HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental Videos All Versions of this Article: 99/5/2329    most recent 01316.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Hayashi, R. Articles by Kawano, K. Search for Related Content PubMed PubMed Citation Articles by Hayashi, R. Articles by Kawano, K.

Eye Movements in Response to Dichoptic Motion: Evidence for a Parallel-Hierarchical Structure of Visual Motion Processing in Primates

Department of Integrative Brain Science, Graduate School of Medicine, Kyoto University, Kyoto, Japan

Submitted 4 December 2007; accepted in final form 9 February 2008

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Brief movements of a large-field visual stimulus elicit short-latency tracking eye movements termed "ocular following responses" (OFRs). To address the question of whether OFRs can be elicited by purely binocular motion signals in the absence of monocular motion cues, we measured OFRs from monkeys using dichoptic motion stimuli, the monocular inputs of which were flickering gratings in spatiotemporal quadrature, and compared them with OFRs to standard motion stimuli including monocular motion cues. Dichoptic motion did elicit OFRs, although with longer latencies and smaller amplitudes. In contrast to these findings, we observed that other types of motion stimuli categorized as non-first-order motion, which is undetectable by detectors for standard luminance-defined (first-order) motion, did not elicit OFRs, although they did evoke the sensation of motion. These results indicate that OFRs can be driven solely by cortical visual motion processing after binocular integration, which is distinct from the process incorporating non-first-order motion for elaborated motion perception. To explore the nature of dichoptic motion processing in terms of interaction with monocular motion processing, we further recorded OFRs from both humans and monkeys using our novel motion stimuli, the monocular and dichoptic motion signals of which move in opposite directions with a variable motion intensity ratio. We found that monocular and dichoptic motion signals are processed in parallel to elicit OFRs, rather than suppressing each other in a winner-take-all fashion, and the results were consistent across the species.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Ocular following responses (OFRs) elicited by moving visual patterns with ultrashort latencies are open-loop ocular motor responses that have been used to investigate visual motion processing in primates (Gellman et al. 1990; Miles et al. 1986). Because the two eyes of primates are frontally positioned, the integration of images from the two retinas may play a role in visual motion processing and driving OFRs. It has been shown that the extent to which OFRs occur is significantly modulated by the binocular disparity of moving patterns (Busettini et al. 1996; Masson et al. 2001). This, however, does not necessarily indicate that the signal encoding motion after binocular integration can elicit OFRs alone because the stimuli used in the studies cited contained monocular motion cues. In this study, we used dichoptic motion stimuli, which evoke the perception of continuous motion through the fusion of the images in each eye, even though neither image is moving (Shadlen and Carney 1986), and examined whether the purely binocular motion signal generated by these stimuli can elicit OFRs. This provides a conclusive test for whether OFRs can be derived solely from the cortical visual motion pathway because input signals from the two eyes are first integrated at primary visual (V1) cortical neurons (Hubel and Wiesel 1968).

There is a wide range of evidence to suggest that visual motion is mediated by at least two independent mechanisms; a first-order motion system is the standard motion-detection mechanism that extracts motion from moving luminance modulations, whereas another mechanism extracts another type of motion that is cued not by luminance but by some other features, such as contrast, spatial or temporal structures, or disparity, which is usually referred to as a second-order or higher-order motion system (Cavanagh and Mather 1989; Chubb and Sperling 1988; Lu and Sperling 2001a). In the following, we will mainly use the term "non-first-order motion" to refer to the latter system because there is a theory claiming that motion processing consists of more than two systems [disparity-defined motion is categorized as a third-order motion (Lu and Sperling 1995)], whereas we are interested only in the distinction between the first-order motion system and the other motion systems. In classical studies, the first-order motion system had been considered as an exclusively monocular process because motion of random-dot patterns was not seen when successive images were presented alternately to the left and right eyes (Braddick 1974). Although the demonstration of a dichoptic motion stimulus (Shadlen and Carney 1986) was first reported as evidence for a dichoptic mechanism of first-order motion, there were criticisms claiming that dichoptic motion stimuli may be detected by a non-first-order motion system, such as a feature-tracking mechanism, which does not require the first-order motion detection after binocular integration (Georgeson and Shackleton 1989; Lu and Sperling 1995; but see Carney 1997 for the counterarguments to such criticisms). Therefore to gain a better understanding of dichoptic motion processing, it is very important to compare the effects of dichoptic motion on ocular motor responses with the effects of other motion modalities. Concerning the possible contribution of non-first-order motion to OFRs, there is one preliminary report that contrast-defined motion elicits OFRs to the same extent as first-order motion (Benson and Guo 1999). In this report, however, the standard method for measuring OFRs was not adopted and the results are inconsistent with other recent reports indicating that non-first-order motion signals are relatively ineffective in eliciting OFRs (Masson et al. 2002; Sheliga et al. 2005) and other types of slow tracking eye movements (Harris and Smith 1992; Hawken and Gegenfurtner 2001). In an attempt to resolve these contradictions, we measured the detailed spatiotemporal frequency tuning of OFRs induced in monkeys by three types of motion stimuli: 1) ordinary first-order motion including monocular motion components, 2) dichoptic motion, and 3) non-first-order motion (contrast-defined, flicker-defined, or disparity-defined motion). The results showed that first-order motion and dichoptic motion elicited OFRs, whereas non-first-order motion did not, indicating the involvement of distinct systems of cortical motion processing in driving OFRs.

Subsequently, following our finding that both monocular first-order motion and dichoptic motion signals are capable of driving OFRs, we further investigated how these two motion signals are integrated to affect the OFRs. For an ordinary motion stimulus, however, the direction of motion at the level of monocular processing always corresponds to the direction of motion generated after binocular integration, and thus it has not been possible to manipulate the two motion signals separately. We solved this problem by using novel motion stimuli recently developed by one of the authors (Hayashi et al. 2007), in which the monocular and dichoptic motion components moved in opposite directions with a variable motion intensity ratio, for OFR recording. The results indicate that monocular and dichoptic motion signals affect the OFRs, depending on the intensity of each motion, without suppressing each other in a winner-take-all fashion.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
The present study consists of two major experiments. In the first experiment, we used three monkeys to detail the spatiotemporal frequency tuning of the OFRs induced by three types of motion stimuli. In the second experiment, we measured the OFRs in both humans and monkeys using our novel dichoptic motion stimuli (Hayashi et al. 2007).

Animal preparation

In experiments 1 and 2, behavioral data were collected from three adolescent rhesus monkeys (Macaca mulatta), weighing 4–7 kg. Many of the methods and procedures used here were similar to those used in previous studies (Miles et al. 1986; Miura et al. 2006). All animals had been previously trained to fixate on a small target spot on a cathode-ray tube (CRT) monitor for a liquid reward. Under sodium pentobarbital anesthesia and aseptic conditions, each monkey was fitted with a pedestal, which was secured to the skull through implanted bolts and dental acrylic, to allow the head to be fixed in a standard stereotaxic position. Scleral search coils were also implanted around both eyes for the purpose of monitoring eye movements using an electromagnetic induction technique (Fuchs and Robinson 1966; Judge et al. 1980). All experimental protocols were approved by the Kyoto University Animal Care and Use Committee.

Data collection and analysis

The experiments were controlled by two PCs, which communicated via an Ethernet using a TCP/IP protocol. One of the PCs, which was running a Real-time EXperimentation software package (REX) developed by Hays et al. (1982), provided the overall on-line control of the experimental protocol as well as acquiring, displaying, and storing the eye movement data. The other PC, which was running Matlab (The MathWorks) subroutines using the Psychophysics Toolbox (Brainard 1997), generated the visual stimuli after receiving a start signal from the REX machine. The AC voltage signals induced in the search coils were converted to separate DC voltage outputs proportional to the horizontal and vertical positions of each eye. The coil output voltages were digitized at 1 kHz with a resolution of 12 bits and calibrated in relation to the eye position by having the animal fixate on small target points located at known eccentricities along the horizontal and vertical meridia. Data were also low-pass filtered (30-Hz cutoff frequency) during off-line processing.

