Effect of binocular rivalry suppression on initial ocular following responses

Abstract

To study the effect of binocular rivalry (BR) suppression on the ocular following response (OFR), we recorded the OFR in both the suppressed and the dominant phases of BR. The BR was established using stationary horizontal/vertical grating patterns presented on two PC monitors. Once a subjective image of a vertical or horizontal grating pattern was perceived, subjects pressed a button to trigger an onset of brief horizontal movement (750 ms) of the vertical grating pattern and an offset of the horizontal pattern. The OFRs were recorded using a scleral search coil system at 1 kHz. The OFRs from the suppressed phases were significantly reduced compared to those from the dominant phases. The OFRs were asymmetrical to temporalward and nasalward motion in most conditions. We suggest that asymmetry of OFRs under the incomplete BR conditions may be a reflection of imbalance binocular inputs and processing in the visual system similar to asymmetrical optokinetic nystagmus in strabismic subjects. The latency of the OFR in deeper suppressed conditions was prolonged, suggesting that the interaction of BR and OFR may occur at multiple stages including an early stage of the visual processing. The OFR may have the potential for objective measurement of BR suppression in clinical evaluation of binocular function.

Introduction

The ocular following response (OFR) elicited by a sudden movement of a large textured pattern operate in a machine-like fashion with ultra-short latencies (<85 ms in humans) to help stabilize the retinal images of objects on the plane of fixation (Miles, Kawano, & Optican, 1986). It uses relatively low-level, pre-attentive cortical processing of visual motion stimuli to extract control signals that initiate eye movements before the observer is even aware that there has been a motion stimulus (Miles, Busettini, Masson, & Yang, 2004; Sheliga, Chen, Fitzgibbon, & Miles, 2005). The OFRs' rapid, reflexive nature might be attributed to low-level motion detectors (Borst & Egelhaaf, 1993). Evidence from monkeys strongly suggests that the OFR is mediated, at least in part, by the medial superior temporal (MST) area and middle temporal (MT) area of the cortex (Takemura & Kawano, 2006; Takemura, Murata, Kawano, & Miles, 2007). Of particular interest here is whether the low-level OFR has any interaction with binocular rivalrous perception. In this study, we utilized perceptual alternations of binocular rivalry (BR) to investigate how and to what extent the OFR varied under the BR condition. 

When each eye sees discordant images, alternating perception of one or another image occurs. This is called binocular rivalry (BR). It has been reported that during the suppressed phase of BR, contrast sensitivity of form and motion discrimination were low and the onset of visual motion produced abnormally long reaction time (Fox & Check, 1968; Li, Freeman, & Alais, 2005; Nguyen, Freeman, & Alais, 2003; Nguyen, Freeman, & Wenderoth, 2001). Unfortunately, the motion detection in those experiments was made based on the subject's subjective judgments because no eye movements were monitored, and it was not clear whether and how eye movements were influenced in those experiments. Other experiments showed a correlation between the eye movements driven by rivalrous motion and the perception of the dominant eye (Fox, Todd, & Bettinger, 1975; Logothetis & Schall, 1990; Wei & Sun, 1998). However, eye movements driven by inputs from the suppressed eye have not been reported. According to electrophysiological and fMRI studies, neural activities at different levels including lateral geniculate nucleus, V1, V3A, LOC, and V5/MT are correlated with perceptual alternations of binocular rivalry (Blake & Logothetis, 2002; Leopold & Logothetis, 1996; Logothetis & Schall, 1989; Moutoussis, Keliris, Kourtzi, & Logothetis, 2005; Polonsky, Blake, Braun, & Heeger, 2000; Wunderlich, Schneider, & Kastner, 2005). We expect that the OFR would be affected by BR suppression partly because both are processed at some common sites (at least at MT and the adjacent area of the MST) in the visual system. 

