Motion induced blindness

• Yoram Bonneh, Department of Human Biology, University of Haifa, Israel

• Tobias Donner, Center for Neural Science, New York University, NY, USA

Motion-induced blindness (MIB) is a phenomenon in which a small but salient object surrounded by a global moving pattern disappears from visual awareness, only to reappear after several seconds (see Figures 1, 2). The term MIB was coined by Bonneh, Cooperman, and Sagi in 2001 to describe a phenomenon that had been observed in some weak form in the 1960s (Grindley & Townsend, 1967), but largely ignored until rediscovered accidentally with modern computer graphics. MIB is a compelling example of visual disappearance in which salient stimuli disappear from awareness and of multistable perception in which physically invariant stimulation leads to fluctuations in perception. It is also one of a few psychophysical phenomena that have been usefully exploited to study the neural basis of visual awareness; other such phenomena include binocular rivalry, flash suppression and visual masking.

A typical MIB stimulus is demonstrated in Figure 1. It consists of high-contrast static yellow patterns (the "targets") surrounded by a pattern of moving blue dots (the "mask"). With steady fixation, observers commonly report several seconds of spontaneous disappearance of one or more of the targets. The targets typically disappear and reappear independently of one another (see below for exceptions). The perceptual suppression of the target(s) is strikingly immediate and complete (i.e., targets are either visible or invisible in an all-or-none fashion).

Figure 1: A demonstration of MIB with typical stimuli used in the laboratory. To view the effect, fixate on the central white dot without moving the eyes and observe what happens to the yellow dots. The demo allows the following stimulus manipulations: Target position; click and drag any of the yellow targets to any position Target shape; click a target to change its shape (circle, horizontal ellipse, vertical ellipses and a rotating ellipse); note that rotating targets can disappear. You can adjust two patches with orthogonal orientation to induce competition. Target Luminance via slider. Target Size via slider; use to see if large patterns can disappear. Mask speed via slider; try zero speed to test for spontaneous fading and maximal speed for incoherent motion. Mask dot number via slider; use to observe how very few dots can make large patterns disappear.

As long as the basic stimulus arrangement (i.e., small salient target surrounded by large moving flow field) is preserved, MIB is remarkably robust. Most observers experience the phenomenon from the first exposure. Some fail to experience it at the beginning, but most improve with practice. Some (e.g., meditating Buddhist monks, as found by Carter et al., 2005) can get the targets to disappear almost indefinitely.

MIB may also occur for more natural stimuli and perhaps even in everyday life. Richard Brown from the San-Francisco science museum, the Exploratorium, created a version of the phenomenon demonstrated in Figure 2. Objects as salient as a silver watch or keys put on a round transparent table can disappear when a textured plate underneath the table is rotated. Interestingly, a strong disappearance effect occurs immediately after the onset of the rotation movement, showing that a transient stimulus is specifically effective in inducing disappearance. Another “Real-person” demonstration was created by Shinsuke Shimojo at Caltech who used a mirror ball (as used in discotheques) to create an optical flow field on the walls of a large dark room (Shimojo, 2008). Watching someone standing at a peripheral position from a distance of a few meters, in this setting, creates an illusion of a floating figure followed by his/her complete disappearance.

Figure 1: Figure 2: A demonstration of MIB with natural stimuli. Fixate in the center of the screen and observe the different objects disappear and reappear. Note that the two upper-left objects may alternately disappear, showing a form of rivalry.

Under which conditions does MIB occur?

The strength of MIB is typically measured in terms of the proportion of time in which the disappearing target is invisible, as well as the average disappearance period, the frequency of the disappearance events (number of events per minute) and the initial fading time. Disappearance can reach 40% of the time and more, with a typical average invisibility period of 1-2 sec. The duration and frequency of disappearance are determined by the properties of the target and the surrounding moving mask (Bonneh et al., 2001). Some of these stimulus parameters are demonstrated in Figure 1.

• Target size - The size of the disappearing pattern should generally be small, but patterns as large as a thumb at arm's-length distance can also disappear depending on their distance from the fixation point (use the Target Size slider in Figure 1 to explore the effect of size).

