Free
Article  |   February 2011
The impact of stimulus complexity and frequency swapping on stabilization of binocular rivalry
Author Affiliations
Journal of Vision February 2011, Vol.11, 6. doi:10.1167/11.2.6
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Kristian Sandberg, Bahador Bahrami, Jonas Kristoffer Lindeløv, Morten Overgaard, Geraint Rees; The impact of stimulus complexity and frequency swapping on stabilization of binocular rivalry. Journal of Vision 2011;11(2):6. doi: 10.1167/11.2.6.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Binocular rivalry occurs when an image is presented to one eye while at the same time another, incongruent, image is presented to the other eye in the corresponding retinotopic location and conscious perception alternates spontaneously between the two monocular views. If a short blank period is inserted between intermittent presentations of rivaling stimuli, perception is stabilized and spontaneous alternations are drastically reduced. Whether the complexity of rivaling stimuli plays a role in stabilization is unknown. We replicated previous findings that swapping the stimuli between eyes across presentations abolishes stabilization for Gabors, but for more complex stimuli (a face and a house in our experiment), stabilization is eye-specific and not disrupted. Phase scrambling the rivaling face and house images did not change the stabilization pattern showing that the pattern can be observed without high-level perceptual content. We conclude that overlaps at low visual stages are the most likely cause of the eye-specific stabilization for both stimulus types. Additionally, we examined the impact of swapping the flicker frequency of the images and found a general impact on stabilization not specific to stimulus type. Taken together, the findings indicate that choice of stimulus features impact greatly on the results obtained in stabilization paradigms.

Introduction
Bistable stimuli such as the Necker cube (Necker, 1832) and Rubin's vase (Rubin, 1915) are visual stimuli that are spontaneously perceived in two different ways and thus allow dissociation of changes in conscious perception from changes in physical stimulation (Andrews, Schluppeck, Homfray, Matthews, & Blakemore, 2002; Lumer, Friston, & Rees, 1998). Binocular rivalry (BR; Breese, 1899) is one form of bistable perception that occurs when an image is presented to one eye while at the same time another, incongruent, image is presented to the other eye. Perception alternates spontaneously between each monocular view every few seconds. 
If a short blank period (a stabilization interval) is inserted between intermittent presentation of a bistable stimulus, the perceptual switch rate drops drastically (Leopold, Wilke, Maier, & Logothetis, 2002; Orbach, Ehrlich, & Heath, 1963). Perception in consecutive trials stabilizes to one of the two alternatives implying the existence of a perceptual memory across subsequent trials (Leopold et al., 2002). This memory effect has been called percept stabilization, percept maintenance, or simply stabilization (Leopold et al., 2002; Sterzer & Rees, 2008). It has recently been suggested that it is mediated by accumulative changes in the sensitivity of neural populations representing each alternative percept and that these changes in sensitivity occur on multiple time scales (Brascamp, Pearson, Blake, & van den Berg, 2009; Brascamp et al., 2008; Noest, van Ee, Nijs, & van Wezel,2007). 
In the above-mentioned studies of stabilization, images are presented to the same eye throughout the experiment, and it is thus impossible to determine whether the stabilization effect occurs due to facilitation of neurons selective for one of the stimuli or alternatively for one eye (i.e., whether interstimulus or interocular competition is modulated). Chen and He (2004) and Pearson and Clifford (2004) disentangled the two by swapping the two stimuli between the two eyes across intermittent presentations. In such a paradigm, perceptual stabilization and what might be called eye signal stabilization can thus be defined separately: Perceptual stabilization occurs when an observer reports seeing the same image on two consecutive trials whereas eye signal stabilization occurs when he reports the stimulus associated with the same eye on two consecutive trials. Chen and He reported that the eye signal appeared to be stabilized and not the percept. Similarly, based on their first experiment, Pearson and Clifford conclude that eye of origin is the most important factor for BR stabilization and make references to the similarity of the findings of the studies. However, Pearson and Clifford do not report complete stabilization of the eye. When their results are plotted instead as eye signal stabilization (Figure 1), a substantial difference between their findings and those on Chen and He can now be seen. Whereas image swapping has no impact on stabilization in Chen and He's study, Pearson and Clifford find stabilization to be reduced to chance levels when averaged across subjects. Thus, the conclusion of Chen and He that the direction of suppression between the eyes is stabilized conflicts with the findings of Pearson and Clifford. Possible causes could be differences in the neural processing of the stimuli or simply differences in the experimental paradigm. Pearson and Clifford used Gabor patches—low-level stimuli processed principally in the primary visual cortex (V1)—while Chen and He used radial and concentric gratings, the processing of which are presumably performed in V4 through the global pooling of local V1 orientations (Wilkinson et al., 2000). 
Figure 1
 
Data reproduced approximately from Chen and He (2004) and Pearson and Clifford (2004), Experiment 1. Results are plotted as eye signal stabilization instead of percept stabilization. Average chance of eye signal stabilization across subjects as a function of image swap is shown.
Figure 1
 