In optokinetic nystagmus (OKN) studies on patients with defective binocular vision, it is generally held that asymmetrical OKN is an indicator of abnormal binocular vision (Steeves, Reed, Steinbach, & Kraft, 1999; Valmaggia, Proudlock, & Gottlob, 2003; Wong, Foeller, Bradley, Burkhalter, & Tychsen, 2003), although different opinions exist (Wright, 1996). However, these studies used subjects with damaged binocular function. The mechanism of asymmetrical OKN remains unclear. In the present study, we focused on the initial open-loop part of OKN, which is referred to as the OFR to demonstrate that OFRs are asymmetrically biased in the nasalward direction under binocular rivalry. The asymmetrical OFRs may help us to better understand the physiological basis of asymmetrical OKN. Because of its reflexive nature, the OFR may be a useful objective measurement of binocular function. 

Methods

Participants

Four subjects (24 to 45 years old) participated in this study. Two of them were experienced scientists and two were paid naive subjects. All four had normal best-corrected visual acuity and were taking no neurologically acting medications. The protocol was approved by the Institutional Review Board of The University of Pittsburgh. All procedures observed the declaration of Helsinki, and informed consent was obtained on all subjects. 

Experiment 1: BR psychophysical experiments

Dichoptic stimuli were established using a mirror haploscope consisting of two computer monitors which had a resolution of 1280 × 1024 pixels and a vertical refresh rate of 70 Hz. A circle stimulus pattern of white-black gratings was displayed on each monitor. The size of the circle subtended 30° at a viewing distance of 47 cm. The width of a white or black bar of square-wave gratings was one degree at this viewing distance. To create BR, the pattern with vertical gratings was presented to the left eye and horizontal gratings to the right eye. The luminance of the patterns was 0.5 cd/m ² for black bars and 85 cd/m ² for white bars. The right pattern and left pattern had equal luminance (see Figure 1A). 

Figure 1

 

Stimulus settings. Subjects viewed dichoptic stimuli through a mirror haploscope. (A) Stimuli used in Experiment 1. (B) stimuli in Experiment 2. (C) Stimulus in Experiment 3. (D) monocular stimuli in Experiment 3. The L/R: right/left pattern. Arrows under left pattern indicate movements of the pattern.

Subjects were seated with their heads stabilized on a chin and forehead rest by means of a head band. When subjects stared at the centers of the two stationary grating patterns, they perceived alternating images in which horizontal or vertical gratings were switching back and forth in the central area, and mixed horizontal/vertical gratings were coexisting in the peripheral area. The dominant central area was usually 20° to 25° by estimation of the subjects. Subjects responded to each of the transitions of the horizontal/vertical images by pressing a button. Sometimes the central area of the perceived image had mixed images of vertical and horizontal gratings. In this case, subjects could stop responding by pressing an error button to abort the mixed trial. There was no time restriction on their waiting time. They could re-start whenever they were ready. Durations of subjective visual dominance and suppression of BR were recorded with the REX-VEX system (from LSR, NEI of NIH, http://www.lsr-web.net). 

Experiment 2: OFR in BR suppressions

The same equal-luminance stimuli for the psychological experiments were used for the OFR eye movement experiment. For each experimental session, subjects performed eye-movement calibration tasks by fixating on a 0.1° dot displayed at different locations with left eye viewing only. Only the left eye was recorded and analyzed. 

