Asymmetric Responses in Cortical Visually Evoked Potentials to Motion Are Not Derived from Eye Movements

Abstract

purpose. Normal neonates and many adults after abnormal visual development have directional preferences for visual stimulus motions; i.e., they give better responses for optokinetic nystagmus (OKN) and visually evoked potentials (VEPs) in one direction than to those in the opposite direction. The authors tested whether the VEP responses were asymmetrical because of abnormal eye movements.

methods. VEPs were recorded from the visual cortices of five macaque monkeys: one normal, one neonate, and three reared with alternating monocular occlusion (AMO). They were lightly anesthetized, followed by paralysis to prevent eye movements. They then had“ jittered” vertical grating patterns presented in their visual fields. The steady state VEPs were analyzed with discrete Fourier transforms to obtain the amplitudes and phases of the asymmetries.

results. The normal, control monkey had small, insignificant amplitudes of its asymmetrical Fourier component and random phases that were not 180^o out of phase across the left and right eyes. The neonatal monkey and the AMO monkeys all had large, significant asymmetries that were approximately 180^o out of phase between the left and right eyes.

conclusions. The neonate and abnormally reared monkeys continued to have asymmetrical responses even after their eyes were paralyzed. Therefore, eye movements cannot be the source of the asymmetrical amplitudes of the VEPs, and the visual cortex is at least one source responsible for asymmetries observed in neonates and adults reared under abnormal visual inputs.

Infant humans and monkeys have asymmetric responses to moving visual stimuli. For example, the slow phase of the monocular optokinetic nystagmus (OKN) response is stronger for motion in the nasalward than in the temporalward direction. ¹ These asymmetrical monocular OKN responses to opposite directions of motion disappear after approximately 5 months or 5 weeks for normal humans and monkeys, respectively. ² However, if infants are exposed to abnormal visual inputs (such as those created by cataracts or strabismus) that disrupt normal binocular development, the asymmetry, as well as other abnormalities, may remain present even in the adult. ³ Therefore, determining the neural source of these asymmetrical responses has clinical relevance for diagnosis and treatment of human patients and also has relevance to basic research aimed at elucidating the mechanisms underlying the normal and abnormal development of motion processing in the brain. 

Initial speculation about the neural source of asymmetric responses to motion focused on the nucleus of the optic tract, which is known to be a primary brain region involved in generating the OKN response. ⁴ However, there is now accumulating evidence that motion asymmetries are more ubiquitous and also can be revealed in measurements of smooth pursuit eye movements, visually evoked potentials to motion (MVEP), and perception of motion. ¹ These latter findings have led some to propose that the visual cortex could be partly or fully responsible for causing the asymmetrical responses and that they are not the result of nystagmus. ^{3 5}  

An alternative explanation for the asymmetric cortical MVEP responses has been proposed that argues asymmetrical movement of the eyes is responsible for motion asymmetry. ⁶ This eye movement explanation takes two forms. The first and most simple explanation is that the cortical asymmetries are accounted for by eye movements that take place during the measurement of the asymmetry. For example, some of the children with motion asymmetry also have latent nystagmus. The nasal drift that is present during latent nystagmus would be expected to cancel the velocity of motion in the nasal direction and enhance velocity in the temporal direction. Thus, the cortical response would be expected to be asymmetrical in response to stimulus motion of equal velocity in the two directions. The second possible explanation is more complex. An individual with asymmetric eye movements, such as latent nystagmus, is subjecting the brain to prolonged periods of asymmetrical motion input that may lead to motion adaptation effects. ^{6 7} Thus, even if the eyes are prevented from moving during measurement of a response to motion, the response may be asymmetric due to the lingering effects of motion adaptation that built up earlier during the rearing period. 

The purpose of the present study was to test directly both forms of the eye movement explanation of motion asymmetry. We recorded MVEPs from anesthetized, paralyzed monkeys in which all eye movements were eliminated during the recording session. Three animals were reared under conditions of binocular visual deprivation and were known to have asymmetrical cortical MVEPs when tested under nonparalyzed conditions (unpublished data). Our hypothesis was that the asymmetrical MVEPs would still be present under paralyzed conditions. If this could be shown to be correct, then it would allow us to reject the first form of the eye movement explanation. 

Second, we recorded from a normal neonatal monkey only 3.5 weeks after birth, also under anesthetized, paralyzed conditions. Our hypothesis was that this neonate would show the expected neonatal asymmetric MVEP response that we have reported previously under nonparalyzed conditions. ² Normal neonatal monkeys do not have asymmetrical eye movements, such as latent nystagmus. Thus, our rationale was that if our hypothesis could be confirmed, then it would allow us to eliminate, at least as a necessary condition, the second form of the eye movement explanation as well. 