The eye position measurements were differentiated in off-line analyses to obtain eye velocity values. Trials with saccadic intrusions during the experiment were omitted. Upward deflections of the stimuli or eyes were designated as positive values. Eye velocity temporal profiles in response to a particular stimulus and synchronized to stimulus onset were averaged. We pooled data from the left and right eyes. The responses to upward and downward motions were also pooled to improve the signal-to-noise ratio by subtracting the mean response to each downward motion stimulus from the mean response to the corresponding upward motion stimulus. The subtraction is also valid to detrend the eye movement, which is irrelevant to motion stimuli, and to examine the effect of motion modality and spatiotemporal frequency regardless of motion direction. The amplitudes of the vertical OFRs were quantified by measuring the changes in vertical eye position during 100-ms time periods commencing 60 ms after the onset of the motion stimuli. The onset latency of the OFRs was distributed around 60 to 100 ms so that these response measures were usually restricted to the period prior to the closure of the visual feedback loop (i.e., twice the reaction time); these values represent the initial open-loop responses. We used the same fixed starting point for all motion conditions for amplitude calculation instead of using different starting points, depending on the onset latency of each motion modality, because for non-first-order motion conditions onset latency was not definable due to no response. The selection of the time window would change the absolute value of the amplitudes of OFRs, whereas it would not critically affect the frequency tuning of the OFRs in the later analysis. To estimate the onset latency of the OFRs elicited by each motion stimulus, we first detected the time at which the velocity profile exceeds threefold the maximum value of the velocity profile during the baseline period (from –130 to 30 ms based on the stimulus onset). We then calculated the zero crossing point of a regression line fitted to the data around the detected time period (±15 ms) as the onset latency (also see Lindner and Ilg 2000).

To draw a spatiotemporal tuning map of the OFR amplitude, we fitted the data with the following two-dimensional, log-scaled Gaussian function as described in the following equation

(1)

where x is the spatial frequency, y is the temporal frequency, A is a scale factor, _x and _y are the window sizes of the Gaussian function, and x_c and y_c are the center of the Gaussian function. is the rotation angle of the Gaussian window in the spatiotemporal frequency space. To draw the spatiotemporal tuning map of the OFR onset latency, we fitted the data with the function A – f(x, y) because the tuning map of the onset latency is concave.

Stereoscopic display system for the monkeys

Two CRT monitors were covered with orthogonal polarizing filters (Edmund Techspec linear polarizing laminated film) and the screen images on the monitors were superimposed with a half mirror. Two orthogonal polarizing filters were also placed in front of the monkey's eye, which dissociated the two monitor images for the dichoptic stimulus presentation. Left-eye images were generated from a PC through the green channel with an 8-bit color resolution, whereas right-eye images were generated through the blue channel. The green signal was then bifurcated into three lines and transmitted to one CRT monitor as RGB signals, whereas the bifurcated blue signal was sent to the other monitor. H- and V-sync signals were also bifurcated into two lines and transmitted to the two CRT monitors. Gamma correction was performed for each monitor and the luminance ranges of the two monitors were matched at 0.25–6.25 cd/m². Two crossed polarizing filters with an average transmission of 0.04% were used; the maximum luminance that passed through the crossed filters from the screen image for the other eye was 0.018 cd/m², which was much darker than the darkest screen dot that passed through the two parallel filters. Human observers with one eye closed did not perceive any coherent motion sensation during the presentation of the dichoptic motion stimulus. We also measured the eye movement responses of three monkeys during the presentation of the dichoptic motion stimulus while occluding the left eye with a black board and observed no OFRs. Therefore cross talk from the two images was very unlikely to affect the results and analysis.

Visual stimuli for experiment 1

In experiment 1, we used three different types of motion stimuli: 1) first-order motion, 2) dichoptic motion, and 3) non-first-order motions with features defined by contrast, flicker, or disparity. The first-order motion stimuli were horizontally oriented sinusoidal gratings with contrasts of 10% (Michelson contrast). We used the term "first-order motion" instead of "(first-order) monocular motion" to refer to these stimuli because, although they contained monocular motion signals, they were observed with the two eyes, thus not presented monocularly.

Dichoptic motion stimuli were generated by decomposing a moving sinusoidal grating into two monocular standing waves for which the phases differed by 90° in both the space and time domains as described in the following equation

(2)

where L(x, t) and R(x, t) are the luminance profiles of inputs from the left and right eyes at position x and time t, respectively. _x, _t, C, and m are spatial frequency, temporal frequency, mean luminance, and amplitude/contrast, respectively. For this stimulus, each monocular component did not contain any directional motion information, whereas a smoothly drifting grating was perceived under the dichoptic observation of the two patterns (see Fig. 1 A for its luminance profile plot over space and time). The monocular patterns of the dichoptic motion stimuli were similar horizontally oriented sinusoidal gratings with contrasts of 10%.

View larger version (63K): [in this window] [in a new window]   FIG. 1. A: luminance profiles of the monocular inputs of a dichoptic motion stimulus and their summation. Each plot depicts luminance profiles over space (the horizontal axis) and time (the vertical axis), indicated by gray scale (dark for low luminance; bright for high luiminance). The left and right eye images are counterphase-flickering sinusoidal gratings (shown as grid patterns in space–time plot) whose phases differ by 90° in both space and time domains. The summation of the 2 monocular images is a moving grating (shown as oblique stripes). B: luminance profiles of the monocular inputs of a monocular/dichoptic (M/D) opposite stimulus and their summation (only one spatiotemporal frequency component is depicted here for simplification, whereas the actual stimuli we used in experiment 2 were broadband). Each monocular image consists of a continuously moving sinusoidal grating and a counterphase-flickering grating (a noise component) whose intensity is scaled by parameter a (described in Eq. 5). Consequently, the left and right images have motion energy moving in one direction (in this figure from right to left). However, by making a noise component in one eye cancel a moving component in the other eye, we can create dichoptic motion whose direction is opposite to the direction of the motion of each eye input. Note that the image contrast in A and B is emphasized to clearly show dark and bright patterns, although it is supposed to be modulated sinusoidally in both space and time.

Behavioral paradigms for experiment 1

During the recording sessions, a monkey sat in a primate chair with its head secured in place and facing the CRT monitor(s) at a distance of 80 cm. The field size of the stimulus was 18 x 18° and the screen refresh rate was 60 Hz. The size of each pixel was 0.037°. The monitor resolution was 640 x 480 pixels. At the beginning of each trial, a fixation point was presented at the center of the screen with a random-dot pattern as the background. The animal was required to gaze at the fixation point for 300 ms + a randomly fluctuating interval (0 to 300 ms). Eye movements were restricted within a ±1° window during the fixation period. The fixation point was then removed and one of eight randomly selected motion stimuli moved upward or downward. After the motion stimulus was presented for 333 ms (20 frames), the screen went blank (neutral gray). The animals were supposed to passively view the presented motion stimuli. If saccadic eye movement was detected during the fixation period, the screen was turned off and the trial was repeated one more time from the beginning. If no saccades were detected during the stimulus presentation period, the data were stored on a hard disk and the animal was given a drop of fruit juice; otherwise, the trial was aborted and fluid was withheld. Also note that the fluid was provided solely as an incentive for the monkey to fixate on the screens and was not contingent on the animal's tracking responses. The mean luminance of the screen was kept constant at 3.25 cd/m² throughout the experiment.

In an additional experiment, we trained our monkeys to report the perceived direction of the motion stimuli by making a saccade after the stimulus presentation toward one of two white points (identical to the fixation point) simultaneously presented at vertically deviated locations of 5° eccentricity (i.e., up or down). To receive a reward, the monkeys were required to make a saccade toward the point in the upper or lower visual field within 300 ms after the stimulus offset and remain fixated for 300 ms within an eye window of ±2° if the presented motion stimulus moved upward or downward, respectively. The correct target location for the saccade was presented after the trial, regardless of whether the monkeys succeeded in the task, by reversing the contrast polarity of the target from white to black for 500 ms, whereas the other point presented at the opposite location as a distracter was removed. The reward was withheld and the same stimulus presentation was repeated if the animal failed to make a saccade toward the target point within the time limit, including when they made a saccade toward the opposite point. We analyzed the eye movement response during the first trial of each presentation and discarded responses obtained during the retrials to calculate the probability of a correct response. In this behavioral experiment, we used a contrast-defined non-first-order motion stimulus for which the spatial frequency was 0.25 cpd and the velocity was 4°/s; this stimulus was also used in the ocular motor response recording experiment.

Visual stimuli and the behavioral paradigm for experiment 2

In experiment 1, we used the simplest configuration of dichoptic motion stimuli, as originally proposed by Shadlen and Carney (1993), consisting of a sinusoidal grating pattern flickering at a single temporal frequency. This narrowband frequency property of Shadlen and Carney's dichoptic motion display, however, makes it difficult to choose a single representative stimulus to compare the monocular and dichoptic motion systems, which have different frequency tuning characteristics. In addition, because two monocular inputs were temporally modulated with a sinusoidal function in a counterphase manner, interocular suppression and disturbed binocular integration should occur, which could potentially reduce the impact of the dichoptic motion signals on the OFRs.