Data reproduced approximately from Chen and He (2004) and Pearson and Clifford (2004), Experiment 1. Results are plotted as eye signal stabilization instead of percept stabilization. Average chance of eye signal stabilization across subjects as a function of image swap is shown.
Recent models of BR propose that rivalry arises from competition between neuronal populations at various stages of the visual system in a multi-level process (Tong, Meng, & Blake, 2006; Wilson, 2003). At early stages—the lateral geniculate nucleus (LGN) and V1—competition between monocular neurons is associated with BR between Gabor patches (Haynes, Deichmann, & Rees, 2005; Haynes & Rees, 2005) similar to those used by Pearson and Clifford (2004). Swapping the stimuli between the eyes will clearly change which population of monocular neurons is stimulated. At higher levels of the visual system, however, processing of stimulus features occurs independently of the eye of origin (although of course some monocular biases may persist in binocularly driven cells in early extrastriate cortex), and BR between images of complex objects (such as faces and houses) may result from competition between binocular neurons representing different stimulus features (Haynes & Rees, 2005; Tong, Nakayama, Vaughan, & Kanwisher, 1998). Since the level of neuronal competition determining the course of binocular rivalry seems to depend on stimulus characteristics, it is conceivable that stabilization of any mnemonic trace during intermittent rivalry might thus also occur in different levels of visual processing. 
This hypothesis would suggest, in contrast to Chen and He (2004), that the mnemonic trace of a high-level stimulus will be independent of eye of origin. This difference between hypothesis and finding could be explained in two ways. One possibility is that an eye-specific feedback mechanism could inhibit eye-specific neurons at early stages of the visual system while stimulus processing proceeds at later stages. Feedback operates during continuous BR (van Boxtel, Alais, & van Ee, 2008). Such a mechanism is consistent with the long-lasting notion of BR suppression being eye-specific (Asher, 1953; du Tour, 1760) and supported by experimental findings that test probes presented to an eye during suppression are more difficult to detect than test probes presented during dominance (Fox & Check, 1972). Alternatively, Chen and He's findings might result from their broadband stimuli (which unlike those of Pearson & Clifford, 2004 consisted of many different orientations) activating contiguous orientation columns and resulting in greater overlap of monocularly biased neurons at early stages of the visual system, thus leaving a mnemonic trace to impact perception on a trial-by-trial basis. 
Thus, the results obtained by Pearson and Clifford might reflect their stimuli being processed by non-overlapping monocularly biased neurons at an early visual stage, possibly V1. In contrast, the results obtained by Chen and He (2004) might be explained by their stimuli activating overlapping monocular neurons or via an eye-specific feedback mechanism from downstream stages of visual perception back to early visual areas. 
Here we conducted three experiments to test these hypotheses. To address the discrepancy between Chen and He (2004) and Pearson and Clifford (2004) directly, we present our results in terms of eye signal stabilization. However, as previous results have been reported in terms of perceptual stabilization, we will refer to those as BR stabilization in general. When distinctions are needed, we refer to eye signal stabilization or perceptual stabilization separately. 
In the literature, the phenomenon of BR stabilization is influenced by factors such as position in visual space, color, and stimulus orientation (Chen & He, 2004; Pearson & Clifford, 2004). For this reason, we also sought to examine whether something not usually considered a stimulus feature proper, the flicker frequency (frequency tagging) of the stimulus, can impact on BR stabilization. Researchers often use images flickering at different frequencies in EEG/MEG experiments in order to “tag” each image so they produce distinct neural responses (Pastor, Artieda, Arbizu, Valencia, & Masdeu, 2003). As mentioned, the flicker frequency of an image has not traditionally been considered a feature of the image as such, but rather a way of attaching a particular neural signature to a stimulus (see, e.g., Lansing, 1964). Such frequency tagging has previously been used in the majority of EEG experiments using BR (Brown & Norcia, 1997; Kamphuisen, Bauer, & van Ee, 2008; Lansing, 1964; Srinivasan, Russell, Edelman, & Tononi, 1999) and may be used in the future in BR stabilization experiments. If frequency swapping affects BR stabilization, however, this would be a potential confound in future EEG/MEG studies of stabilization. 
We examined all goals of the study in three experiments. In Experiment 1, we replicated the findings of Pearson and Clifford (2004) using Gabor patches. Additionally, we conducted further analyses that indicated that our findings did indeed reflect a mnemonic trace in monocular neurons. In Experiment 2, we replicated the findings of Chen and He (2004) using higher level (face and house) stimuli. In Experiment 3, we tested whether phase scrambling the faces and houses while keeping the power spectra identical to Experiment 2 (thus removing the high-level meaningful content of rivaling images) modulates eye-specific stabilization. In Experiments 1–2, the role of frequency swapping was examined. In Experiment 3, only the impact of image swapping was examined and the two stimuli were flickering at the same frequency. 
Before reporting our findings, it should be noted that this classical type of BR for which the stabilization effect is found is distinct from other types of rivalry caused by different images being presented to each eye. For instance, swapping rivaling stimuli between the eyes every 300 ms or so does not cause perception to alternate rapidly, but instead produces regular rivalry alternations, i.e., stimulus rivalry (Logothetis, Leopold, & Sheinberg, 1996). Moreover, information from one half of an image presented to one eye can be grouped with the other half of an image presented to the other eye and thus cause perception to alternate between the grouped images and not the eyes (Kovács, Papathomas, Yang, & Fehér, 1996). The workings of these phenomena are distinct from conventional BR in that the eye-of-origin component has been neutralized (Pearson & Clifford, 2004). An explanation of these phenomena (and stabilization within such paradigms) is beyond the scope of the present study. 
Experiments 1 and 2
Methods
Experiments 1 and 2 sought to replicate the findings of previous experiments as well as conduct additional analyses and examine the role of frequency tagging on BR stabilization. 
Participants
Eight healthy young adults with normal or corrected-to-normal vision gave informed consent to participate in each experiment. In Experiment 1, the age of the participants was between 22 and 36 years; four participants were female. In Experiment 2, the age of the participants was between 24 and 34 years; three participants were female. The experiments were reviewed by the local ethics committee. 
Apparatus
Stimuli were generated using the MATLAB toolbox Cogent (www.vislab.ucl.ac.uk/Cogent/). They were displayed on a CRT monitor (17″ in Experiment 1, 19″ in Experiment 2) with a screen resolution of 1024 × 768 at a refresh rate of 60 Hz. Participants viewed the stimuli through a mirror stereoscope positioned at approximately 50 cm from the monitor. 
Stimuli
In Experiment 1, stabilization of a green and a red Gabor patch (contrast = 100%, spatial frequency = 4 cycles/degree, standard deviation of the Gaussian envelope = 10 pixels) was examined (Figure 2). The green Gabor patch was tilted 45 degrees counterclockwise and the red Gabor patch was tilted 45 degrees clockwise. In Experiment 2, stabilization of a red face and a green house was examined. In both experiments, each stimulus was presented within an annulus (inner/outer r = 2/3 degrees of visual angle) consisting of randomly oriented lines. In the center of the circle was a small circular fixation dot. The luminance of the stimuli was set for each participant (see Procedure section). 
Figure 2
 