Finally, as a control, we made similar measurements in a normal juvenile monkey that had been reared under normal conditions. Our hypothesis was that this juvenile would show a symmetric cortical MVEP response under the paralyzed conditions. This was done to rule out the unlikely possibility that all monkeys show an asymmetrical MVEP response when measured under anesthetized, paralyzed conditions. All three of our hypotheses were confirmed. We conclude that asymmetrical cortical MVEP responses do not depend on asymmetric eye movements. 

Methods

Subjects

Five Macaca mulatta monkeys were studied in accordance with ARVO and NIH guidelines for use of animals in research. These monkeys were obtained and reared at the Yerkes Regional Primate Research Center. Three were reared with a daily alternating monocular occlusion (AMO), using opaque soft contact lenses. That is, starting on the day of birth, an opaque occluder contact lens was placed on one eye for 24 hours, followed by removal and placement on the fellow eye for the next 24 hours. Using this protocol, each eye was deprived of vision every other day, and no binocular input was allowed. Compliance with the lens wear protocols is shown in Table 1 . These animals were recorded at 1 to 2 years of age. The fourth monkey was a normally reared neonate recorded 3.5 weeks after birth. The fifth monkey was a normal juvenile reared under normal binocular visual conditions and recorded at 4 years of age. 

All monkeys were given ophthalmic examinations to ensure that no eye pathology was present and that the eyes wearing contact lenses remained in good condition. The three AMO monkeys all exhibited an exotropia, but had acuities within the normal range for each eye. 

Recordings

Eye movements were stabilized with a standard preparation used to eliminate eye movements during electrophysiological recordings from the visual system. Briefly, ketamine (15 mg/kg) was used to initially anesthetize each monkey, followed by insertion of an intravenous needle and infusion of sodium pentobarbital (5–10 mg/kg) for further anesthesia, which was titrated to the proper level by closely monitoring blood pressure, heart rate, and electroencephalogram (EEG). A rectal temperature probe monitored and a heating blanket maintained the body temperature at 38°C. After intubating the trachea, gallamine triethiodide was given (10 mg/kg) to block muscle movements, and the monkey was ventilated artificially to maintain the expired CO₂ level at the pre-paralysis level (∼5%). A 70%/30% mixture of N₂O/O₂ was given as the inspired respiratory gas. 

Contact lenses were placed on the eyes to prevent corneal drying. For the control monkey and two of the AMO monkeys, subdermal needle electrodes were inserted to record VEPs. Two of these electrodes were placed over each occipital lobe near the representation of the foveal region of the striate cortex, i.e., approximately 1 cm lateral to the midline and just above the nuchal ridge. A reference electrode was placed over the parietal cortex at the midline, and a ground electrode was placed over the forehead. The differential signals (sampled at 450 Hz) from the subdermal occipital electrode leads were amplified, filtered (1–100 Hz), and fed into a computer system for monitoring, storage, and analysis. The amplitudes of these signals varies. This system has been thoroughly described previously and carries out a Fourier analysis of the steady state waveforms of the electroencephalograms (see below). ⁸  

The third AMO monkey first had scalp recordings as before and then received a craniotomy over the left occipital cortex to record the VEPs with bipolar, low-impedance microelectrodes inserted into the cortex. The microelectrodes were held 300 μm vertically apart to restrict the differential recordings to the striate cortex. The electrodes were driven into the striate cortex with a microdrive and recordings were obtained at various depths. Some of the data from this monkey have been reported previously. ⁵  

Stimuli

Each monkey faced a video monitor (19-inch diagonal) at 30- to 50-cm distance from the eyes. A vertical sinewave grating pattern (0.25–3 cycles/deg) was generated on the screen (80% contrast and average luminance of 110 candela/m²) and moved horizontally back and forth through a 90^o phase shift at a frequency of 6 Hz in a square wave manner. Stimulus frequency was an even division of the video frame rate (60 Hz). This appears as a horizontal “jitter” to a human observer and has been found to generate a large amplitude VEP for both humans and monkeys. ^{2 8} Each trial consisted of 10 seconds of the stimulus and 5 to 15 trials were vector-averaged. 