Subjects were first presented with two stationary dichoptic stimuli (horizontal grating pattern for the right eye and vertical grating pattern for the left eye) and then they started to perceive the alternation of the two patterns. However, subjects in the eye movement experiment did not need to react to each transition of the BR. They waited as long as needed for cued dominant images with long durations and then responded to them by pressing a button to trigger a motion of OFR and an offset of horizontal pattern. A small (1° × 0.3°) horizontal or vertical black bar was presented as a cue in the middle of the right pattern for 500 ms at the beginning of each trial to inform the subjects to wait for a perception of a cued image with horizontal or vertical gratings. Subjects were instructed to wait for a long reliable perception of a cued image and to allow any short transitions to pass. There was no time restriction for this. When an image of horizontal gratings was cued, the trial was labeled as a suppressed trial for the left eye, and when an image of vertical gratings was cued, the trial was labeled as a dominant trial. Once a subject perceived a cued image, he/she quickly pushed a button to trigger a brief horizontal movement of the vertical grating pattern either temporally or nasally at a speed of 10°, 20°, 30°, 40°, or 50°/sec. See the sequence of stimuli in Figure 2. Only the left eye was stimulated by OFR motion. As in the psychophysical experiments, the perceived image was usually dominated by horizontal or vertical gratings in the central area of 20° to 25° by the subjects' estimation. To ensure that the BR suppression and dominance were reliable and consistent, we excluded mixed trials and trials with short transitions. When the central area of the image was mixed with gratings of un-cued orientation, it was considered a mixed trial. When a perceived image altered in orientation before or at the same time the triggering button was pressed, it was ruled as a short trial. The mixed trials and short trials were labeled for offline exclusion by pressing an error button after the motion had been executed. Thus, the selected trials were reliable and consistent throughout the experiment, although the BR was incomplete under the equal luminance condition. At the beginning of the horizontal movement of the vertical pattern, the horizontal pattern was turned off, the movement of the vertical grating pattern lasted for 750 ms. 

Figure 2

 

Stimulus sequence. A. OFR during suppressed phase of BR; B. OFR during dominant phase of BR. SBs: subjects. H: horizontal. V: vertical. The double headed arrows indicate movements of the vertical pattern.

Subjects were instructed to keep a clear image of the moving pattern and not to subjectively follow any individual bar after the motion was triggered. A control condition using a stationary pattern was intermixed with the experimental conditions in order to eliminate some default eye movements. A photo cell was placed on the upper-left corner of one monitor to catch the first frame of any stimulus movements in order to accurately align the eye movement with the onset of moving stimuli. 

For the purpose of comparison, the OFR under monocular viewing condition was recorded. In this condition, a single vertical grating pattern was presented to the subject's left eye without any BR stimuli on the right side. 

Experiment 3: OFR in complete BR suppressions

The purpose of these experiments was to create a complete BR suppression on one of the two patterns. The pattern to be completely suppressed was the left vertical grating pattern which also moved to evoke the OFR. The luminance of the right pattern was the same as that of the equal luminance experiment (0.5 cd/m ² for black bars and 85 cd/m ² for white bars), but the luminance of the left pattern was reduced to 0.17 cd/m ² for black bars and 0.3 cd/m ² for white bars. When the left dim stimulus and the right bright stimulus were presented to the left and the right eye, the left image was rarely dominant due to the stronger competition of the bright image in the right eye ( Figure 1C). Subjects pressed the button to trigger the onset of horizontal movements of the dim pattern when they were certain that the bright horizontal grating pattern was dominant without seeing any image of other orientation. While obtaining a complete suppressed phase of the dim vertical pattern was easy, it was very difficult to obtain a dominant phase of the vertical pattern under this unequal luminance condition. To have a condition where the dim vertical pattern was not suppressed, the bright horizontal pattern was replaced with a dim blank field that had a luminance of 0.25 cd/m ². This actually created a monocular viewing condition ( Figure 1D). Under this monocular condition, subjects could clearly see the vertical grating pattern. Onset of movement was triggered by subjects when they fixated at the centre of the vertical pattern and got a clear image of the pattern. The pattern moved at velocities of 5, 10, 20, 30, and 40°/sec in nasalward or temporalward direction. A lower velocity of 5°/sec was added and a high velocity of 50°/sec was excluded because the velocity of 50°/sec did not induce reliable OFR under the low luminance condition. All conditions, including the suppressed, monocular, nasalward and temporalward, were randomized. 

Data acquisition

Durations of dominant/suppressed phase of the BR were recorded with the REX-VEX real-time experimental system and were automatically measured with software programs written in MatLab offline. 