Binocular rivalry stimuli. Stimuli were dichoptically presented to the eyes of the participant using a mirror stereoscope. Stimuli were flickering at a rate between 6 and 15 Hz. First row: Grating stimuli used in Experiment 1. Second row: Face/house stimuli used in Experiment 2. Third row: Phase-scrambled face/house stimuli used in Experiment 3.
Figure 2
 
Binocular rivalry stimuli. Stimuli were dichoptically presented to the eyes of the participant using a mirror stereoscope. Stimuli were flickering at a rate between 6 and 15 Hz. First row: Grating stimuli used in Experiment 1. Second row: Face/house stimuli used in Experiment 2. Third row: Phase-scrambled face/house stimuli used in Experiment 3.
Flickering was applied by removing the stimulus from screen at a certain rate. For all participants, the first image flickered at a frequency of 6 Hz. For one half of the participants (selected at random), the second image flickered at 7.5 Hz; for the other half of the participants, it flickered at 15 Hz. No systematic difference was observed in the response pattern as an effect of fast or slow flicker, and the two subsets of the data were analyzed as one. 
Procedure
Participants looked into the mirror stereoscope while the fixation circles around the stimuli were displayed, and the position of the circles as well as the mirrors of the stereoscope was calibrated until the circles fused. When stimuli were displayed, participants reported what they saw using one of three buttons, each corresponding to a report option specific to the experiment. In Experiment 1, the report options were left (counterclockwise)-tilted grating, right (clockwise)-tilted grating, and mixed perception. In Experiment 2, the options were face, house, and mixed perception. 
It is possible that BR stabilization in some cases is not a result of perceptual memory but instead perceptual bias (Carter & Cavanagh, 2007). To minimize bias, we increased the chances that participants would report each percept equally often during the experiment by adjusting the relative luminance of the images for each participant before each experiment. The starting luminance for each image was maximum screen value, and one value was decreased until the participant reported seeing both images equally often (±7%) during a 1-min-long continuous presentation. This was done for separately for each eye. 
The experiments consisted of 32 blocks presented in a pseudorandom order. Each block consisted of 20 trials. In Experiment 1, each trial consisted of a stimulation period of approximately 700 ms followed by a blank screen (the stabilization interval) lasting 1800 ms. The trial duration was selected so that participants experienced a stable perceptual state, i.e., the percept had time to stabilize and did not switch during the stimulation period. As indicated by early findings that complex stimuli switch at a slower rate (Rogers, Rogers, & Tootle, 1977), we found that the face/house rivalry stabilized and switched slower than grating rivalry. For that reason, the stimulation period was slightly longer in Experiment 2. Stimulus durations of approximately 1500 ms and blank periods of 1500 ms were used. 
Between successive presentations of the BR stimuli separated by a stabilization interval, we swapped either the eye to which each image was presented, or the flicker frequency associated with one of the images, in order to determine whether this had an effect on BR stabilization. The experiment employed a factorial design with stabilization proportion as the dependent variable and frequency swap and image swap as independent variables (Figure 3). Hence, there were four experimental conditions. In the first, the image was always presented to the same eye with the same tagging frequency within each block (eye and frequency were counterbalanced across blocks). In the second condition, the image was always presented to the same eye within each block (counterbalanced across blocks), but the tagging frequency was swapped between eyes between trials. In the third condition, the image was swapped between eyes between trials, but the tagging frequency always displayed to the same eye (counterbalanced across blocks). In the fourth condition, both image and tagging frequency were swapped between eyes between trials. 
Figure 3
 