To solve these problems and explore the gross effects of monocular and dichoptic motion on OFRs in experiment 2, we introduced three types of broadband motion stimuli with motion signals that contained all of the spatiotemporal frequency components: 1) dichoptic motion stimuli with monocular components that were dynamic random-dot patterns without any coherent motion, 2) monocular motion stimuli without interocular pattern correlation, and 3) motion stimuli in which the monocular and dichoptic motion components moved in opposite directions (hereafter, we will refer to the third type of stimuli as M/D opposite stimuli). The broadband nature of these three types of motion stimuli allowed us to investigate the monocular and dichoptic motion detection systems despite the differences in their frequency properties. Furthermore, the monocular components of these motion stimuli were characterized by constant luminance and contrast throughout the presentation, thereby avoiding unwanted artifacts due to interocularly unbalanced inputs.

The principle to generate broadband dichoptic motion (the first stimulus) is the same as that of the original dichoptic motion proposed by Shadlen and Carney (1993): broadband dichoptic motion is produced by having counterphase modulation in all combination of spatial and temporal frequencies, which are in spatiotemporal quadrature between the two eyes, which is given in the following equation (also see Eq. 2 in Hayashi et al. 2007 for details)

(3)

where I(x, t) is the luminance profile of an arbitrary spatiotemporal random-dot pattern. H_x/t denotes the Hilbert transformation in the dimension of x or t, which is the operator that shifts all of the frequency components by 90°. Movie S1 is an example of such broadband dichoptic motion stimulus.¹

The interocularly uncorrelated monocular motion stimuli (the second stimuli) used can be described as follows

(4)

where I(x, t) and I2(x, t) are mutually independent random noise generated by different seeds. The monocular inputs of this stimulus are the so-called quadrature motion (Carney and Shadlen 1993), which is generated by adding dynamic random noise and its spatiotemporal quadrature pair, thus mathematically equivalent as binocular summation of dichoptic motion. Because the random seed used to generate the input pattern to one eye was different from the seed used to generate the pattern for the other eye, there was no correlation between the inputs from the two eyes and no moving component emerged after interocular integration. We used the term monocular motion to mean that no constant motion emerged from integrating the image components delivered to one eye with those delivered to the other eye. Movie S2 is an example of such monocular motion.

The method for generating the M/D opposite motion stimuli (the third stimuli), in which monocular and binocular motions moved in opposite directions, is given in the following equation (or see Eq. 3 from Hayashi et al. 2007 for details)

(5)

As shown in Fig. 1B, the monocular inputs of those motion stimuli consist of moving components and counterphase-flickering noise components (note that the actual stimuli we used in experiment 2 were broadband, whereas Fig. 1B depicts only one spatial and temporal frequency modulation for simplification). Therefore the monocular inputs have motion energy in the direction of moving components. However, a noise component in one eye is arranged to cancel a moving component in the other eye to produce motion energy toward the direction opposite to the monocular motion after binocular integration. Movie S3 is an example of such M/D opposite motion stimuli.

An interesting aspect of the M/D opposite motion stimuli is that we are able to manipulate the intensity of noise components in the monocular inputs (i.e., the intensity of the monocular and dichoptic motion components) by varying a single parameter, a. We generated the motion stimuli listed in Table 1 for the first block of sessions in experiment 2. Stim1 was interocularly uncorrelated monocular motion (described in Eq. 4) and Stim8 was pure dichoptic motion (described in Eq. 3). Stim2 to Stim7 were M/D opposite motion stimuli generated based on Eq. 5 by setting parameter a at 1, 1.5, 2, 3, 5, and 10, respectively. In our experimental configuration, monocular motion intensity decreased and dichoptic motion intensity increased as parameter a increased; thus the monocular and dichoptic motion signals were inversely related. As control stimuli, we created motion stimuli with monocular motion intensities that were exactly the same as Stim1 to Stim8, whereas the dichoptic motion intensities of the stimuli were all zeroed using different random seeds to generate each eye's input. We will refer to these stimuli as Stim1' to Stim8'. Note that Stim1 and Stim1' were identical and Stim8' was simply a dynamic random-noise pattern without any consistent motion signal.

Human subjects

In addition to recording the monkeys' behavior, we also measured OFRs using human subjects in experiment 2. Three healthy male subjects whose visual acuity and stereoscopic depth detection performance were normal (or corrected to a normal level) participated in the experiments. All of these subjects were authors but they had no prejudice about the results and did not know the results of the animal experiments before participating in these experiments.

Recording eye movements of human subjects

We used an electromagnetic search-coil technique to record eye movements as was done in the animal experiments. A coil embedded in a silastin scleral ring (Skalar Medical, Delft, The Netherlands) was placed in the right eye of the subject after the application of one or two drops of oxybuprocaine. The experimental setup for the human subjects was basically the same as that used for the monkeys except for the stimulus display system. We used a customized CRT rear projector with orthogonal polarizing filters. Left-eye images were output from the display PC through the green channel with an 8-bit color resolution, whereas right-eye images were output through the blue channel. Stimuli were rear projected from the CRT projector, in which the blue gun was replaced with a green gun covered with a horizontally polarizing filter, whereas the original green gun was covered with a vertically polarizing filter. Images were then segregated using eyeglasses that covered the left and right eyes with vertically and horizontally polarizing filters, respectively. Subjects viewed the image screen (80.72 x 80.72°; 256 x 256-pixel resolution; 1 pixel = 0.38°, 60-Hz refresh rate; maximum luminance = 1.79 cd/m²) from a distance of 50 cm. Thus the screen size was wider than that in the monkey setup. Each subject's head was restricted on a chinrest with a headband and recordings were made with the room lights dimmed. Subjects were instructed to look at the center of the screen and were asked not to attentively pursue the moving patterns. All experiments were conducted in accordance with the principles embodied in the Declaration of Helsinki (code of ethics of the World Medical Association) under the approval of the ethics committee of Graduate School of Medicine, Kyoto University.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Experiment 1

We measured the OFRs elicited by standard first-order motion, which included monocular motion cues, dichoptic motion (purely binocular motion), and contrast-defined, disparity-defined, and flicker-defined non-first-order motion. In this experiment, we tested a variety of spatiotemporal frequency conditions for each motion type to attempt to reveal how the effects of the frequency properties of visual motion processing on the OFRs depended on the motion types.

Eye velocity temporal profiles for each motion type

The averaged eye velocity temporal profiles for each spatiotemporal frequency condition and each motion type are plotted in Fig. 2. Dichoptic motion (blue lines) elicited OFRs similar to the results observed with first-order motion (red lines), although the response gains were smaller and the onset latencies were longer. On the other hand, non-first-order motion (green lines) did not elicit OFRs regardless of the spatiotemporal frequency conditions and whether or not the non-first-order features were defined by contrast or disparity. These results were consistent across all of the tested monkeys.

View larger version (27K): [in this window] [in a new window]   FIG. 2. Averaged temporal eye velocity profiles for each spatiotemporal frequency recorded from 3 monkeys (A: M1; B: M2; C: M3). Red, blue, and green lines denote the eye velocity profiles in response to first-order motion, dichoptic motion, and non-first-order motion, respectively. Note that each line is not the eye velocity profile for a single trial but represents the averaged responses for each frequency condition. We have not indicated whether the non-first-order features were defined by contrast modulation or disparity modulation in the plot of the waveforms elicited by the 2 non-first-order stimuli (green lines).

ONSET LATENCY DIFFERENCE.  The onset latency of the OFRs was determined from the eye velocity profiles as described in METHODS. We were able to determine the latencies for 37 of the 40 conditions for first-order motion and 16 of the 32 conditions for dichoptic motion. In Fig. 3 A, the mean latencies from the three monkeys are plotted as functions of the temporal frequency. The results show that the onset latencies of the OFRs elicited by first-order motion (red lines) were shorter than those evoked by dichoptic motion (blue lines), with the exception of one spatial frequency condition (0.125 cpd) that resulted in very weak OFRs. The averaged onset latency across all of the frequency conditions was 74 ms for first-order motion and 94 ms for dichoptic motion.