Horizontal and vertical eye movements of the left eye were recorded with an electromagnetic technique using scleral search coils embedded in a silastin ring (Remmel Labs). Following application of 1–2 drops of anesthetic (Proparacaine HCl), coils were placed in the left eye. The AC voltages induced in the scleral search coils were led off to a phase-locked amplifier that provided separate DC voltage outputs proportional to the horizontal and vertical positions of the eye with corner frequencies (−3 dB) at 1 K Hz. Peak-to-peak voltage noise levels were equivalent to an eye movement of 1–2 min of arc. Eye movement data, along with timing marks and stimulus codes, were recorded using REX-VEX real-time experimental system (Laboratory of Sensorimotor Research, NEI of NIH). Usually 20 blocks of data were collected in each experimental session. Each block consisted of 22 trials {5 speeds × 2 directions × 2 viewing conditions [dominance (or monocular) & suppression] + 2 control condition}. More than 100 trials for each condition were obtained from each subject in multiple sessions. Coil wearing time for each session was approximately 30 minutes. All stimulus conditions were randomized. 

Data analysis

The horizontal and vertical eye positions obtained from the calibration were each fitted with a third-order polynomial to linearize the horizontal eye position data recorded during the experiment. Horizontal eye position signals, once smoothed with a cubic spline function of weight 10 ⁷, were then two-point backward differentiated to derive the horizontal eye velocity traces. Rightward movements (nasalward for the left eye) were defined as positive. The latency of initial OFR was estimated using a mathematical algorithm (onset of OFR was decided with a relative velocity criterion of 1.5°/sec) and visual inspection of the mean of the velocity profiles afterward. The amplitude of the initial OFR was calculated by averaging about 100 trials of responses in an open-loop time window of 90–170 ms for equal luminance experiments. For unequal luminance conditions, the open-loop time windows were 120–240 and 200–320 ms for monocular and suppressed condition, respectively. The OFR in the suppressed and dominant phases as well as in the monocular conditions were analyzed and statistically compared. The effect of BR suppression was expressed as a percentage that is calculated by dividing the difference in the OFR velocity between the dominant and the suppressed responses with the dominant responses. For the equal luminance experiments, the following formula ( Equation 1) was used:  

$P=(\mathrm{D}\mathrm{o}\mathrm{m}-\mathrm{S}\mathrm{u}\mathrm{p})/\mathrm{D}\mathrm{o}\mathrm{m}\times 100\%,$

(1)

where the Dom represents the OFR velocity from dominant phases; the Sup from suppressed phases. P is the percentage of the suppression. 

For the unequal luminance experiments, monocular OFR replaced dominant OFR. The following formula (Equation 2) was used:  

$P=(\mathrm{M}\mathrm{o}\mathrm{n}-\mathrm{S}\mathrm{u}\mathrm{p})/\mathrm{M}\mathrm{o}\mathrm{n}\times 100\%,$

(2)

where the Mon stands for velocity of the monocular OFR; the Sup is for the suppressed OFR. P is the percentage of the suppression. 

The asymmetrical index (AI) of the temporal–nasal asymmetry of OFR was calculated with the following formula ( Equation 3):  

$\mathrm{A}\mathrm{I}=(\mathrm{N}\mathrm{a}\mathrm{s}-\mathrm{T}\mathrm{e}\mathrm{m})/\mathrm{N}\mathrm{a}\mathrm{s}\times 100\%,$

(3)

where the Nas represents velocity of the nasalward OFR, and the Tem is for the temporalward OFR. 

Results

Experiment 1: Psychophysical BR experiment

The mean of durations of BR dominance and suppression for three subjects was 1.34 s based on the total 4828 trials. The data are consistent with a Gamma distribution fitting with a high correlation ( R ² = 0.97). See more information in 1. 

Experiment 2: OFR in BR suppressions

To show normal OFR responses, monocular OFR velocity traces from subject CK are displayed in Figure 3. It is clear that the latency of OFRs is about 85 ms. 