The four experimental conditions employed in Experiments 13. In condition 1, the stimuli were displayed in the same way on every trial; in condition 2, the flicker frequencies of the stimuli are swapped between eyes on consecutive trials; in condition 3, the images were swapped between eyes on consecutive trials; in condition 4, both flicker frequency and image were swapped between eyes on consecutive trials. In Experiment 3, identical flicker frequencies were used for the two stimuli, thus effectively leaving only two conditions: No image swap and image swap.
Figure 3
 
The four experimental conditions employed in Experiments 13. In condition 1, the stimuli were displayed in the same way on every trial; in condition 2, the flicker frequencies of the stimuli are swapped between eyes on consecutive trials; in condition 3, the images were swapped between eyes on consecutive trials; in condition 4, both flicker frequency and image were swapped between eyes on consecutive trials. In Experiment 3, identical flicker frequencies were used for the two stimuli, thus effectively leaving only two conditions: No image swap and image swap.
Calculation of stabilization
As mentioned above, the main predictor of stabilization is eye of origin. For this reason, we calculated stabilization from the perspective of the eye, not the percept. Participants were asked to report the identity of the image rather than the eye of origin as utrocular discrimination is rarely possible, and eye signal stabilization was thus calculated from the reported percept. 
For blocks in which the stimuli were not swapped, eye signal stabilization was identical to percept stabilization and was thus given by 
s t a b i l i z a t i o n e y e = s t a b i l i z a t i o n p e r c e p t ,
(1)
where stabilizationpercept = 1 if the same percept was reported before and after the stabilization interval, and stabilizationpercept = 0 if the reported percept changed across such an interval. When the stimuli were swapped between eyes between presentations, perception followed either the eye or the percept; i.e., if the same percept was reported, the reported image was presented to a different eye before and after the break in the presentation, and if a different percept was reported, those stimuli were presented to the same eye before and after the break. For blocks in which the stimuli were swapped, eye signal stabilization was thus given by 
s t a b i l i z a t i o n e y e = 1 s t a b i l i z a t i o n p e r c e p t .
(2)
 
Results
Average eye signal stabilization was calculated for each participant using only trials in which the participant reported a single clear percept, i.e., trials during which participants reported multiple or mixed percepts were excluded from the analysis. We excluded the possibility of perceptual bias causing the results by examining that participants reported the images equally often: Images 1 and 2 were reported on 41% vs. 59% of the trials in Experiment 1 and on 47% vs. 53% in Experiment 2. Importantly, all images were stabilized above chance level on −80% vs. 66% in Experiment 1 and 79% vs. 75% in Experiment 2
Data from Experiments 1 and 2 were then analyzed in a single mixed analysis of variance with eye signal stabilization as the dependent variable and image swap and frequency swap as related conditions and experiment as the unrelated condition (Figure 4). We found an effect of image swap (F(1,14) = 23.36, p < 0.0001) but also an interaction between image swap and experiment (F(1,14) = 27.67, p < 0.0001). An effect of frequency swap was also found (F(1,14) = 9.20, p < 0.01), and there was a just significant interaction between image swap and frequency swap (F(1,14) = 5.49, p < 0.05). There was no interaction between frequency swap and experiment (F(1,14) = 0.15, p = 0.71) or between image swap, frequency swap, and experiment (F(1,14) = 1.88, p = 0.19). There was no effect of experiment alone on stabilization (F(1,14) = 0.90, p = 0.36). Additional one-way ANOVAs revealed that for rivalry between Gabor patches (Experiment 1), image swap disrupted BR stabilization highly significantly (F(1) = 38.9, p < 0.0001), whereas there was no such effect for face/house stimuli (Experiment 2; F(1) = 0.11, p = 0.74). Thus, this analysis (1) replicated the findings that that image swap affects BR stabilization for Gabor patches (Pearson & Clifford, 2004) but not for rivaling images with more complex structure (Chen & He, 2004) and (2) indicated that frequency swapping has an impact on eye signal stabilization independently of whether Gabor patches or face/house stimuli are used. 
Figure 4
 
Average chance of eye signal stabilization across subjects as a function of image swap and frequency swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals. (A) Results of Experiment 1 (rivalry between Gabor patches). (B) Results of Experiment 2 (rivalry between a face and a house).
Figure 4
 