• Target position - Targets as close as 1 deg off fixation can easily disappear, but fixated patterns rarely or never disappear (drag one of the yellow patches in Figure 1a close to the central point to explore). Disappearance increases with the distance from the fixation point (eccentricity). At large eccentricity (e.g., above 4 deg), patterns (especially of low contrast) disappear without any mask, a phenomenon known as Troxler fading or Troxler effect. Interestingly, disappearance is not uniform across space. Patterns in the upper-left quadrant of the visual field tend to disappear more for most observers.

• Target movement and flicker - Maximal disappearance occurs when the target is static, but a dynamic target can also disappear (click on one of the targets in Figure 1 three times to obtain a small rotating ellipse and explore disappearance of dynamic patterns). Most strikingly, a slowly moving peripheral target can disappear in one quadrant and reappear in another. Flickering targets disappear especially at fast rates (e.g., 3 Hz flickering dots disappear 25% of the time).

• Spatial layout of targets - When different parts such as small patches form a group, they tend to disappear together (click once on each of two of the patches in Figure 1 and drag them to arrange a collinear configuration in the upper left side). When the parts appear as belonging to different objects, they tend to disappear in alternation (click once more on one of the patches to create an orthogonal arrangement, see also the two adjacent objects in Figure 2). In general, this type of common fate within objects conforms to the Gestalt principles of perceptual organization (Bonneh et al., 2001).

• Target luminance and contrast - For a small bright patch on a black background, the brighter the patch is, the more it disappears (Bonneh et al., 2001; use the Target Luminance slider in Figure 1). This paradoxical finding is in disagreement with the Troxler effect (peripheral fading without a mask, see also below). For contrast patches on a gray background (e.g., as in Figure 3) disappearance rate appears to be invariant to contrast (Bonneh & Cooperman, unpublished data).

• Mask - The type of moving mask used affects disappearance but many types of masks can induce some disappearance as long as the mask surrounds the target. Effective masks typically use coherent motion, a rotating pattern on a 2D surface (e.g., as in Figure 1a, 1b), and a random dot 3D sphere (also used in the original study by Bonneh et al., 2001). A drifting pattern along one dimension (e.g., from left to right), random incoherent motion, and local surrounding flicker (Kawabe & Miura, 2007) can also induce some disappearance, especially for more peripheral targets. However, such masks are less effective. The number of dots in the mask need not be large (use the Mask dots num slider in Figure 1a to reduce the number of dots to minimum). The mask contrast could also be very low as long as the mask is visible.

• Target-mask relation - The mask need not be close to the target in order to induce disappearance. The target could be surrounded by an empty "protection zone" (as in Figure 1) of few degrees and readily disappear (Bonneh et al., 2001). However, the mask should surround or partially surround the target. The depth ordering of the mask and target has a significant effect on disappearance. More disappearance occurs when the mask is presented in front of the target, such as via binocular stereo disparity, and is thus interpreted as occluding it (Graf et al., 2002).

Figure 2: Figure 3: An example of a MIB display used to study the effect of microsaccades on MIB. The time course of fixational eye movements of one observer, including drift and microsaccades is superimposed on the animated stimuli. Note the relatively large eye movements and the abrupt horizontal jumps which prevent any image stabilization.

Is MIB caused by retinal stabilization?

Images stabilized on the retina are known to disappear in a way similar to that observed in MIB (Ditchburn & Ginsborg, 1952). This is assumed to occur via adaptation. It is possible that the moving mask in MIB alters or reduces fixational eye movements causing image stabilization. However, a careful investigation revealed that this is not so (Bonneh et al., 2010). First, the pattern of microsaccades, which are the main component of fixational eye movements, does not change significantly in the presence of the rotating mask used in MIB. Second, microsaccades, like any other abrupt visual transient, tend to trigger reappearance and may even interfere with disappearance (the disappearance is more likely when microsaccades are absent). However, crucially, microsaccades are neither necessary nor sufficient to account for MIB, but have a modulatory effect. In addition, it was found that the rate of microsaccades was modulated by illusory perceptual events of disappearance and reappearance in a way similar to its response to physical removal and onset. This demonstrates that the oculomotor system is informed about visibility in MIB. It remains to be established whether the same principles hold for the effect of microsaccades in other visual disappearance phenomena (see below). See Figure 3 for an animated illustration of the nature and magnitude of fixational eye movements during MIB.

Does MIB depend on attention?

