Adaptation to transients disrupts spatial coherence in binocular rivalry

When the two eyes are presented with incompatible images, the visual system fails to create a single, fused, coherent percept. Instead, it creates an ongoing alternation between each eye’s image; a phenomenon dubbed binocular rivalry (BR). Such alternations in awareness are separated by brief, intermediate states during which a spatially mixed (incoherent) pattern of both images is perceived. A recent study proposed that the precedence of mixed percepts positively correlates with the degree of adaptation to conflict between the eyes. However, it neglected the role of visual transients, which covaried with the degree of conflict in the stimulus design. We here study whether the presence of visual transients drive adaptation to interocular conflict and explain incidence rates of spatially incoherent BR. Across three experiments we created several adaptation conditions in which we systematically varied the frequency of transients and the degree of conflict between the eyes. Transients consisted of grating orientation reversals, blanks, and plaids. The results showed that the pattern of variations in the fractions mixed percepts across conditions was best explained by variations in the frequency of visual transients, rather than the degree of conflict between the eyes. We propose that the prolonged presentation of transients to both eyes evokes a chain of events consisting of (1) the exogenous allocation of attention to both images, (2) the increase in perceptual dominance of both rivalling images, (3) the speed up of adaptation of interocular suppression, and eventually (4) the facilitation of mixed perception during BR after adaptation. Author summary When one eye is presented with an image that is distinct from the image presented to the other eye, the eyes start to rival and suppress each other’s image. Binocular rivalry leads to perceptual alternations between the images of each eye, during which only one of the images is perceived at a time. However, when the eyes exert weak and shallow mutual suppression, observers tend to perceive both images intermixed more often. Here we designed an experiment and a model to investigate how stereoscopic stimuli can be designed to alter the degree of interocular suppression. We find that prolonged and repeated observations of strong visual transients, such as sudden changes in contrast, can facilitate the adaptation to suppression between the eyes, resulting in that observers report more mixed percepts. This novel finding is relevant to virtual- and augmented reality for which it is crucial to design stereoscopic environments in which binocular rivalry is limited.


73
When the two images are distinct, the visual system is unable to fuse them into a coherent percept.

74
Instead, the distinct mental representations of both eyes compete for priority to visual awareness. This 75 results in the perception of unending perceptual alternations between the two images over time, a purely 76 internally (mentally) driven process because the physical environment is kept stable.

77
BR has been heavily exploited by psychologists, neuroscientists, and philosophers for a variety of 78 reasons. One reason is that the dynamic properties of BR provide information on what type of images 79 dominate more strongly or break into visual awareness faster (e.g., 1, 2). Such research is necessary in 80 order to understand why people sometimes fail to notice objects (e.g., in traffic), how image-parts are 81 grouped into ensemble objects (i.e., Gestalt principles), and why certain objects in the visual 82 environment receive sensory priority (e.g., advertisements). BR is also the primary method used to study 83 the interaction between the sensory processing of stimulus properties and other cognitive high-level 84 functions such as attention, numerosity and emotions (3-7). Furthermore, studies have revealed that a 85 variety of brain regions and processes underlie changes in the content of visual awareness during BR 86 (8-11). Using BR to find the neural loci of consciousness and to identify the distinct processing stages 87 of the stream of consciousness remains an ongoing line of research. Lastly, BR serves as a tool to 88 examine to what degree information, that falls outside the scope of awareness, is processed and affects 89 behavior (e.g., 12 adaptation condition did not (see Figure 1, a modification of Figure 6    We first aimed to test whether the adaptation type in the preceding adaptation phase affected the spatial 296 stability of rivalry in the test phase. Indeed, the fraction mixed percepts during rivalry significantly 297 varied across adaptation types (Figure 2d; repeated measures ANOVA: F(4,25) = 11.17, p < .001).

298
Qualitative inspection of the pattern of results suggested that the original conflict adaptation condition 299 produced the highest fraction mixed percepts while the conditions with a plaid produced the lowest 300 fraction.   and right panel in Figure 3a), as well as two novel conditions for which the contrast of tilted gratings 357 were set at 50% and 25% (see second and third panel in Figure 3a). These two conditions specifically 358 affected the degree of perceptual orientation and monocular contrast transients (see dotted and dashed 359 lines in Figure 3b) and, based on the findings in Experiment 1, we predict that the decrease in contrast 360 should weaken adaptation and decrease the fraction mixed percepts (Figure 3c).

508
In line with this knowledge and an initial proposal (35), they suggested that the sudden disappearance 509 of an image during binocular rivalry activates mnemonic processes dedicated to hold the previously seen 510 image in memory and prioritize it for visual awareness the moment it reappears. This memory process 511 is not restrained to only the most recent image but likely holds and biases perception based on images 512 that are observed for at least the last sixty seconds (51). As an image is prioritized, it will also exert 513 stronger suppression to the rivalling image. As the case in the current study, when both images are 514 subject to transients, both will be prioritized and will mutually inhibit each other, that is strengthen 515 interocular suppression and proliferate its adaptation.

516
It is not unlikely that the effects of working memory and attention on interocular suppression interact.

517
The sudden aspect of transients may (involuntarily and unconsciously) both draw attention and 518 strengthen the (mnemonic) representations of previously seen images, therewith enhancing their 519 inhibitory influence on competing images. However, neither explanation requires adaptation of a 520 specialized conflict detection mechanism. In the model put forth in Said & Heeger, this mechanism is 521 based on the idea of ocular opponency neurons (34, 52, 53). Although such neurons appear likely 522 candidates for involvement in binocular rivalry, and the initial prediction of the model by Said & Heeger 523 that included a conflict detection mechanism explained their data well, the results reported here cannot 524 be unified under that model. As such, we currently see no evidence that mechanisms based on ocular 525 opponency neurons should be included in models of binocular rivalry.

526
It is important to note that in our study the intermittent presentation of blanks had a stronger effect on 527 adaptation than the intermittent presentation of plaids. A similar effect has been reported before (54),

528
showing that the presentation of interleaved blanks enhanced the temporal stabilization of rivalry more 529 than plaids. As blanks are more distinct from the orthogonal images and therefore more conspicuous, it 530 makes sense that intermittent presentation of blanks adapted interocular suppression stronger than 531 plaids. This conclusion may appear at odds with our observation that the monocular contrast transients 532 (i.e., blanks) disrupted the spatial coherence of rivalry slightly weaker than perceptual orientation 533 transients. Note however that the monocular contrast transients were not visible but the perceptual 534 orientation transients were visible to the observer. As the visibility of transients is positively linked to 535 the degree of drawing attention exogenously (55) and the suppressive strength of an evoked traveling 536 dominance wave (26), it is not unexpected that the perceptually visible orientation transients adapted 537 interocular suppression most. Our observation that a relatively high rate of orientation transients (e.g., 538 see rivalry condition) increased the fraction mixed percepts more than a relatively low rate (e.g., see