Average chance of eye signal stabilization across subjects as a function of image swap and frequency swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals. (A) Results of Experiment 1 (rivalry between Gabor patches). (B) Results of Experiment 2 (rivalry between a face and a house).
BR stabilization of face/house stimuli thus followed the eye of origin, as eye signal stabilization did not change as a function of image swap. In contrast, BR stabilization of Gabor patches was disrupted and the reported percept appeared random when images were swapped between eyes across presentations. In order to confirm that the reported percepts were indeed random, we tested the hypothesis that BR stabilization was different from chance for the two image swap conditions of Experiment 1, the most important of these being condition 4 in which both stimulus features (image identity and flicker frequency) were swapped between trials. We found that BR stabilization did not differ from chance neither for condition 3 (t(7) = 0.94, p = 0.38) nor condition 4 (t(7) = 0.01, p = 0.99). In comparison, BR stabilization in conditions 1 (t(7) = 26.37, p < 0.0001) and 2 (t(7) = 13.20, p < 0.0001) of Experiment 1 clearly differed from chance, and it did so in all conditions of Experiment 2 (condition 1: t(7) = 7.50, p = 0.0001, condition 2: t(7) = 6.81, p < 0.0005, condition 3: t(7) = 11.51, p < 0.0001, and condition 4: t(7) = 3.75, p < 0.01 (all tests uncorrected)). 
If the pattern observed in Experiment 1 was caused by the stabilization memory trace being stored in monocular neuronal populations, then swapping the images between eyes across trials would effectively be identical to temporally interleaving two unrelated rivalry pairs as has been done by Maier, Wilke, Logothetis, and Leopold (2003) in a study of ambiguous figure/motion perception. In their study, stabilization was found in spite of the interleaved stimuli. In the image swap conditions of our experiments, the presented stimuli could similarly be considered trials with interleaved stimuli as one rivalry pair was presented on trials 1, 3, 5, etc. and another on trials 2, 4, 6, etc. If stabilization occurs in spite of interleaved stimuli for BR—as was observed for ambiguous figures/motion—we would thus expect a high degree of stabilization between the reported percept between every pair of consecutive odd and even presentations, i.e., stabilization should occur across trial 1, 3, 5, etc. and separately across trial 2, 4, 6, etc. As shown in Figure 5, this was indeed the case in our data. Eye signal stabilization did not vary across condition (F(3,28) = 0.30, p = 0.83), and the average degree of stabilization (93%) different highly significantly from chance (t(31) = 39.7, p < 0.0001). 
Figure 5
 
Data from Experiment 1. Eye signal stabilization is calculated for every odd and every even pair of trials. Average chance of eye signal stabilization across subjects as a function of the four experimental conditions is shown. Note that although conditions are labeled by what is usually swapped in that condition, nothing is swapped across every other trial in any of the conditions. Contrary to stabilization across every single trial, stabilization is not disrupted across every other trial in any condition, and it is thus unaffected by the intermediate presentation of an image that activates a different subset of monocular neurons. The finding is in line with our proposal that the memory trace is situated in monocular neurons in the early stages of visual processing.
Figure 5
 
Data from Experiment 1. Eye signal stabilization is calculated for every odd and every even pair of trials. Average chance of eye signal stabilization across subjects as a function of the four experimental conditions is shown. Note that although conditions are labeled by what is usually swapped in that condition, nothing is swapped across every other trial in any of the conditions. Contrary to stabilization across every single trial, stabilization is not disrupted across every other trial in any condition, and it is thus unaffected by the intermediate presentation of an image that activates a different subset of monocular neurons. The finding is in line with our proposal that the memory trace is situated in monocular neurons in the early stages of visual processing.
Discussion
We found that the effects of image swapping on BR stabilization differed substantially when comparing Gabor patches and face/house stimuli. For Gabor patches, swapping the images between presentations disrupted stabilization, with the reported percept being determined no more consistently than chance on consecutive trials. Analysis of stabilization on odd or even pairs of trials confirmed that eye signals were stabilized across every other trial. This suggests that a mnemonic trace was established. We propose that the most likely explanation for this pattern of stabilization is that non-overlapping populations of monocular early visual neurons carry such a mnemonic signal. This interpretation is consistent with the recent findings that stabilization of Gabor patches decreases as a function of distance in eccentricity and that perception is almost chance at eccentricities where no overlap is to expected in the receptive field of V1 neurons (Knapen, Brascamp, Adams, & Graf, 2009). In contrast to the pattern observed for Gabor patches, image swapping had no impact on eye signal stabilization for face/house stimuli. 
We found that swapping the tagging frequency of the rivaling stimuli disrupted BR stabilization and that no difference was observed between experiments. Though the disruption was quite small, it is potentially important as it demonstrates that frequency tagging can, in fact, be considered an object feature that influences the processing of the stimuli. This is consistent with a previous finding showing that frequency tagging can bias perception in the interocular switching paradigm (Silver & Logothetis, 2007). Acknowledging this property of frequency tagging will be important in any future EEG and MEG studies examining electro- or magnetoencephalic correlates of BR stabilization, especially if image characteristics are swapped between trials. Special notice should be paid to the observation that in Experiment 2, frequency swapping was the only factor impacting on eye signal stabilization. 
The results of Experiment 2 replicated those of Chen and He (2004). The results, however, do not allow us to distinguish between the hypotheses that BR stabilization for the face/house stimuli is caused by mnemonic traces in overlapping monocular populations of early visual areas or, alternatively, by an eye-specific feedback mechanism from later downstream stages of visual processing back onto early visual neurons. In order to distinguish these possibilities, an additional experiment was conducted. 
Experiment 3
To identify whether feedback mechanisms working at later visual stages was necessary for eye-specific stabilization (as observed in Experiment 2), we tested whether meaningful content of rivaling is necessary or rather having a complex (even though meaningless) image structure should be sufficient for maintaining stabilization as we saw in Experiment 2. We therefore removed all high-level meaningful content from the images by phase scrambling the rivaling images and conducted a follow-up experiment. 
Methods
Experiment 3 was identical to Experiment 2 with the following changes: 9 participants were tested, and instead of meaningful face/house stimuli, phase-scrambled versions of the same images were used (Figure 2). Stimulus durations of 1200 ms were found to allow perception to stabilize on each trial and prevent perceptual switches within trials for most subjects. This need for slightly lower stimulus durations for successful BR compared to Experiment 2 could be causes by decreased complexity of the stimuli due to phase scrambling. Flicker rates were always 7.5 Hz as the impact of frequency tagging was not examined in this experiment. 
Results
Experiment 3 had effectively only two conditions, “No image swap” and “Image swap.” The results are plotted in Figure 6. A paired t-test revealed no difference in eye signal stabilization between conditions (t(8) = 0.65, p = 0.53), and the average proportion of stabilized percepts was above chance (t(17) = 5.16, p = 0.0001). The results of Experiment 2 were thus replicated. 
Figure 6
 
