be unique to patients with WWE, the authors postulated an asymmetry of cerebral horizontal gaze tone to explain the WWE phenomenon. Such an imbalance could also account for the nystagmus seen in Patient 1 and that previously described in another 420 WWE patient (6). Central horizontal gaze tone in humans remains a hypothetical construct, without identified cellular or network correlates. In nonhuman primates, periarcuate neurons in the frontal eye fields, which receive vestibular inputs, are involved in ipsilateral smooth pursuit as part of a distributed circuit that includes medial superior temporal and posterior parietal cortical areas and subcortical projections to the thalamus, pons, vermis, and floccular lobe of the cerebellum (12,13). In humans, nystagmus resulting from large cortical lesions, termed “pursuit paretic nystagmus,” was initially described in patients who had undergone hemispherectomy (14). The fast phase of this nystagmus beats away from the side of the lesion (15), as seen in Patient 1 and the other reported case (6). An imbalance of smooth pursuit mechanisms was proposed to account for the contralateral deviation of the eyes on eyelid closure in the patient reported by Sharpe et al (5). Similarly, during relatively alert states, Patient 1 in our report had spontaneous contralateral drift of the eyes (or increased drift velocity) reliably seen on eyelid closure. Despite a growing body of evidence pointing away from the thalamus as the sole lesional cause of WWE, the thalamus is almost invariably involved in all cases of WWE. While the localization of WWE is difficult due to the typically large lesions affecting multiple structures, a punctate thalamic infarct without mass effect has been shown to produce WWE (6). The non-thalamic lesions reported to cause WWE almost Gilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 3. Patient 2. A. EEG during period of unresponsiveness with WWE. B. EEG 2 minutes earlier during relatively alert state (able to obey commands), without WWE. See text for details. EEG, Electroencephalography; WWE, wrong way eyes. invariably resulted in midline shift, indirectly affecting the thalamus (2,3). In the case series of Jokkura et al (2), patients with non-thalamic lesions developed WWE after a delay of 2– 4 days, while those with thalamic lesions or compression had a delay of no more than one day. The patients we report did not have involvement of the thalamus at onset, but also had a delay of 3 and 6 days before developing WWE. This discrepancy might suggest that those patients with lesions distant from the thalamus may require a longer period to accumulate enough edema to affect the ipsilateral thalamus. Given that the thalamus in most cases seems affected by mass effect, it is reasonable to expect that the contralateral thalamus could also be affected simultaneously, even if to a lesser extent, which could account for many of the features that have been interpreted to support the “midline shift hypothesis.” The cases we describe may be due to unique circumstances whereby a lesion produced just enough edema to cause ipsilateral thalamic dysfunction resulting in fluctuating consciousness, without affecting the FIG. 4. Examples of eye movements in Patient 1 during state of relative alertness. Patient supine in hospital bed, being called to look toward examiner standing at the patient’s left side. Bipolar EOG shows large volitional leftward saccades followed immediately by slow conjugate rightward deviation of the eyes (*). The velocity of the rightward drift increases abruptly when the patient blinks (^) between the first and second saccades. Right-beating jerk nystagmus noted intermittently, most evident superimposed on slow spontaneous rightward eye drift (^^). EOG, electrooculography. Gilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 421 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 5. During relatively alert state without WWE in Patient 1, EOG shows conjugate rightward eye drift on eyelid closure. EOG, electrooculography; WWE, wrong way eyes. contralateral thalamus, and consistent with the absence of EEG abnormalities indicative of contralateral thalamic dysfunction. The thalamus has been implicated in smooth pursuit mechanisms (16), which may be anatomically and mechanistically relevant to the WWE phenomenon. In conclusion, the EEG findings obtained in these 2 patients add significantly to the body of evidence showing that seizures are not the cause of WWE. Our findings implicate a single dysfunctional hemisphere being sufficient to produce WWE, arguing against the idea that disruption of contralateral descending horizontal gaze pathways is the cause of this phenomenon. Instead, these data support a pursuit imbalance and the smooth pursuit asymmetry hypothesis. Although the ipsilateral thalamus may be implicated, further study is needed to better understand the specific neuroanatomical structures in the dysfunctional hemisphere that are most responsible for the WWE phenomenon. STATEMENT OF AUTHORSHIP Conception and design: K. Gilani, F. J. Tyndel, R. A. Wennberg; Acquisition of data: K. Gilani, F. J. Tyndel, R. A. Wennberg; Analysis and interpretation of data: K. Gilani, F. J. Tyndel, R. A. Wennberg. Drafting the manuscript: K. Gilani, F. J. Tyndel, R. A. Wennberg; Revising the manuscript for intellectual content: K. Gilani, F. J. Tyndel, R. A. Wennberg. Final approval of the completed manuscript: K. Gilani, F. J. Tyndel, R. A. Wennberg. REFERENCES 1. Fisher CM. Some neuro-ophthalmological observations. J Neurol Neurosurg Psychiatry. 1967;30:383–392. 422 2. Johkura K, Nakae Y, Yamamoto R, Mitomi M, Kudo Y. Wrongway deviation: contralateral conjugate eye deviation in acute supratentorial stroke. J Neurol Sci. 2011;308:165–167. 3. Tijssen CC. Contralateral conjugate eye deviation in acute supratentorial lesions. Stroke. 1994;25:1516–1519. 4. Keane JR. Contralateral gaze deviation with supratentorial hemorrhage. Arch Neurol. 1975;32:119–122. 5. Sharpe JA, Bondar RL, Fletcher WA. Contralateral gaze deviation after frontal lobe haemorrhage. J Neurol Neurosurg Psychiatry. 1985;48:86–88. 6. San Pedro Murillo E, Rodríguez López A, Morales García E, Herrero San Martin A. Wrong-way deviation due to thalamic ischemic stroke: a new association. Neurologia. 2021;36:479–480. 7. Leichnetz GR. The prefrontal cortico-oculomotor trajectories in the monkey. J Neurol Sci. 1981;49:387–396. 8. Pessin MS, Adelman LS, Prager RJ, Lathi ES, Lange DJ. “Wrong-way eyes” in supratentorial hemorrhage. Ann Neurol. 1981;9:79–81. 9. Gloor P, Ball G, Schaul N. Brain lesions that produce delta waves in the EEG. Neurology. 1977;27:326–333. 10. Brenner RP. The interpretation of the EEG in stupor and coma. Neurologist. 2005;11:271–284. 11. Lekwuwa GU, Barnes GR. Cerebral control of eye movements. I. The relationship between cerebral lesion sites and smooth pursuit deficits. Brain. 1996;119:473–490. 12. Gottlieb JP, Bruce CJ, MacAvoy MG. Smooth eye movements elicited by microstimulation in the primate frontal eye field. J Neurophysiol. 1993;69:786–799. 13. Fukushima K, Sato T, Fukushima J, Shinmei Y, Kaneko CRS. Activity of smooth pursuit-related neurons in the monkey periarcuate cortex during pursuit and passive whole-body rotation. J Neurophysiol. 2000;83:563–587. 14. Sharpe JA. Pursuit defect nystagmus. Ann Neurol. 1979;6:458. 15. Sharpe JA, Lo AW, Rabinovitch HE. Control of the saccadic and smooth pursuit systems after cerebral hemidecortication. Brain. 1979;102:387–403. 16. Tanaka M. Involvement of the central thalamus in the control of smooth pursuit eye movements. J Neurosci. 2005;25:5866–5876. Gilani et al: J Neuro-Ophthalmol 2023; 43: 417-422 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.