View larger version (43K): [in this window] [in a new window]   FIG. 3. A: dependence of the mean onset latencies on the temporal frequency (first-order motion: red lines; dichoptic motion: blue lines). Symbols and solid lines denote the latency under a fixed spatial frequency condition. Frequency conditions in which the onset latency was identified for all 3 animals were chosen to calculate the mean onset latency. B: heat map plots of the spatiotemporal frequency tuning of the mean onset latency. Heat maps were created by fitting the data with the 2-dimensional (2D) log-scaled Gaussian function described in Eq. 1. The positions of the black dots show sampled points, whereas the color of the surrounding disc represents the measured mean onset latency.

OFR AMPLITUDE DIFFERENCE.  Figure 4 A depicts how the OFR amplitude changed when either first-order motion or dichoptic motion was used under the same spatiotemporal frequency condition. Amplitude values were normalized within each individual with the maximum amplitude among all the tested conditions. Note that the mean luminance and contrast of each eye's input for the first-order motion were identical with those for dichoptic motion. If the amplitudes for both motions were equal, the data should align on the black dotted line (slope = 1). The data are primarily distributed under this line and the slope of the regression line (black solid line) was 0.26 (R² = 0.4, P < e^–11), indicating that the OFR amplitude in response to dichoptic motion was lower than that elicited by first-order motion.

View larger version (15K): [in this window] [in a new window]   FIG. 4. A: a comparison of the normalized ocular following response (OFR) amplitudes induced by first-order motion (horizontal axis) and dichoptic motion (vertical axis). The OFR amplitudes were normalized within an individual using the maximum amplitude obtained with the first-order motion condition. Symbols represent the spatiotemporal frequency conditions for an individual monkey. Almost all of the points fell below the line with a slope of 1 (black dotted line), indicating that the OFR amplitude in response to dichoptic motion was lower than that elicited by first-order motion. B: dependence of the mean OFR amplitude on temporal frequency. Tuning curves for first-order motion, dichoptic motion, and non-first-order motion are depicted in red, blue, and green, respectively. Symbols connected with solid lines represent average OFR amplitudes from 3 monkeys with a fixed spatial frequency; we tested all 5 spatial frequency conditions for first-order motion, although the 0.75-cpd condition was not tested for dichoptic and non-first-order motions. C: dependence of the mean OFR amplitude on stimulus speed.

Figure 5 depicts the heat map plots of the spatiotemporal frequency tuning of each motion type. OFR amplitudes were normalized within each individual using the maximum value of OFR amplitude across all motion conditions. The plots indicate that high OFR amplitude areas extend to a higher temporal frequency domain in the maps for first-order motion, whereas the tuning maps for dichoptic motion tend to center around lower temporal frequencies. The temporal frequency at the peak of the fitted surface was 13.7 Hz for first-order motion and 10.5 Hz for dichoptic motion; the temporal-frequency peak shifted by 3.2 Hz between the two conditions. The observed lower temporal frequencies of the OFRs elicited by dichoptic motion are consistent with a previous finding that the perception of dichoptic motion is mediated by a temporally inferior system compared with ordinary first-order motion (Lu and Sperling 1995). The spatial frequency peak was at 0.39 cpd for first-order motion and at 0.65 cpd for dichoptic motion. The higher spatial frequency property of OFRs evoked by dichoptic motion may reflect the fact that binocular integration is well developed in the central visual areas. Note that data were normalized within each individual (thus the maximum OFR amplitude value of a first-order motion condition) rather than within each motion type because non-first-order motions did not elicit the OFRs and normalizing data with an almost null value distort the tuning plots. Furthermore, tuning plots for the dichoptic motion condition did not change their shape even if the plots were normalized with the maximum OFR amplitude of dichoptic motion conditions.

View larger version (37K): [in this window] [in a new window]   FIG. 5. Heat map plots of spatiotemporal frequency tuning of the OFR amplitude for each motion type (arranged in rows) recorded from 3 monkeys (arranged in columns). The horizontal axes of the maps represent spatial frequency, whereas the vertical axes represent temporal frequency. OFR amplitudes were normalized using the maximum OFR amplitude within an individual animal and colors represent the scales of the normalized OFR amplitudes (increasing from blue to red, indicated in the color bar). The positions of the black dots show the sampled points, whereas the color of the surrounding discs represents the measured OFR amplitudes. Heat maps are drawn by fitting the data with the 2D log-scaled Gaussian function described in Eq. 1. Non-first-order motion did not show any systematic modulation, whereas first-order and dichoptic motions exhibited clear tuning that was dependent on the frequency. Modulation patterns were fairly consistent across the monkeys.

EFFECT OF NON-FIRST-ORDER MOTION ON OFRS.  As shown in Fig. 2, we did not observe any significant eye movement responses with short latencies during the presentation of contrast-defined and disparity-defined motions. The heat map plots depicted in Fig. 5 (bottom three rows) further demonstrate that no systematic spatiotemporal tuning for the non-first-order motion conditions was observed regardless of the non-first-order feature used (contrast-defined, disparity-defined, and flicker-defined motion).

IS NON-FIRST-ORDER MOTION INVISIBLE TO MONKEYS?  It is possible that we did not observe the short-latency OFRs in response to non-first-order motion stimuli because the monkeys were unable to see non-first-order motion stimuli. In an additional experiment, we trained our monkeys to report the perceived direction of a motion stimulus by making a saccade toward the direction of the perceived motion. The results showed that the monkeys were able to accurately determine the direction of contrast-defined non-first-order motion. Indeed, we observed a >80% correct response rate; the performances of the three monkeys were, respectively, 82.8, 85.7, and 82.1%. The results are consistent with a previous study, which reported that monkeys correctly identified the direction of non-first-order motion defined by flicker modulation and were able to eye-track non-first-order moving targets (Ilg and Churan 2004). Consequently, we concluded that the lack of OFRs in response to non-first-order motion was not because non-first-order motion was invisible to the monkeys.

SHORT SUMMARY.  In summary, experiment 1 showed that non-first-order motion did not elicit OFRs regardless of whether the non-first-order stimulus was defined by contrast, flicker, or disparity. This finding is intriguing because non-first-order motion can readily evoke motion sensation in humans and monkeys, which can accurately determine the direction of non-first-order motion. Dichoptic motion, on the other hand, elicited OFRs similar to first-order motion. We also found, however, that the OFRs elicited by dichoptic motion showed a reduction in the response amplitude and an increase in the response latency compared with those elicited by first-order motion. Considering that first-order motion contains monocular motion cues, whereas dichoptic motion lacks such cues, the difference in the OFRs shown here can be attributed to a difference between monocular motion processing and dichoptic motion processing. In the next set of experiments, we further investigated the effects of monocular and dichoptic motion signals on OFRs using a new type of motion stimuli that we have recently developed.

Experiment 2

To investigate the interactions between the monocular and dichoptic motion detection systems independent of the difference in their frequency tuning, we measured OFRs in monkeys and humans during the presentation of three types of motion stimuli that consisted of dynamic random-dot textures including broadband frequency components: 1) a broadband dichoptic motion stimulus, 2) an interocularly uncorrelated monocular motion stimulus, and 3) M/D opposite motion stimuli in which the monocular and dichoptic motion components moved in opposite directions. In the following analysis, we will first focus on the OFRs elicited by the dichoptic motion stimulus and the responses elicited by the monocular motion stimulus to examine the differences between the two mechanisms for motion processing. Then, we will investigate the time course of the OFRs elicited during the presentation of the M/D opposite motion stimuli. Finally, we will explore the way in which monocular and dichoptic motion signals interact by decomposing the effects of the two motion components on the OFRs.

Figure 6, A–F compares the OFR waveforms elicited by interocularly uncorrelated monocular motion (continuous lines) and the waveforms elicited by dichoptic motion (dotted lines) in three human subjects (A, B, and C) and three monkeys (D, E, and F). The results revealed that OFRs were elicited more quickly by monocular motion than by dichoptic motion. Furthermore, the data from the three monkeys (Fig. 6, D–F) showed that the eye velocity values of the OFRs induced by dichoptic motion reached the level of the eye velocity values evoked by monocular motion about 150 ms after stimulus onset, indicating that the increased onset latency for dichoptic motion was not caused by a weak dichoptic motion signal that did not evoke vigorous OFRs. The direct comparison between the latencies associated with monocular motion and dichoptic motion (Fig. 6G) clearly shows that the latency in response to dichoptic motion was greater than that in response to monocular motion. Mean onset latency was 91 ± 23 ms for monocular motion (111 ms for the human subjects and 71 ms for monkeys) and 123 ± 43 ms (158 ms for the human subjects and 89 ms for monkeys). Note that the mean luminance and contrast of each eye's input for the monocular motion stimulus were identical to those for the dichoptic motion stimulus. It is also noticeable that the long latency time for dichoptic motion observed in human subjects is similar to the long latency time for vergence eye movements elicited by stepwise shift of dichoptically presented random-dot patterns (Van Der Steen and Kanai 2002).