Results of Experiment 3 (rivalry between phase-scrambled face/house images). Average chance of eye signal stabilization across subjects as a function of image swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals.
Figure 6
 
Results of Experiment 3 (rivalry between phase-scrambled face/house images). Average chance of eye signal stabilization across subjects as a function of image swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals.
Discussion
Experiment 3 replicated the findings of Experiment 2 using phase-scrambled face/house images. We thus found that BR stabilization of eye of origin was not dependent on high-level processing but also occurs when stimuli lack meaningful, high-level feature/object content. Overall, the experiments of the study support the notion that BR stabilization reflects a memory trace in early visual areas and that this memory trace is located in non-overlapping populations of monocular neurons for rivalry between Gabor patches but in overlapping populations of monocular neurons for the other stimulus types tested. 
Conclusions
We have identified a previously unnoticed difference in the stabilization patterns of Gabor patches and other stimuli. In two experiments using the same experimental design, we replicated the finding that swapping the stimuli between eyes across presentations caused chance perception for Gabors, but that for other stimuli (a face and a house in our case) stabilization was eye-specific. Further analyses revealed that perception was stabilized completely on every odd pair and every even pair of trials for Gabor patches, consistent with a mnemonic signal associated with each Gabor being stored in non-overlapping populations of monocular neurons. Experiment 3 examined whether face/house stabilization was caused by an eye-specific feedback mechanism. This was performed by phase scrambling the images thus obscuring high-level features. In this experiment, we still found no difference in the stabilization pattern. We thus conclude that overlapping neuronal populations at early stages of visual processing are the most likely cause of the eye-specific stabilization. Examining rivalry between a Gabor patch and a broadband stimulus would be an interesting line of future research. 
Additionally, we examined the impact of swapping the flicker frequency of the images and found a general impact on stabilization not specific to stimulus type. This finding demonstrates that frequency tagging should be considered an object feature that influences the processing of the stimulus. Taken together, the findings indicate that choice of stimulus features impact greatly on the results obtained in stabilization paradigms. 
Acknowledgments
This work was supported by the Wellcome Trust (GR), by the MindBridge project, funded by the European Commission under the Sixth Framework Programme (BB), and by a Starting Grant, European Research Council (KS and MO). We thank Karl Friston for helpful discussions at an early stage of the study. 
Commercial relationships: none. 
Corresponding author: Kristian Sandberg. 
Email: krissand@rm.dk. 
Address: Cognitive Neuroscience Research Unit, Hammel Neurorehabilitation and Research Center, MindLab, Aarhus University, Voldbyvej 15, Hammel 8450, Denmark. 
References
Andrews T. J. Schluppeck D. Homfray D. Matthews P. Blakemore C. (2002). Activity in the fusiform gyrus predicts conscious perception of Rubin's vase–face illusion. Neuroimage, 17, 890–901. [CrossRef] [PubMed]
Asher H. (1953). Suppression theory of binocular vision. British Journal of Ophthalmology, 37, 37–49. [CrossRef] [PubMed]
Brascamp J. W. Pearson J. Blake R. van den Berg A. V. (2009). Intermittent ambiguous stimuli: Implicit memory causes periodic perceptual alternations. Journal of Vision, 9, (3):3, 1–23, http://www.journalofvision.org/content/9/3/3, doi:10.1167/9.3.3. [PubMed] [Article] [CrossRef] [PubMed]
Brascamp J. W. Knapen T. H. J. Kanai R. Noest A. J. van Ee R. van den Berg A. V. et al. (2008). Multi-timescale perceptual history resolves visual ambiguity. PLoS ONE, 3, e1497. [CrossRef] [PubMed]
Breese B. B. (1899). On inhibition. Psychological Monographs, 3, 1–65. [CrossRef]
Brown R. J. Norcia A. M. (1997). A method for investigating binocular rivalry in real-time with the steady-state VEP. Vision Research, 37, 2401–2408. [CrossRef] [PubMed]
Carter O. Cavanagh P. (2007). Onset rivalry: Brief presentation isolates an early independent phase of perceptual competition. PLoS ONE, 2, e343. [CrossRef] [PubMed]
Chen X. He S. (2004). Local factors determine the stabilization of monocular ambiguous and binocular rivalry stimuli. Current Biology, 14, 1013–1017. [CrossRef] [PubMed]
du Tour E. F. (1760). Discussion d'une question d'optique. Mémoires de Mathématique et de Physique Présentés par Divers Savants (l'Académie des Sciences), 3, 514–530.
Fox R. Check R. (1972). Independence between binocular rivalry suppression duration and magnitude of suppression. Journal of Experimental Psychology, 93, 283–289. [CrossRef] [PubMed]