Figure 3

 

An example of individual initial ocular following responses (OFR) from monocular viewing conditions for nasalward movement at 20°/sec. The thin lines are velocity traces from individual trial and the thick line is their average. The latency of the responses is indicated as about 85 ms.

The mean velocity of monocular OFR to temporalward and nasalward movements at various speeds in an open-loop window (90–170 ms) is summarized and presented in Figure 4 for each subject. The maximum OFR occurred at stimulus velocity of 20–30°/sec, but the OFR decreased as the stimulus velocity increased. 

Figure 4

 

Mean velocity of OFR to temporalward and nasalward movements at various speeds under monocular viewing conditions. Data from four subjects were calculated from open-loop window of 90–170 ms. Vertical axes shows response velocity. Positive values: nasalward movement; negative: temporalward. Error bars are ±1 SD. No significant difference between nasalward and temporalward OFR ( p > 0.05).

It is also shown that the monocular OFRs are symmetrical for temporalward and nasalward motions. A statistical comparison between temporalward and nasalward responses shows no significant difference (all p > 0.05). 

Effects of BR on OFR

The perceived alternations of horizontal and vertical pattern mainly occurred in the central area of the images. The area of dominance usually occupied about 20–25 degrees of the central area by subjects' estimation. 

Figure 5 displays examples of OFR velocity traces for suppressed and dominant conditions. The latencies of the OFR for suppressed conditions seemed longer than those for the dominant conditions; however, there was no significant difference ( p > 0.05). The mean velocities of the OFR for suppressed (bars with thin-line pattern) and dominant (bars with thick-line pattern) phases in an open-loop time window (90–170 ms) are shown in Figure 6 for four subjects. The velocities of the OFR in all subjects were found to be significantly lower in suppressed phases than in dominant phases at almost all speeds tested (all p < 0.001). The temporalward OFRs were significantly lower than the nasalward OFRs in most suppressed conditions and in some dominant conditions for three subjects and in suppressed conditions only for one subject ( p < 0.05). For the purpose of direct comparison among subjects, the effects of suppression were normalized with Equation 1 and displayed in Figure 7 for all subjects. The OFR velocity was reduced by 20% to almost 40% for the nasalward OFRs and by 40–55% for the temporalward OFRs. A significant difference between the nasalward and temporalward OFRs for velocities of 30°/sec and above ( p < 0.05) was also demonstrated. To show the asymmetry of the OFR, the nasalward and the temporalward OFRs were normalized with Equation 3. Figure 8 displays the summarized data for four subjects. For the lower stimulus velocity, the asymmetry of the OFR for dominant phases was close to that for suppressed phases. For the higher stimulus velocity, the OFR asymmetry for suppressed phases was significantly larger than that for dominant phases ( Figure 8). 

Figure 5

 

Examples of individual OFR under dominance (green) and suppressed (red) conditions for nasalward movement at 20°/sec. The black and blue thick lines are the mean for dominant and suppressed condition respectively. The latency of the responses is indicated as about 85 ms.

Figure 6

 

Mean velocity of OFR for suppressed and dominant phases, as well as nasalward and temporalward conditions for four subjects (open loop window: 90–170 ms). Error bars are ±1 SD. The OFR for suppressed condition (bars with thin lines) are significantly smaller than the OFR for the dominant condition (bars with thick lines) for all subjects in almost all conditions ( p < 0.001). The temporalward OFR are significantly lower than nasalward OFR in all suppressed conditions for all subjects and some dominant conditions for 3 subjects and in suppressed phases only for one subject (DY). Sup-tem: suppressed phase and temporalward motion. Sup-nas: suppressed phase and nasalward motion. Dom-tem: dominat phase and temporalward motion. Dom-nas: dominance phase and nasalward motion.

Figure 7

 

Normalized effects of BR suppression four subjects. Square: nasalward movement; spade: temporalward. Error bars are ±1 SD. The OFR decreased by 20 to 55% for both nasalward and temporalward conditions. It is also shown a significant difference between nasalward and temporalward OFR for velocities above 30°/sec ( p < 0.05).