One early hypothesis is that MIB might reflect spontaneous attention shifts, the dynamics of which are induced by the moving mask (Bonneh et al., 2001). As discussed below (see Neural basis of MIB), this hypothesis is consistent with fMRI responses measured in dorsal and ventral extrastriate cortex during MIB (Donner, Sagi, Bonneh, & Heeger, 2008). Psychophysical measurements suggest a complex effect of attention on the dynamics of MIB. Withdrawing attention from both the target and the mask by a demanding task at fixation slows down the bistable process and prolongs the mean invisible time (Scholvinck & Rees, 2009). On the other hand, directing attention to the target increases the probability and duration of its disappearance (Scholvinck & Rees, 2009), analogous to an increase in target contrast (Bonneh et al., 2001). It remains an open question whether MIB requires some residual amount of top-down attention, or whether these perceptual dynamics can unfold independent of top-down attention, as has been observed for other bistable phenomena (Lee, Blake, & Heeger, 2007; Pastukhov & Braun, 2007).

Related visual phenomena

There are a number of other conditions and phenomena, in which visual stimuli are not registered in conscious perception and become invisible (Kim & Blake, 2005). Invisibility of a physical object could be explicit (seeing it disappear) or implicit (failing to notice it) and could refer to its perceptual existence or to its properties and identity. Implicit invisibility is known to occur when stimuli are presented too briefly, or at low contrast, or are embedded in a noisy background or are followed by a masking pattern. In these conditions, unlike MIB, stimuli do not disappear, but their sensory representation is degraded or processing disrupted. Implicit invisibility of stimuli or their properties also occurs when stimuli escape our focus of attention, such as in change blindness, inattentional blindness and the attentional blink.

MIB belongs to a different class of phenomena in which the lack of perception has an explicit nature of "visual disappearance", as if stimuli are erased in front of the observers' eyes. The most prominent among these phenomena are binocular rivalry, in which dissimilar patterns presented to different eyes disappear in alternation, and Troxler fading, in which low-contrast peripheral stimuli disappear under strict fixation. Other phenomena include flash suppression in which disappearance in one eye is induced by a flash presented to the other, images stabilized on the retina that fade away, meta-contrast masking, and perceptual filling-in of "artificial scotomas", in which static peripheral patches disappear when "filled-in" by dynamic surrounding noise (Ramachandran & Gregory, 1991). Of the above, the "artificial scotoma" is most similar to MIB, but there are some notable differences, such as disappearance with "protection zones" without filling-in by the dynamic mask (Figure 1) and the disappearance in alternation of superimposed contour patterns, e.g. ellipses (Bonneh et al., 2001, Figure 3c).

A further important feature of MIB is that the two distinct perceptual states (target visible or invisible) alternate spontaneously, in a seemingly stochastic fashion (following a gamma distribution). This links MIB to other multi-stable perception phenomena, such as binocular rivalry, in which perception alternates spontaneously between distinct states in the face of constant physical stimulation (Blake & Logothetis, 2002).

What is the locus of perceptual suppression in MIB?

At which level of the visual processing hierarchy does the perceptual suppression occur? Much of the current research into MIB focuses on this question as reviewed below. The hope is that the answer will not only provide insights into the nature of MIB, but also reveal some general principles underlying visual awareness.

The explicit nature of the target disappearance and its retinotopic specificity (multiple targets at different locations disappear in an uncorrelated fashion, see Figure 1), suggest that MIB involves early, local sensory mechanisms, such as those in primary visual cortex or even in the retina. This idea of local low-level suppression is further supported by the evidence for the involvement of adaptation (Gorea & Caetta, 2009), filling-in (Hsu, Yeh, & Kramer, 2006), and motion streak suppression (Wallis & Arnold, 2009) in MIB. However, for several reasons, explanations based entirely on early perceptual suppression do not seem sufficient. First, a slowly moving target can also disappear and, after a few seconds, reappear at a different location (even in the other visual hemifield!). This argues against adaptation of localized sensory neuronal populations being the sole mechanism at play. Second, as noted above, the moving mask can be quite far (a few degrees of visual angle) from the target and still induce target disappearance. This rules out a local masking mechanism, based on lateral connections in the retina. Third, the brighter the target, the more it disappears, which rules out a simple "gain-control" mechanism (discarding low-contrast stimuli in a cluttered environment) as the basis of MIB (Bonneh et al., 2001). Note that the latter property distinguishes MIB qualitatively from Troxler fading discussed above (Livingstone & Hubel, 1987).