Haynes J. Deichmann R. Rees G. (2005). Eye-specific effects of binocular rivalry in the human lateral geniculate nucleus. Nature, 438, 496–499. [CrossRef] [PubMed]
Haynes J. Rees G. (2005). Predicting the stream of consciousness from activity in human visual cortex. Current Biology: CB, 15, 1301–1307. [CrossRef] [PubMed]
Kamphuisen A. Bauer M. van Ee R. (2008). No evidence for widespread synchronized networks in binocular rivalry: MEG frequency tagging entrains primarily early visual cortex. Journal of Vision, 8, (5):4, 1–8, http://www.journalofvision.org/content/8/5/4, doi:10.1167/8.5.4. [PubMed] [Article] [CrossRef] [PubMed]
Knapen T. Brascamp J. Adams W. J. Graf E. W. (2009). The spatial scale of perceptual memory in ambiguous figure perception. Journal of Vision, 9, (13):16, 1–12, http://www.journalofvision.org/content/9/13/16, doi:10.1167/9.13.16. [PubMed] [Article] [CrossRef]
Kovács I. Papathomas T. V. Yang M. Fehér A. (1996). When the brain changes its mind: Interocular grouping during binocular rivalry. Proceedings of the National Academy of Sciences of the United States of America, 93, 15508–15511. [CrossRef] [PubMed]
Lansing R. W. (1964). Electroencephalographic correlates of binocular rivalry in man. Science, 146, 1325–1327. [CrossRef] [PubMed]
Leopold D. A. Wilke M. Maier A. Logothetis N. K. (2002). Stable perception of visually ambiguous patterns. Nature Neuroscience, 5, 605–609. [CrossRef] [PubMed]
Logothetis N. K. Leopold D. A. Sheinberg D. L. (1996). What is rivalling during binocular rivalry? Nature, 380, 621–624. [CrossRef] [PubMed]
Lumer E. D. Friston K. J. Rees G. (1998). Neural correlates of perceptual rivalry in the human brain. Science, 280, 1930–1934. [CrossRef] [PubMed]
Maier A. Wilke M. Logothetis N. K. Leopold D. A. (2003). Perception of temporally interleaved ambiguous patterns. Current Biology: CB, 13, 1076–1085. [CrossRef] [PubMed]
Necker L. A. (1832). Observations on some remarkable optical phenomena seen in Switzerland; and on an optical phenomenon which occurs on viewing a figure of a crystal or geometrical solid. Philosophical Magazine Journal Science, 1, 329–337.
Noest A. J. van Ee R. Nijs M. M. van Wezel R. J. (2007). Percept-choice sequences driven by interrupted ambiguous stimuli: A low-level neural model. Journal of Vision, 7, (8):10, 1–14, http://www.journalofvision.org/content/7/8/10, doi:10.1167/7.8.10. [PubMed] [Article] [CrossRef] [PubMed]
Orbach J. Ehrlich D. Heath H. (1963). Reversibility of the Necker cube: I. An examination of the concept of “satiation of orientation”. Perceptual and Motor Skills, 17, 439–458. [CrossRef] [PubMed]
Pastor M. A. Artieda J. Arbizu J. Valencia M. Masdeu J. C. (2003). Human cerebral activation during steady-state visual-evoked responses. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 23, 11621–11627. [PubMed]
Pearson J. Clifford C. (2004). Determinants of visual awareness following interruptions during rivalry. Journal of Vision, 4, (3):6, 196–202, http://www.journalofvision.org/content/4/3/6, doi:10.1167/4.3.6. [PubMed] [Article] [CrossRef]
Rogers R. L. Rogers S. W. Tootle J. S. (1977). Stimulus complexity and rate of alternation in binocular rivalry. Perceptual and Motor Skills, 44, 669–670. [CrossRef] [PubMed]
Rubin E. (1915). Synsoplevede figurer. Studier i psykologisk Analyse. Første del. København, Denmark: Gyldendalske Boghandel, Nordisk Forlag.
Silver M. A. Logothetis N. K. (2007). Temporal frequency and contrast tagging bias the type of competition in interocular switch rivalry. Vision Research, 47, 532–543. [CrossRef] [PubMed]
Srinivasan R. Russell D. P. Edelman G. M. Tononi G. (1999). Increased synchronization of neuromagnetic responses during conscious perception. The Journal of Neuroscience: The Official Journal of the Society for Neuroscience, 19, 5435–5448. [PubMed]
Sterzer P. Rees G. (2008). A neural basis for percept stabilization in binocular rivalry. Journal of Cognitive Neuroscience, 20, 389–399. [CrossRef] [PubMed]
Tong F. Meng M. Blake R. (2006). Neural bases of binocular rivalry. Trends in Cognitive Sciences, 10, 502–511. [CrossRef] [PubMed]
Tong F. Nakayama K. Vaughan J. T. Kanwisher N. (1998). Binocular rivalry and visual awareness in human extrastriate cortex. Neuron, 21, 753–759. [CrossRef] [PubMed]
van Boxtel J. J. A. Alais D. van Ee R. (2008). Retinotopic and non-retinotopic stimulus encoding in binocular rivalry and the involvement of feedback. Journal of Vision, 8, (5):17, 1–10, http://www.journalofvision.org/content/8/5/17, doi:10.1167/8.5.17. [PubMed] [Article] [CrossRef] [PubMed]
Wilkinson F. James T. W. Wilson H. R. Gati J. S. Menon R. S. Goodale M. A. (2000). An fMRI study of the selective activation of human extrastriate form vision areas by radial and concentric gratings. Current Biology: CB, 10, 1455–1458. [CrossRef] [PubMed]
Wilson H. R. (2003). Computational evidence for a rivalry hierarchy in vision. Proceedings of the National Academy of Sciences of the United States of America, 100, 14499–14503. [CrossRef] [PubMed]
Figure 1
 