View larger version (13K): [in this window] [in a new window]   FIG. 6. A–F: averaged eye velocity temporal profiles elicited by interocularly uncorrelated monocular motion (continuous lines) and dichoptic motion (dotted lines). The figures in the top row are the OFR waveforms from 3 human subjects (A, B, and C for S1, S2, and S3, respectively) and the figures in the middle row are the OFR waveforms from 3 monkeys (D, E, and F for M1, M2, and M3, respectively). The onset latency of the OFR waveforms evoked by monocular motion was shorter than the latency of those elicited by dichoptic motion. Open circles indicate estimated stimulus onset time points. G: a comparison of the onset latencies for monocular and dichoptic motions. Each symbol represents data for an individual human subject (open circles) or monkey (crosses). All of the points fell above the line with a slope of 1 (black dotted line), indicating that the onset latency of the OFR waveform elicited by dichoptic motion was consistently longer than the latency of the OFR waveform elicited by monocular motion.

View larger version (24K): [in this window] [in a new window]   FIG. 7. Differences in the mean eye velocities for the monocular and dichoptic motion conditions. Waveforms were constructed by subtracting the OFR waveforms for dichoptic motion from those for monocular motion. A, B, and C represent data from human subjects (S1, S2, and S3, respectively), whereas D, E, and F represent data from monkeys (M1, M2, and M3, respectively). The results show that eye velocity was initially higher for monocular motion than for dichoptic motion in each of the subjects and monkeys. The open circles indicate the peak at which the proceeding response elicited by monocular motion reached a maximum compared with the response to dichoptic motion.

To investigate how the interaction between monocular and dichoptic motion signals affects the OFRs, we analyzed the time course of the eye movements elicited by M/D opposite motion stimuli, in which the monocular and dichoptic motion signals moved in opposite directions. Figure 8, A–F depicts the OFR waveforms elicited by monocular motion, dichoptic motion, and M/D opposite motion stimuli recorded from three human subjects (A, B, and C) and three monkeys (D, E, and F). The results show that the OFR waveforms gradually changed their forms as a function of motion intensity. Furthermore, under some conditions (marked by the open circles in the figure), the OFRs elicited by M/D opposite motion stimuli initially followed the direction of the monocular motion components (positive eye velocity), whereas they later followed the direction of the dichoptic motion components (negative eye velocity). This finding is confirmed by the fact that eye acceleration waveforms became biphasic (first positive peak, consistent with the direction of monocular motion, followed by negative peak, consistent with the direction of dichoptic motion) for some conditions. Eye acceleration plots, however, are not shown here because the initial monocular effect is too small in some eye velocity data to derive reliable acceleration data.

View larger version (19K): [in this window] [in a new window]   FIG. 8. A–F: OFR waveforms elicited by monocular motion (Stim1), dichoptic motion (Stim8), and M/D opposite motion stimuli (Stim2 to Stim7). A, B, and C represent data from human subjects (S1, S2, and S3, respectively), whereas D, E, and F represent data from monkeys (M1, M2, and M3, respectively). Line colors from red to blue represent the types of stimuli (Stim1 to Stim8). Dichoptic motion intensity increases and monocular motion intensity decreases as the color changes from red to blue. We plot the OFRs as positive if the eyes followed the direction of the monocular motion, whereas the negative values are associated with the direction of dichoptic motion. Note that the monocular and dichoptic motion components always moved in opposite directions for the M/D opposite motion stimuli. Open circles indicate a condition in which the OFRs initially followed the monocular motion signals and later moved to follow the direction of the dichoptic motion signals. G: a comparison of normalized mean eye velocities at the time of the maximum monocular motion effect and at 250 ms. Symbols represent the data for a given stimulus condition. Eye velocity was normalized within an individual using the velocity observed at 250 ms and was then averaged across all 3 human subjects and 3 monkeys. Positive eye velocity indicates that the eyes followed the direction of the monocular motion signal, whereas negative eye velocity indicates that the eyes moved in the opposite direction. Data from the Stim3 and Stim4 conditions are located in the 4th quadrant; i.e., eye velocity was positive around 100 ms and negative at 250 ms for the stimulus conditions.

Interaction between monocular and dichoptic motion signals: parallel or winner-take-all?

Next, we explored how the interaction between monocular and dichoptic motion signals changed their motion intensity-mediated effects on the OFRs. As we described in METHODS, we generated interocularly uncorrelated motion stimuli that were associated with identical monocular motion intensities as those used for the tested M/D opposite motion stimuli (monocular motion control stimuli: Stim1' to Stim8'). Therefore in addition to measuring the OFRs induced by M/D opposite motion stimuli (Stim1 to Stim8, depicted in Fig. 9 A), we also measured the OFRs elicited by these monocular motion control stimuli (Stim1' to Stim8') and estimated the effect of the monocular motion signals on the OFRs by calculating the size of the eye shifts that occurred during a 50-ms time window after the onset latency induced by monocular motion (Fig. 9B). We then subtracted the OFR waveforms elicited by monocular motion control stimuli from the waveforms elicited by M/D opposite motion (Fig. 9C). Because the residual responses can be regarded as the response modulation due to dichoptic motion signals, we estimated the effect of the dichoptic motion signals on the OFRs as the amplitude of the residual waveforms. For this measurement, we calculated the size of the eye shifts during a 50-ms time window after the onset latency in response to purely dichoptic motion.

View larger version (22K): [in this window] [in a new window]   FIG. 9. Method for decomposing the effects of monocular and dichoptic motion signals in the responses to M/D opposite motion and results. A: the OFR waveforms elicited by M/D opposite motion (Stim2 to Stim8) in human subject S1. B: the waveforms elicited by interocularly uncorrelated monocular motion control stimuli (Stim2' to Stim8') in human subject S1. C: the residual waveforms after the subtraction of B from A. Plots for Stim1 and Stim1' are removed from these figures because the conditions are identical and subtraction of one from the other is not informative. D: mean normalized amplitude of the OFRs (all 3 human subjects and 3 monkeys) elicited by monocular motion control stimuli (red continuous lines) and mean normalized amplitude of the OFRs elicited by dichoptic motion components of M/D opposite motion stimuli (blue dotted lines) are plotted as a function of motion intensity (dichoptic motion intensity – monocular motion intensity). Error bars indicate the SEs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
In the present study, we tested whether OFRs can be elicited by purely binocular motion signals (dichoptic motion signals) in the absence of any consistent monocular motion signals and examined the direct contribution of binocular cortical processing to ocular motor responses. We also revisited an issue about whether pure non-first-order motion can elicit OFRs. Furthermore, we used a novel motion display, in which monocular and dichoptic motion components move in opposite directions with a variable intensity ratio, to dissociate the effects of monocular and dichoptic motion signals on eye movement responses. This allowed us to explore the interaction between these two types of motion signals.

We found that 1) non-first-order motion did not elicit OFRs while evoking a sensation of motion; 2) dichoptic motion elicited OFRs as did first-order motion; 3) the detection of dichoptic motion required a longer integration time compared with monocular first-order motion; and 4) monocular and dichoptic motion signals did not interact in a winner-take-all manner to elicit OFRs when they moved in opposite directions, suggesting that two signals are processed in a parallel manner.

Does non-first-order motion elicit OFRs?