Analyses

A discrete Fourier transform was carried out on the steady state VEPs to obtain the amplitudes of the EEG components at various temporal frequencies. The two frequencies of interest were 6 Hz (the fundamental or F1 frequency) and 12 Hz (the stimulus doubled or F2 frequency). Frequencies on either side of both of these frequencies were used to estimate the “noise” or non–stimulus-derived levels of the EEG. ⁸ In addition to the amplitudes of the EEG components measured at various frequencies, the phase of each frequency component was also measured during the stimulus presentation. Phase coherence was used as an indicator that the measured signal was derived from the stimulus and not just part of the normal EEG (using the Circle t-test, which is similar to a Student’s t-test but is used for circular values). ⁸  

Eye Movement Measurements

We performed a detailed study of oculomotility in one alert monkey raised with AMO (RCl5). We used an electromagnetic method employing scleral search coils to measure accurately movements of both eyes. Our system (CNC Engineering, Seattle, WA) is drift-free and capable of measuring eye movements to 15 minutes of arc resolution. We precisely calibrated the signal obtained from each eye coil by requiring the monkey to monocularly fixate a stationary target at different eccentricities on a tangent screen 57 cm distant. Each eye was calibrated to an accuracy of at least 0.5° for target positions within ±20° (horizontal and vertical) of straight ahead gaze. In addition, the monkey was trained to perform accurate saccadic, smooth pursuit, and vergence eye movements on a small diameter (0.1°) laser spot. 

Results

The MVEP was obtained monocularly while the other eye was blocked from viewing the stimulus. Control trials also were obtained to ensure that there was no EEG component at our stimulus frequency when both eyes were blocked from viewing the stimuli. During each cycle of the grating pattern movement (first in one direction and then back in the opposite direction), an MVEP is generated in a normal animal. As a result, there is a large signal obtained at the F2 frequency (12 Hz), which is twice the stimulus frequency (6 Hz). Furthermore, this signal always exhibits phase coherence. This demonstrates that our system is properly recording the EEG evoked by our visual stimulus. For normally reared juvenile or adult monkeys, including our control for this experiment, there is no significant amplitude for the F1 signal (6 Hz) and thus no asymmetry of the MVEP for one direction of the movement (Fig. 1) . For the AMO and neonate monkeys, a significant component of the EEG was obtained at the F1 frequency. This demonstrates that there was a stronger MVEP generated in one direction of the stimulus movement than the opposite direction (Fig. 1) . ⁸  

Both the F1 and F2 components of the Fourier analyses are expected to show phase coherence if they are stimulus-driven. Furthermore, when each eye is tested monocularly, the phase of the F2 for each eye is expected to be similar because each eye generates an MVEP at about the same time relative to the stimulus movements in both directions. However, if each eye has a preference for the temporal-to-nasal direction (or the reverse), then the F1 phases are expected to occur at approximately 180^o apart for the two eyes. This is because the temporal-to-nasal directions are opposite for the two eyes. 

The normal and the AMO monkeys all exhibited similar phases for the F2 signals from the two eyes. The three AMO monkeys also all exhibited approximately a 180^o phase shift of the F1 signal between the two eyes (Fig. 1 ; 156^o± 12^o; mean ± SD). 

The third AMO monkey was recorded with a bipolar set of microelectrodes that were 300 μm vertically apart and were driven through the striate cortex perpendicular to the pial surface to record the MVEPs. Asymmetrical MVEPs were first recorded in layers II/III with amplitudes increasing as the electrodes were driven deeper. At the deepest position of the track, a small lesion was made by current passed through the electrode at the end of the recording session. The lowest position of the electrodes gave the largest F1 signal that we recorded, and the track produced was observed to end in upper layer IV of the striate cortex. It is possible that greater amplitude signals might have been obtained if the electrode had been moved deeper into the infragranular layers. The data from this monkey have been published previously, ⁵ but are included it here to emphasize that the MVEP asymmetry is derived at least partially from the striate cortex and demonstrate that the cortical asymmetries matches the MVEP asymmetries of our other AMO monkeys. 

Eye movement studies were performed for AMO monkey RCl5 when the animal was between 1.5 and 2.5 years of age. These measurements revealed a 12° exotropia of the left eye (57 cm viewing distance). Furthermore, a strong preference for fixation with the right eye was present, despite normal acuity in each eye. Finally, a leftward (slow phase) nystagmus of 5°/s was present, but the direction or velocity of drift did not change, regardless of which eye was viewing. Therefore, this animal had no measurable latent nystagmus. 

Discussion

Our results show that asymmetrical cortical MVEP responses do not depend on eye movements. Visually deprived adult monkeys that exhibit an MVEP asymmetry when recorded under unparalyzed conditions continue to have asymmetrical MVEPs when tested under paralyzed conditions. Normal monkeys, including our control monkey, have symmetrical responses when tested under the same conditions. Furthermore, our recordings from a neonate recorded under paralyzed conditions only 3.5 weeks after birth demonstrate that asymmetrical cortical MVEP responses do not require a prior period of motion adaptation, as might occur in an individual that grows up with asymmetrical eye movements. 

It should be emphasized that our results do not rule out eye movement explanations for the cortical motion asymmetries that are present in some cases. ⁶ In other words, our neonatal data demonstrate that eye movements are not a necessary condition for an asymmetry to be present, but do not rule out the possibility that they might be a sufficient condition. At the least, our results demonstrate that eye movements cannot explain all cortical motion asymmetries. 