Figure 8

 

Temporal-nasal asymmetrical index (AI) for four subjects. Square: dominant conditions; spade: suppressed condition. Error bars: ±1 SD. A significant difference is shown between dominance and suppressed conditions for velocities above 30°/sec ( p < 0.05).

The OFR from dominant conditions was also compared with the OFR from monocular conditions. 

In three subjects, the monocular OFR was significantly larger than the OFR in the dominant condition ( p < 0.01). In one subject, the monocular responses were smaller than the dominant responses ( p < 0.01). 

Experiment 3: OFR in complete BR suppressions

Figure 9 shows examples of OFR velocity traces from subject MZ. Dashed lines represent the mean velocity traces from suppressed phases and solid lines represent the velocity traces from monocular conditions. Under the lower luminance condition, the latency of the OFR was prolonged for both monocular conditions (116.8 ± 7.9 ms) and suppressed conditions (194.8 ± 12.8 ms) compared with the higher luminance condition (85 ms). The mean velocities of the OFRs from suppressed conditions were significantly lower than those from monocular conditions (all p < 0.004) except for one condition from subject CK for nasal movements at 30°/sec. The effects of BR on the OFRs for all subjects were calculated with Equation 2 and are presented in Figure 10. The OFRs were reduced by 60% to 85% compared with the monocular OFR under the same luminance conditions. There was no significant difference between nasal and temporal OFR ( p > 0.05). 

Figure 9

 

Examples of OFR velocity traces for various stimulus speeds. Each trace is an average of multiple trials for subject MZ under the unequal luminance conditions. Dotted lines represent suppressed OFR and solid lines represent monocular OFR. The number next to each trace is the stimulus velocity for the response. Different latencies are demonstrated for two different conditions.

Figure 10

 

Normalized effects of BR suppression under unequal luminance condition four subjects. The open loop window for monocular condition was 120–240 ms and that for suppressed condition was 200–320 ms. Square: nasalward movement; spade: temporalward. Error bars are ±1 SD.

Discussion

The present BR psychophysical experiments showed that the duration of dominant/suppressed phase of the BR in most trials lasted more than 1 s (1.34 s). This was similar to those reported in the literature (Sun, Tong, Yang, Tian, & Hung, 2002). The duration of 1s was long enough to allow the OFR motion stimuli to be applied during suppressed or dominant phases. According to a previous study, a subject required a reaction time of about 400 ms to press a button that triggered an appearance of the visual target (Hertle et al., 2002). In the present study, the initiation of the OFR motion was triggered in the middle of a suppressed or dominant phase after a reaction time of about 400 ms. In the case of fast suppressed or dominant transition, subjects aborted the trials at the end of the trial. The present OFR eye-movement experiment, unlike previous psychophysical BR experiments, did not require subjects to respond to each of the alternations; instead, the subjects were allowed to wait, as long as needed, for a reliable phase of dominance and to respond only to the reliable BR trials. Therefore, the BR suppression and the dominance were reliable despite being incomplete. It is true that the transitions of two phases or mixed phases are also important in BR studies (Brascamp, van Ee, Noest, Jacobs, & van den Berg, 2006; Erkelens & Collewijn, 1985; Mueller & Blake, 1989); however, the present study was not designed to cover these aspects. 