Benson and Guo (1999) reported that eye movement velocity was indistinguishable to first- and non-first-order (second-order in their term) motion in a monkey OFR study. However, subsequent reports using reversed-phi motion stimuli and missing-fundamental stimuli (Masson et al. 2002; Sheliga et al. 2005) have indicated that non-first-order motion signals are less effective than first-order motion signals in eliciting OFRs. The studies of other types of eye movement responses, such as optokinetic nystagmus (OKN) (Harris and Smith 1992, 2000) and pursuit eye movements (Butzer et al. 1997; Hawken and Gegenfurtner 2001; Lindner and Ilg 2000), have also shown a very limited effect of non-first-order motion on the initiation of eye movements at short latency. The apparent discrepancy between the finding of Benson and Guo and the results of subsequent studies could be attributable to the fact that the former used their own unique paradigm for measuring OFRs, making direct comparison difficult. In the present study, we revisited this issue using the standard paradigm for measuring OFRs and directly measured the effect of non-first-order motion on OFRs. We observed that the OFRs were not elicited at a short latency by non-first-order motion—regardless of whether the non-first-order features were defined by contrast, flicker, or disparity modulation—supporting the results of the more recent OFR studies. It is difficult to determine why the effect of non-first-order motion of OFRs reported by Benson and Guo should differ from the results of this and other later studies since, as mentioned earlier, the experimental paradigms used were not standard. It is very unlikely, however, that the observed discrepancy is due to the size of stimulus or to other parameters, such as the spatiotemporal frequency, because these parameters were similar in both the present study and that of Benson and Guo. One possibility is that the OFRs elicited by non-first-order motion stimuli in the study of Benson and Guo were due to the intrusion of first-order motion cues because their method for generating non-first-order motion stimuli may not have eliminated all potential first-order artifacts. Benson and Guo used non-first-order motion stimuli composed of contrast-modulated "static" noise texture patterns, although it has been reported that the detection of such motion stimuli is inadvertently mediated by the first-order motion detection mechanism (Ledgeway and Hutchinson 2006; Smith and Ledgeway 1997). Furthermore, the authors did not verify that they had eliminated the first-order artifacts using a psychophysical technique (for technical details see Lu and Sperling 2001b). We observed that the slightest addition of first-order components (1.25% contrast) to a non-first-order motion stimulus resulted in significant OFRs, indicating that cancellation of first-order components is critical for measuring the effects of purely non-first-order motion stimuli on OFRs. The finding that no OFR was elicited by flicker-modulated non-first-order motion stimuli, which are thought to be free of first-order artifacts as a result of imperfect monitor calibration, is strong evidence supporting our suggestion. Moreover, our results were consistent across all three monkeys tested, whereas only one monkey was tested by Benson and Guo. The main focus of the present study, however, was not to identify factors causing the results to differ from those of Benson and Guo (1999), but to investigate dichoptic motion processing in comparison with first-order and non-first-order motion processes from the viewpoint of reflexive ocular motor control. Our results clearly show that the effect of non-first-order motion on OFRs is quite limited compared with the effects of first-order and dichoptic motions.

In real-life situations, first- and non-first-order motion cues are typically present together, and non-first-order signals, which are mediated by a slow detection mechanism, are thought to help the detection of a moving target in a noisy environment (Smith and Scott-Samuel 2001). Our results indicate that this type of slow supplementary motion system is not used for eliciting quick ocular motor reflexes; however, they leave open the possibility that non-first-order motion signals are used to generate other types of eye movement responses, such as voluntary eye tracking and the closed-loop eye movement observed during OKN.

Dichoptic motion as first-order motion

It was once thought that dichoptic motion stimuli are detected by a higher-level feature-tracking mechanism (thus non-first-order motion mechanism) rather than the first-order motion detection system (Georgeson and Shackleton 1989, 1992; Lu and Sperling 1995). Carney (1997), however, convincingly demonstrated that dichoptic motion can be processed by a binocular first-order system using a psychophysical technique that makes figural cues for motion perception unavailable. Our findings that dichoptic motion elicited OFRs similar to first-order motion, whereas different types of non-first-order motion did not elicit OFRs, are consistent with the hypothesis that there is a first-order motion system for dichoptic motion processing.

Disparity-defined motion as higher-order motion

We tested the effects of motion from a disparity modulation pattern on OFRs because this motion stimulus is another type of motion that is perceived only after binocular integration. Although Lu and Sperling (1995) concluded that disparity-defined stereoscopic motion is higher-order motion, which is presumably detected by a feature-tracking–based mechanism, relying on the fact that the temporal frequency limit for detecting the stereoscopic motion is far below the limit of the first-order motion system (Lu and Sperling 1995), another study claimed that disparity-defined stereoscopic motion is processed by first-order motion sensors (Patterson 1999). Our results showing that disparity-defined motion as well as other non-first-order motions did not elicit OFRs provide further support for the hypothesis that disparity-defined motion is detected by a processing system distinct from that used to process first-order motion.

Non-first-order motion is processed independently from first-order motion

We assumed that first- and non-first-order motions are initially detected separately based on recent findings made using electrophysiology (Albright 1992; Chaudhuri and Albright 1997; Mareschal and Baker 1998a,b; O'Keefe and Movshon 1998; Zhou and Baker 1993, 1994), MRI imaging (Dumoulin et al. 2003; Smith et al. 1998), clinical pathology (Plant et al. 1993; Vaina and Cowey 1996; Vaina et al. 1999), and psychophysics (Baker Jr 1999; Derrington and Badcock 1985; Ledgeway and Smith 1994; Nishida and Ashida 2000; Nishida et al. 1997; Smith and Ledgeway 1997; for a review of the separate motion systems theory also see Lu and Sperling 2001a). However, we should also note that a single mechanism can theoretically detect both first-order motion and some types of non-first-order motions (to be specific, second-order motions) (Johnston et al. 1992; Taub et al. 1997; Victor and Conte 1992) and some studies support a model in which motion is detected by a single pathway (a human fMRI study: Seiffert et al. 2003; a clinical study: Greenlee and Smith 1997; and psychophysics studies: Hock and Gilroy 2005; Taub et al. 1997). Therefore the debate about whether first- and second-order motions are processed in a single pathway or multiple pathways has not yet been resolved. Our OFR data, however, clearly showed segregation between first-order motion and any type of non-first-order motions, as reflected by reflexive ocular motor responses, supporting a dual-pathway model.

Two cortical pathways for visual motion processing revealed by the OFRs

Our finding that OFRs can be elicited by dichoptic motion signals in the absence of any monocular motion component provides compelling psychophysical evidence supporting the idea that OFRs can be cortically initiated, which has been suggested in previous physiological studies (Kawano et al. 1994; Takemura et al. 2007). Moreover, assuming that dichoptic motion is processed by a first-order system (Carney 1997), we can conclude that OFRs are elicited only by a monocular or binocular first-order motion signal, whereas higher-level motion signals that include non-first-order motion have no effect on these responses. These results are consistent with those of previous studies that used reversed-phi and missing-fundamental stimuli to show that the OFRs followed the direction of the first-order motion components rather than the direction of the overall pattern movement, the detection of which requires higher-order motion processing (Masson et al. 2002; Sheliga et al. 2005). Taking into account the fact that the detection of non-first-order motion requires relatively high spatial frequency filtering to extract carrier texture or salient features before motion detection (Lu and Sperling 2001a), whereas OFRs are elicited by motion signals that are characterized by relatively low spatial frequencies (Miles et al. 1986), we hypothesized that cortical visual motion processing is split into at least two pathways (most likely based on spatial frequency tuning). One pathway, which is selective for low spatial frequencies, is a genuinely first-order motion system that elicits reflexive ocular motor responses. On the other hand, all motion pathways including those processing non-first-order motion should be integrated at a certain level for further motion perception (Wilson and Kim 1994; Wilson et al. 1992; Zanker and Hupgens 1994), and motion signals at the level of this integration cannot be used to elicit OFRs. It should be noted that MT/MST neurons, which are thought to play crucial roles for driving OFRs (Kawano et al. 1994; Takemura et al. 2007), are also known to show limited response modulation in response to non-first-order motion and are thought to primarily be involved in the analysis of first-order motion (O'Keefe and Movshon 1998).

Neural substrates for dichoptic motion detection

Physiological studies have shown that the receptive fields of some binocular simple and complex cells in V1 and V2 have a space–time structure, encoding both stereoscopic depth and motion (Anzai et al. 2001). Such binocular neurons, which have response profiles that can be described with a motion-stereo-hybrid energy model (Qian and Andersen 1997), are thought to mediate a variety of psychophysical effects (Hayashi et al. 2003, 2004), such as the Pulfrich-type effect (Carney et al. 1989) and depth from motion parallax. Although no electrophysiological study has explicitly measured neural responses to dichoptic motion, it seems reasonable that dichoptic motion would be detected by the same neural population that is selective for both motion direction and zero disparity because the motion energy of dichoptic motion emerges after a simple summation of the inputs to the left and right eyes. It has been reported that the gain of OFRs varies depending on binocular disparity added to an inductive moving pattern; the gain is maximum at zero disparity for humans and at a slight crossed (near) disparity for monkeys (Busettini et al. 1996; Masson et al. 2001; Takemura et al. 2000). The hypothesis that the detection of dichoptic motion is achieved by part of the stereoscopic depth processing pathway and that such a joint coding mechanism of disparity and motion can drive OFRs would account for the gain modulation of OFRs mediated by binocular disparity.