Data reproduced approximately from Chen and He (2004) and Pearson and Clifford (2004), Experiment 1. Results are plotted as eye signal stabilization instead of percept stabilization. Average chance of eye signal stabilization across subjects as a function of image swap is shown.
Figure 1
 
Data reproduced approximately from Chen and He (2004) and Pearson and Clifford (2004), Experiment 1. Results are plotted as eye signal stabilization instead of percept stabilization. Average chance of eye signal stabilization across subjects as a function of image swap is shown.
Figure 2
 
Binocular rivalry stimuli. Stimuli were dichoptically presented to the eyes of the participant using a mirror stereoscope. Stimuli were flickering at a rate between 6 and 15 Hz. First row: Grating stimuli used in Experiment 1. Second row: Face/house stimuli used in Experiment 2. Third row: Phase-scrambled face/house stimuli used in Experiment 3.
Figure 2
 
Binocular rivalry stimuli. Stimuli were dichoptically presented to the eyes of the participant using a mirror stereoscope. Stimuli were flickering at a rate between 6 and 15 Hz. First row: Grating stimuli used in Experiment 1. Second row: Face/house stimuli used in Experiment 2. Third row: Phase-scrambled face/house stimuli used in Experiment 3.
Figure 3
 
The four experimental conditions employed in Experiments 13. In condition 1, the stimuli were displayed in the same way on every trial; in condition 2, the flicker frequencies of the stimuli are swapped between eyes on consecutive trials; in condition 3, the images were swapped between eyes on consecutive trials; in condition 4, both flicker frequency and image were swapped between eyes on consecutive trials. In Experiment 3, identical flicker frequencies were used for the two stimuli, thus effectively leaving only two conditions: No image swap and image swap.
Figure 3
 
The four experimental conditions employed in Experiments 13. In condition 1, the stimuli were displayed in the same way on every trial; in condition 2, the flicker frequencies of the stimuli are swapped between eyes on consecutive trials; in condition 3, the images were swapped between eyes on consecutive trials; in condition 4, both flicker frequency and image were swapped between eyes on consecutive trials. In Experiment 3, identical flicker frequencies were used for the two stimuli, thus effectively leaving only two conditions: No image swap and image swap.
Figure 4
 
Average chance of eye signal stabilization across subjects as a function of image swap and frequency swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals. (A) Results of Experiment 1 (rivalry between Gabor patches). (B) Results of Experiment 2 (rivalry between a face and a house).
Figure 4
 
Average chance of eye signal stabilization across subjects as a function of image swap and frequency swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals. (A) Results of Experiment 1 (rivalry between Gabor patches). (B) Results of Experiment 2 (rivalry between a face and a house).
Figure 5
 
Data from Experiment 1. Eye signal stabilization is calculated for every odd and every even pair of trials. Average chance of eye signal stabilization across subjects as a function of the four experimental conditions is shown. Note that although conditions are labeled by what is usually swapped in that condition, nothing is swapped across every other trial in any of the conditions. Contrary to stabilization across every single trial, stabilization is not disrupted across every other trial in any condition, and it is thus unaffected by the intermediate presentation of an image that activates a different subset of monocular neurons. The finding is in line with our proposal that the memory trace is situated in monocular neurons in the early stages of visual processing.
Figure 5
 
Data from Experiment 1. Eye signal stabilization is calculated for every odd and every even pair of trials. Average chance of eye signal stabilization across subjects as a function of the four experimental conditions is shown. Note that although conditions are labeled by what is usually swapped in that condition, nothing is swapped across every other trial in any of the conditions. Contrary to stabilization across every single trial, stabilization is not disrupted across every other trial in any condition, and it is thus unaffected by the intermediate presentation of an image that activates a different subset of monocular neurons. The finding is in line with our proposal that the memory trace is situated in monocular neurons in the early stages of visual processing.
Figure 6
 
Results of Experiment 3 (rivalry between phase-scrambled face/house images). Average chance of eye signal stabilization across subjects as a function of image swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals.
Figure 6
 
Results of Experiment 3 (rivalry between phase-scrambled face/house images). Average chance of eye signal stabilization across subjects as a function of image swap. Eye signal stabilization chance is the chance that the reported percept was displayed to the same eye on two consecutive trials. The horizontal dashed line indicates chance perception. Error bars represent 95% confidence intervals.
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×