Our eye movement recordings in one AMO monkey revealed no measurable latent nystagmus. However, a conjugate slow phase drift (leftward) was present when tested in the light or dark. The slow phase drift velocity was always low (<5°/s) and did not prevent accurate fixation or eye movements. Whether such a unidirectional slow phase drift could contribute to the motion asymmetry measured in this animal is uncertain. However, the unidirectional nature of oculomotor drift for both eyes did not correlate with MVEP data showing a 180° phase difference across the eyes in this animal. Furthermore, the other AMO monkeys generated asymmetrical evoked potentials under paralyzed conditions during which no eye movements were possible. 

Important questions concerning motion asymmetries include their localization within the brain and their relationship to binocularity. The latter question arises because disruptions of normal binocular input during development result in a number of abnormalities of the visual system such as amblyopia, strabismus, and latent nystagmus. ⁹ If normal binocular inputs are major factors for achieving symmetrical MVEPs derived from directional motion, then the striate cortex would be the first stage of visual processing where the neurons have a combination of directional selectivity and binocularity as properties of their receptive fields. 

On the other hand, the middle temporal visual cortical area (MT or V5) appears to be particularly important for processing motion information. This area receives direct projections from the striate cortex and also has projections to the NOT in the brain stem. ¹⁰ It might be expected, therefore, that in the presence of motion asymmetry, directional sensitivity in MT might be affected. However, in a recent study of the MT cortex of strabismic monkeys, a complete representation of directional preferences was found in the recorded population of single units. ¹¹ This was despite finding a significant asymmetry in smooth pursuit performance favoring nasally directed tracking. Unfortunately, the researchers were unable to identify whether there was any difference in the directional sensitivity for the critical population of lamina V neurons that project to oculomotor brain stem areas involved in eye movements. Thus, their results do not rule out the possibility that it may have been only a subpopulation of neurons within MT that had a directional preference. 

Our finding of an asymmetry in V1 suggests another possibility. The NOT is known to be a controlling area for OKN functioning and other horizontal motion effects, and the striate cortex itself has a direct, ipsilateral projection to the NOT. ⁴ A subpopulation of neurons within V1 might be responsible for producing motion asymmetries that are measured in the VEPs and passed on to the controlling neurons of the NOT. 

In summary, we have shown that there is a motion asymmetry neural signal derived from the striate cortex that is not produced by eye movements. This cortical asymmetry could be the basis for many abnormalities that are seen in monkeys and humans that receive abnormal visual inputs during development. 

 Supported by the National Institutes of Health Grants RR-00165 and EY-05975.

 Submitted for publication June 3, 1998; accepted August 11, 1998.

 Proprietary interest category: N.

 Corresponding author: James R. Wilson, Yerkes Research Center, Emory University, Atlanta, GA 30322. E-mail: [email protected]

Table 1.

 

 

Individual Monkey Backgrounds

Table 1.

 

Individual Monkey Backgrounds

Monkey	Treatment	Lens compliance (%)			Length of lens wear (mo)	Age at recording (y)
				OD			OS
RTl3	AMO	48	50	4	1.75
RCl5	AMO	48	49	6	1.5
RWo5	AMO	48	51	4	1.1
RSd6	Normal	—	—	—	3.5*
RLm3	Normal	—	—	—	4.0

 AMO stands for alternating monocular occlusion. Lens compliance is the percentage of total time during rearing that an occluder lens was worn on the eye.

*  Age in weeks.

Figure 1.

 

Polar plots showing the amplitudes and phases of the asymmetrical component of visually evoked potentials recorded from the visual cortices of five macaque monkeys: one normal adult (RLm3), one normal neonate (RSd6), and 3 monkeys (RTl3, RCl3, RWo5) reared with alternating monocular occlusion (AMO). The right eye’s phases and amplitudes are designated by a small solid circle and the left eye’s by a small open circle. The larger circles around the end points are 1 SE for the vector of the phase and amplitude trial values. Note that the phases of the VEPs of the neonate and AMOs are approximately 180^o apart. All eyes for these monkeys had significant phase coherence (P < 0.05) for the asymmetrical component. Also note that the amplitudes for the asymmetrical responses are much larger than those of the normal control monkey, which did not have a significant phase coherence for the asymmetries for either eye. All the monkeys had large amplitudes and very significant (P < 0.01) phase coherences for the F2 (symmetrical) component for both eyes (not shown because of large values beyond the scales needed to clearly show the asymmetrical values).

 

The authors thank Jean Torbit for secretarial assistance and the veterinarians and caretakers of the Yerkes Regional Primate Research Center for expert assistance with care of the animals used in this project.