If the detection of dichoptic motion relies on a shared mechanism with disparity processing, the two processes are expected to show similar temporal properties compared with the processing of monocularly detectable stimuli. Indeed, the finding that the detection of dichoptic motion requires a longer integration time than that required for the detection of monocular motion agrees with the facts that the temporal resolution of disparity modulation (Norcia and Tyler 1984) is poorer than that of contrast modulation (Kelly et al. 1976). One may claim that the latency difference between OFRs to either monocular or dichoptic motions (37 ms for human subjects and 17 ms for monkeys) is too large if dichoptic motion is reconstructed from a simple summation of the monocular inputs. Nienborg et al. (2005) measured the temporal frequency tuning of V1 binocular neurons in response to disparity and contrast modulation and reported that the temporal cutoff frequency for disparity modulation is 2.5-fold poorer than that for contrast modulation. The low-pass filtering nature of such a disparity process can be explained, according to their argument, by modeling the computation of disparity detection as a cross-correlation processing, i.e., band-pass monocular temporal filtering followed by nonlinear rectification. They also reported that the temporal cutoff frequency was highly correlated with rise time of V1 neurons' activity; thus the temporal cutoff frequency difference between disparity and contrast modulations could end up characterized by a 30-ms latency difference of neuronal activities in V1. In this sense, a 17-ms OFR latency difference in monkeys is not too large. In the same paper, human psychophysical experiments revealed that the temporal cutoff frequency for disparity modulation in human subjects was about twofold poorer than that estimated from monkey's V1 neurons; thus the larger OFR latency delay to dichoptic motion in human subjects in our experiment would not be totally unexplainable by the low-pass filtering nature of binocular neurons. Nonetheless, we think further electrophysiological recordings from V1 or other areas' neurons during the viewing of dichoptic motion are necessary to clarify exactly what neural mechanism is responsible for the OFR latency delay in response to dichoptic motion.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
This work was supported by Japan Society for the Promotion of Science Grant in Aid of Scientific Research JSPS16GS0312.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ The online version of this article contains supplemental data.

Address for reprint requests and other correspondence: R. Hayashi, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-Cho, Sakyo-Ku, Kyoto, 606-8501, Japan (E-mail: rhayashi{at}brain.med.kyoto-u.ac.jp)

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Albright TD. Form-cue invariant motion processing in primate visual cortex. Science 255: 1141–1143, 1992.[Abstract/Free Full Text]

Anzai A, Ohzawa I, Freeman RD. Joint-encoding of motion and depth by visual cortical neurons: neural basis of the Pulfrich effect. Nat Neurosci 4: 513–518, 2001.[Web of Science][Medline]

Baker CL Jr. Central neural mechanisms for detecting second-order motion. Curr Opin Neurobiol 9: 461–466, 1999.[CrossRef][Web of Science][Medline]

Barthelemy FV, Masson GS. Spatial integration of motion for human and monkey ocular following: effect of spatial frequency and eccentricity. Soc Neurosci Abstr 735.735/L733, 2006.

Benson PJ, Guo K. Stages in motion processing revealed by the ocular following response. Neuroreport 10: 3803–3807, 1999.[Web of Science][Medline]

Braddick O. A short-range process in apparent motion. Vision Res 14: 519–527, 1974.[CrossRef][Web of Science][Medline]

Brainard DH. The Psychophysics Toolbox. Spat Vis 10: 433–436, 1997.[Web of Science][Medline]

Busettini C, Masson GS, Miles FA. A role for stereoscopic depth cues in the rapid visual stabilization of the eyes. Nature 380: 342–345, 1996.[CrossRef][Medline]

Busettini C, Miles FA, Schwarz U. Ocular responses to translation and their dependence on viewing distance. II. Motion of the scene. J Neurophysiol 66: 865–878, 1991.[Abstract/Free Full Text]

Butzer F, Ilg UJ, Zanker JM. Smooth-pursuit eye movements elicited by first-order and second-order motion. Exp Brain Res 115: 61–70, 1997.[CrossRef][Web of Science][Medline]

Carney T. Evidence for an early motion system which integrates information from the two eyes. Vision Res 37: 2361–2368, 1997.[CrossRef][Web of Science][Medline]

Carney T, Paradiso MA, Freeman RD. A physiological correlate of the Pulfrich effect in cortical neurons of the cat. Vision Res 29: 155–165, 1989.[CrossRef][Web of Science][Medline]

Carney T, Shadlen MN. Dichoptic activation of the early motion system. Vision Res 33: 1977–1995, 1993.[CrossRef][Web of Science][Medline]

Cavanagh P, Mather G. Motion: the long and short of it. Spat Vis 4: 103–129, 1989.[Medline]

Chaudhuri A, Albright TD. Neuronal responses to edges defined by luminance vs. temporal texture in macaque area V1. Vis Neurosci 14: 949–962, 1997.[Web of Science][Medline]

Chubb C, Sperling G. Drift-balanced random stimuli: a general basis for studying non-Fourier motion perception. J Opt Soc Am A Opt Image Sci Vis 5: 1986–2007, 1988.[CrossRef][Web of Science]

Derrington AM, Badcock DR. Separate detectors for simple and complex grating patterns? Vision Res 25: 1869–1878, 1985.[CrossRef][Web of Science][Medline]

Dumoulin SO, Baker CL Jr, Hess RF, Evans AC. Cortical specialization for processing first- and second-order motion. Cereb Cortex 13: 1375–1385, 2003.[Abstract/Free Full Text]

Fuchs AF, Robinson DA. A method for measuring horizontal and vertical eye movement chronically in the monkey. J Appl Physiol 21: 1068–1070, 1966.[Free Full Text]

Gellman RS, Carl JR, Miles FA. Short latency ocular-following responses in man. Vis Neurosci 5: 107–122, 1990.[Web of Science][Medline]

Georgeson MA, Shackleton TM. Monocular motion sensing, binocular motion perception. Vision Res 29: 1511–1523, 1989.[CrossRef][Web of Science][Medline]

Georgeson MA, Shackleton TM. No evidence for dichoptic motion sensing: a reply to Carney and Shadlen. Vision Res 32: 193–198, 1992.[CrossRef][Web of Science][Medline]

Harris LR, Smith AT. Motion defined exclusively by second-order characteristics does not evoke optokinetic nystagmus. Vis Neurosci 9: 565–570, 1992.[Web of Science][Medline]

Harris LR, Smith AT. Interactions between first- and second-order motion revealed by optokinetic nystagmus. Exp Brain Res 130: 67–72, 2000.[CrossRef][Web of Science][Medline]

Hawken MJ, Gegenfurtner KR. Pursuit eye movements to second-order motion targets. J Opt Soc Am A Opt Image Sci Vis 18: 2282–2296, 2001.[Web of Science][Medline]

Hayashi R, Maeda T, Shimojo S, Tachi S. An integrative model of binocular vision: a stereo model utilizing interocularly unpaired points produces both depth and binocular rivalry. Vision Res 44: 2367–2380, 2004.[CrossRef][Web of Science][Medline]

Hayashi R, Miyawaki Y, Maeda T, Tachi S. Unconscious adaptation: a new illusion of depth induced by stimulus features without depth. Vision Res 43: 2773–2782, 2003.[CrossRef][Web of Science][Medline]

Hayashi R, Nishida S, Tolias A, Logothetis NK. A method for generating a "purely first-order" dichoptic motion stimulus. J Vis 7: 1–10, 2007.[Medline]

Hays AV, Richmond BJ, Optican LM. A UNIX-based multiple process system for real-time data acquisition and control. WESCON Conf Proc Session 2: 1–10, 1982.