The size of the BR stimuli in psychophysical studies was usually small. It was 2° or 4° in diameter in Kovács's (2000) study, 8° in Tong and Engel's (2001) work, and 0.31° in Brascamp et al.'s (2006) work. However, in OKN eye movement studies, the stimulus sizes were large. For example, Sun and his colleagues (2002) used stimuli of 23° × 30° (height × width), Wei & Sun (1998) used stimuli of 18° in diameter, and Erkelens and Collewijn (1985) used 30° × 30°. In the present study, we used circle stimuli of 30° in diameter. It is true that the stimuli were quite large compared with those used in the psychophysical BR studies. With these large stimuli, only the central area of about 20 to 25° was perceived as the dominant area. Obviously, suppression/dominance was incomplete. We tried smaller sizes of the stimuli; however no complete suppression could be achieved, even with the stimuli of 20°. A stimulus of 20° was already too small to induce normal OFRs. It is important to obtain OFRs similar to the previously reported ones for reasonable comparisons. The OFRs under non-suppressed conditions (Figures 3 and 4) were shown to be similar to that in the previous reports (Busettini, Miles, Schwarz, & Carl, 1994; Gellman, Carl, & Miles, 1990; Masson, Busettini, Yang, & Miles, 2001; Yang & Miles, 2003). 

The OFR was significantly reduced when motion was initiated in BR suppressed phases ( Figures 5 and 6). The reduction of the OFR was as much as 55% compared with dominant phases ( Figure 7) for the equal luminance condition; the OFR was reduced by 85% compared with monocular conditions for the lower luminance condition ( Figure 10). The effects of BR on the OFR have not been previously reported although a reduction of OKN eye movements during motion rivalry was previously studied (Logothetis & Schall, 1990). However, in that study, the OKN eye movements were driven by the dominant eye. The decrease in OKN velocity might be the result of incomplete dominance of the dominant eye although the authors had a different interpretation (Logothetis & Schall, 1990). In our study, the OFR from the dominant phases was lower than the OFR from the monocular condition in three of four subjects (p < 0.001), and the OFR from the dominance phases was higher than the OFR from the monocular condition in one of the four subjects (p < 0.001). This is similar to the results of Logothetis and Schall (1990). The reduction of the OFR was also in agreement with the findings that motion sensitivity examined with subjective motion perception was lower during BR suppression (Fox & Check, 1968; Li et al., 2005; Nguyen et al., 2001, 2003). 

Figures 6, 7, and 8 clearly show that the effect of BR suppression on the temporalward OFR was greater than that on the nasalward OFR. This asymmetrical effect of BR suppression on the OFR may be explained through the physiological properties of the cells in visual pathways that are involved in both OFR and BR processing. Since it has been shown that the OFR is the initial part of the OKN slow phase (Miles et al., 2004), we think the present OFR study may be related to optokinetic nystagmus (OKN) studies. 

Many OKN studies on animals (Distler & Hoffmann, 1992; Hoffmann, Distler, & Erickson, 1991; Hoffmann, Distler, & Markner, 1996; Ohmi, Howard, & Eveleigh, 1986; Schiff, Cohen, Büttner-Ennever, & Matsuo, 1990; Yang & Sun, 1998) and strabismic subjects (Valmaggia et al., 2003; Wong et al., 2003) generally agree that asymmetrical OKN is caused by a lack of functional input from the visual cortex to the nucleus of the optic tract (NOT) in the pretectum and the related dorsal terminal nucleus (DTN) of the accessory optic system. The NOT-DTN cells are the main link between sensory input from the retina and the motor output of OKN through the brainstem. The direct pathway from the retina through the contralateral NOT-DTN is sufficient to mediate OKN when patterns move temporward. The indirect pathway from the retina through the ipsilateral LGN and striate cortex to the ipsilateral NOT-DTN is necessary for consistent OKN when patterns move nasalward (Harris, Lewis, & Maurer, 1993; Hoffmann et al., 1991, 1996; Montarolo, Precht, & Strata, 1981; Strong, Malach, Lee, & Van Sluyters, 1984; Zee, Tusa, Herdman, Butler, & Gücer, 1987). It is possible that when the temporalward preferred motion detectors in cortical/indirect pathway were engaged in BR, the nasalward preferred motion detectors in the subcortical/direct pathway would bias the nasalward responses. It is interesting that the asymmetrical OFR, not only occurred in the suppressed conditions but also appeared in some of the dominant conditions in the present study. This phenomenon is consistent with findings in OKN studies in patients with strabismus and amblyopia (Lewis, Maurer, Chung, Holmes-Shannon, & Van Schaik, 2000; Steeves et al., 1999; Valmaggia et al., 2003; Wright, 1996; Yang & Sun, 1998); the asymmetrical OKN was not only shown in the suppressed strabismic eye, but also in the non-suppressed eye. Additionally, incomplete dominance of the BR may play a role in producing an asymmetrical OFR under the dominant condition. We believe that asymmetry of OFRs under the incomplete BR conditions may be a reflection of imbalance binocular inputs and processing in the visual system similar to asymmetrical OKN in strabismic subjects. 