Hubel DH, Wiesel TN. Receptive fields and functional architecture of monkey striate cortex. J Physiol 195: 215–243, 1968.[Abstract/Free Full Text]

Ilg UJ, Churan J. Motion perception without explicit activity in areas MT and MST. J Neurophysiol 92: 1512–1523, 2004.[Abstract/Free Full Text]

Johnston A, McOwan PW, Buxton H. A computational model of the analysis of some first-order and second-order motion patterns by simple and complex cells. Proc Biol Sci 250: 297–306, 1992.[Abstract/Free Full Text]

Judge SJ, Richmond BJ, Chu FC. Implantation of magnetic search coils for measurement of eye position: an improved method. Vision Res 20: 535–538, 1980.[CrossRef][Web of Science][Medline]

Kawano K, Shidara M, Watanabe Y, Yamane S. Neural activity in cortical area MST of alert monkey during ocular following responses. J Neurophysiol 71: 2305–2324, 1994.[Abstract/Free Full Text]

Kelly DH, Boynton RM, Baron WS. Primate flicker sensitivity: psychophysics and electrophysiology. Science 194: 1077–1079, 1976.[Abstract/Free Full Text]

Ledgeway T, Hutchinson CV. Is the direction of second-order, contrast-defined motion patterns visible to standard motion-energy detectors: a model answer? Vision Res 46: 556–567, 2006.[CrossRef][Web of Science][Medline]

Ledgeway T, Smith AT. Evidence for separate motion-detecting mechanisms for first- and second-order motion in human vision. Vision Res 34: 2727–2740, 1994.[CrossRef][Web of Science][Medline]

Lindner A, Ilg UJ. Initiation of smooth-pursuit eye movements to first-order and second-order motion stimuli. Exp Brain Res 133: 450–456, 2000.[CrossRef][Web of Science][Medline]

Lu ZL, Sperling G. The functional architecture of human visual motion perception. Vision Res 35: 2697–2722, 1995.[CrossRef][Web of Science][Medline]

Lu ZL, Sperling G. Three-systems theory of human visual motion perception: review and update. J Opt Soc Am A Opt Image Sci Vis 18: 2331–2370, 2001a.[Web of Science][Medline]

Lu ZL, Sperling G. Sensitive calibration and measurement procedures based on the amplification principle in motion perception. Vision Res 41: 2355–2374, 2001b.[CrossRef][Web of Science][Medline]

Mareschal I, Baker CL Jr. Temporal and spatial response to second-order stimuli in cat area 18. J Neurophysiol 80: 2811–2823, 1998a.[Abstract/Free Full Text]

Mareschal I, Baker CL Jr. A cortical locus for the processing of contrast-defined contours. Nat Neurosci 1: 150–154, 1998b.[CrossRef][Web of Science][Medline]

Masson GS, Busettini C, Yang DS, Miles FA. Short-latency ocular following in humans: sensitivity to binocular disparity. Vision Res 41: 3371–3387, 2001.[CrossRef][Web of Science][Medline]

Masson GS, Yang DS, Miles FA. Reversed short-latency ocular following. Vision Res 42: 2081–2087, 2002.[CrossRef][Web of Science][Medline]

Miles FA, Kawano K, Optican LM. Short-latency ocular following responses of monkey. I. Dependence on temporospatial properties of visual input. J Neurophysiol 56: 1321–1354, 1986.[Abstract/Free Full Text]

Miura K, Matsuura K, Taki M, Tabata H, Inaba N, Kawano K, Miles FA. The visual motion detectors underlying ocular following responses in monkeys. Vision Res 46: 869–878, 2006.[CrossRef][Web of Science][Medline]

Nienborg H, Bridge H, Parker AJ, Cumming BG. Neuronal computation of disparity in V1 limits temporal resolution for detecting disparity modulation. J Neurosci 25: 10207–10219, 2005.[Abstract/Free Full Text]

Nishida S, Ashida H. A hierarchical structure of motion system revealed by interocular transfer of flicker motion aftereffects. Vision Res 40: 265–278, 2000.[CrossRef][Web of Science][Medline]

Nishida S, Ledgeway T, Edwards M. Dual multiple-scale processing for motion in the human visual system. Vision Res 37: 2685–2698, 1997.[CrossRef][Web of Science][Medline]

Norcia AM, Tyler CW. Temporal frequency limits for stereoscopic apparent motion processes. Vision Res 24: 395–401, 1984.[CrossRef][Web of Science][Medline]

O'Keefe LP, Movshon JA. Processing of first- and second-order motion signals by neurons in area MT of the macaque monkey. Vis Neurosci 15: 305–317, 1998.[CrossRef][Web of Science][Medline]

Patterson R. Stereoscopic (cyclopean) motion sensing. Vision Res 39: 3329–3345, 1999.[CrossRef][Web of Science][Medline]

Plant GT, Laxer KD, Barbaro NM, Schiffman JS, Nakayama K. Impaired visual motion perception in the contralateral hemifield following unilateral posterior cerebral lesions in humans. Brain 116: 1303–1335, 1993.[Abstract/Free Full Text]

Qian N, Andersen RA. A physiological model for motion-stereo integration and a unified explanation of Pulfrich-like phenomena. Vision Res 37: 1683–1698, 1997.[CrossRef][Web of Science][Medline]

Shadlen M, Carney T. Mechanisms of human motion perception revealed by a new cyclopean illusion. Science 232: 95–97, 1986.[Abstract/Free Full Text]

Sheliga BM, Chen KJ, Fitzgibbon EJ, Miles FA. Initial ocular following in humans: a response to first-order motion energy. Vision Res 45: 3307–3321, 2005.[CrossRef][Web of Science][Medline]

Sheliga BM, Kodaka Y, FitzGibbon EJ, Miles FA. Human ocular following initiated by competing image motions: evidence for a winner-take-all mechanism. Vision Res 46: 2041–2060, 2006.[CrossRef][Web of Science][Medline]

Smith AT, Greenlee MW, Singh KD, Kraemer FM, Hennig J. The processing of first- and second-order motion in human visual cortex assessed by functional magnetic resonance imaging (fMRI). J Neurosci 18: 3816–3830, 1998.[Abstract/Free Full Text]

Smith AT, Ledgeway T. Separate detection of moving luminance and contrast modulations: fact or artifact? Vision Res 37: 45–62, 1997.[CrossRef][Web of Science][Medline]

Smith AT, Scott-Samuel NE. First-order and second-order signals combine to improve perceptual accuracy. J Opt Soc Am A Opt Image Sci Vis 18: 2267–2272, 2001.[Web of Science][Medline]

Takemura A, Inoue Y, Kawano K. The effect of disparity on the very earliest ocular following responses and the initial neuronal activity in monkey cortical area MST. Neurosci Res 38: 93–101, 2000.[CrossRef][Web of Science][Medline]

Takemura A, Murata Y, Kawano K, Miles FA. Deficits in short-latency tracking eye movements after chemical lesions in monkey cortical areas MT and MST. J Neurosci 27: 529–541, 2007.[Abstract/Free Full Text]

Taub E, Victor JD, Conte MM. Nonlinear preprocessing in short-range motion. Vision Res 37: 1459–1477, 1997.[CrossRef][Web of Science][Medline]

Vaina LM, Cowey A. Impairment of the perception of second order motion but not first order motion in a patient with unilateral focal brain damage. Proc Biol Sci 263: 1225–1232, 1996.[Abstract/Free Full Text]

Vaina LM, Cowey A, Kennedy D. Perception of first- and second-order motion: separable neurological mechanisms? Hum Brain Mapp 7: 67–77, 1999.[CrossRef][Web of Science][Medline]

Van Der Steen J, Kanai R. Binocular eye movement responses to dichoptically presented horizontal and/or vertical stimulus steps. Ann NY Acad Sci 956: 487–491, 2002.[Web of Science][Medline]

Victor JD, Conte MM. Evoked potential and psychophysical analysis of Fourier and non-Fourier motion mechanisms. Vis Neurosci 9: 105–123, 1992.[Web of Science][Medline]

Wilson HR, Ferrera VP, Yo C. A psychophysically motivated model for two-dimensional motion perception. Vis Neurosci 9: 79–97, 1992.[Web of Science][Medline]

Wilson HR, Kim J. A model for motion coherence and transparency. Vis Neurosci 11: 1205–1220, 1994.[Web of Science][Medline]

Zanker JM, Hupgens IS. Interaction between primary and secondary mechanisms in human motion perception. Vision Res 34: 1255–1266, 1994.[CrossRef][Web of Science][Medline]

Zhou YX, Baker CL Jr. A processing stream in mammalian visual cortex neurons for non-Fourier responses. Science 261: 98–101, 1993.[Abstract/Free Full Text]

Zhou YX, Baker CL Jr. Envelope-responsive neurons in areas 17 and 18 of cat. J Neurophysiol 72: 2134–2150, 1994.[Abstract/Free Full Text]

This article has been cited by other articles:

				K. Miura, Y. Sugita, K. Matsuura, N. Inaba, K. Kawano, and F. A. Miles The Initial Disparity Vergence Elicited With Single and Dual Grating Stimuli in Monkeys: Evidence for Disparity Energy Sensing and Nonlinear Interactions J Neurophysiol, November 1, 2008; 100(5): 2907 - 2918. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Supplemental Videos All Versions of this Article: 99/5/2329    most recent 01316.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Hayashi, R. Articles by Kawano, K. Search for Related Content PubMed PubMed Citation Articles by Hayashi, R. Articles by Kawano, K.