Another important observation in the current study is the elongation of the OFR latency. It has been known that the eye movement latency became longer under low luminance (low contrast) conditions than under high luminance conditions (Miles & Kawano, 1987). Logothetis and Schall (1990) reported prolonged latencies of OKN under binocular motion rivalry conditions although they looked at the responses driven by the dominant eye. However, this is the first demonstration that the OFR latency was longer under suppressed conditions (194.8 ± 12.8 ms). Based on the data obtained from lesions and single unit recordings in monkeys, the medial superior temporal (MST) area and the middle temporal (MT) area in which many neurons provide major input to MST mediate (at least in part) the OFR (Takemura et al., 2006, 2007). Research on BR using single-unit recordings in awake monkeys and fMRI in human subjects showed that neural activity related to binocular rivalry occurred at several stages of the visual pathway including lateral geniculate nucleus (LGN), V1, V3A, LOC, and V5/MT (Blake & Logothetis, 2002; Leopold & Logothetis, 1996; Logothetis & Schall, 1989; Moutoussis et al., 2005; Wunderlich et al., 2005). It is also interesting that only the late part of an open-loop vergence was influenced by a stereoscopic perception from slant (Sheliga & Miles, 2003), while the whole OFRs were delayed in our experiments. The elongation of the OFR latency may suggest that interaction of BR and OFR may occur at multiple stages including an early visual processing stage, such as LGN. However, the MST and the MT may play important roles since the MST and MT are involved in both BR and OFR processing. Additionally, the fact that the elongation of the OFR latency occurred in deeper suppression may indicate that the latency of the OFR could be an objective indicator of BR depth. 

Conclusion

In conclusion, we found that initial OFRs were significantly reduced at suppressed phases of binocular rivalry compared with those from dominant phases or monocular conditions. Asymmetrical OFRs biasing to nasalward motion under BR was unexpectedly observed. We suggest that asymmetry of OFRs under the incomplete BR conditions may be a reflection of unbalanced binocular inputs and processing in the visual system similar to asymmetrical OKN in strabismic subjects. The latency of OFR was prolonged under deeper suppression conditions, suggesting that interaction of BR and OFR may occur at multiple processing stages including an early stage of the visual processing. The approaches may have the potential for objective measurement of BR suppression in clinical evaluation of binocular function. 

Appendix A

The following formula was used for Gamma distribution function:  

$f\left(x\right)={\lambda }^{\gamma }/\Gamma \left(\gamma \right)x^{\gamma -1}{e}^{-\lambda x},$

(A1)

where, Γ( γ) = ( γ − 1)!, γ and λ (gamma and lambda) are the parameters in the Gamma distribution function. 

Acknowledgments

We thank Drs. John McClurkin and Arthur Hays, Boris Sheliga and Edmond FitzGibbon, and Frederick Miles from LSR, NEI, NIH for their helping with setting up the search coil system. We thank Mr. Matthew Kaufman for his contribution in preparing the manuscript. This project is partly supported by NEI EY015797 Grant and The Research to Prevent Blindness. 

This study was presented in part at the 2006 ARVO annual meeting (Zhu, Yang, & Hertle, 2006). 

Commercial relationships: none. 

Corresponding author: Dongsheng Yang. 

Email: [e-mail]. 

Address: 3705 5th Avenue, Children's Hospital of Pittsburgh, Room 2953, Pittsburgh, PA 15213, USA.