Another line of evidence, which focuses on the fate of the invisible stimuli, further suggests that MIB does not only result from the suppression of low-level sensory representations. Stimuli rendered invisible by MIB can still produce negative afterimages (Hofstoetter, Koch, & Kiper, 2004) and orientation-selective adaptation (Montaser-Kouhsari, Moradi, Zandvakili, & Esteky, 2004). These results suggest that a cortical representation of the invisible target persists in the retina and primary visual cortex during MIB. Targets that perceptually group into coherent wholes or Gestalts, such as collinear line segments or proximal dots, tend to disappear and reappear in a coherent fashion (Bonneh et al., 2001). Further, when the perceptual grouping configuration between targets is changed while they are invisible, their reappearance depends on the new configuration: Targets forming a good Gestalt reappear in unity and targets forming a bad Gestalt reappear independently of one another (Mitroff & Scholl, 2005). Again, this suggests that a representation of the invisible stimuli remains available at the level of perceptual grouping during MIB.

Taken together, the points discussed in this section suggest that MIB involves both lower and higher levels of visual processing. Mechanisms at neither of the two levels seem sufficient to explain all properties of MIB. Thus, MIB may critically depend on the interaction between these processing levels.

Neuronal interactions underlying MIB

According to one hypothesis, the spontaneous target disappearance in MIB results from competitive interactions between populations of cortical neurons processing the target and the mask (Donner et al., 2008; Keysers & Perrett, 2002). The segregation between these neuronal populations is particularly pronounced within higher stages of the extrastriate visual cortex. Here, the static target and the moving mask are processed by separate visual pathways, the ventral and dorsal pathways, respectively. During target disappearance, functional magnetic resonance imaging (fMRI) responses to the target decrease in the ventral pathway (such as area V4) and responses to the mask increase in the dorsal pathway (such as areas V5/MT or V3A) (Donner et al., 2008). These opposite modulations in ventral and dorsal pathways do not occur during a physical “replay” of the target disappearance, suggesting that they may play a causal role in MIB.

Studies of MIB and related phenomena like generalized flash suppression have also observed modulations of neuronal population activity in primary visual cortex (V1) during target disappearance. These V1 modulations are evident in the low-frequency range (< 30 Hz) of the local field potential (Maier et al., 2008; Wilke, Logothetis, & Leopold, 2006) as well as in the fMRI response (Donner et al., 2008; Hsieh & Tse, 2009; Maier et al., 2008; Scholvinck & Rees, 2009, 2010), but not in spiking activity (Libedinsky, Savage, & Livingstone, 2009; Maier et al., 2008; Wilke et al., 2006). Similar modulations have been observed in the visual thalamus (Wilke, Mueller, & Leopold, 2009). The V1 modulations during target disappearance are spatially nonspecific, that is, globally expressed throughout the retinotopic map, perhaps reflecting neuromodulatory feedback triggered by the perceptual transition (Donner et al., 2008).

Why do salient targets disappear in MIB?

The phenomenological properties and the neurophysiology of MIB suggest that it involves both high- and low-level visual processes. At the sensory level, the lack of transient stimulation of the targets may cause decay of the response due to adaptation mechanisms. This occurs all the time, even without the surrounding motion, but the system normally knows not to interpret this decay as a physical disappearance and to fill-in the missing representation across time. However, if there is a hint that suggests otherwise, e.g. with a moving pattern that seems to occlude the static pattern, then the system may interpret the sensory decay as a real disappearance. Thus, MIB may be the result of the interaction between sensory level processes and an interpretation or decisional level process.

References

• Blake, R., & Logothetis, N. K. (2002). Visual competition. Nat Rev Neurosci, 3(1), 13-21.
• Bonneh, Y. S., Cooperman, A., & Sagi, D. (2001). Motion-induced blindness in normal observers. Nature, 411(6839), 798-801.
• Bonneh, Y. S., Donner, T. H., Heeger, D. J., & Sagi, D. (In preparation). Motion-induced blindness and Troxler fading: similarities and differences.
• Bonneh, Y. S., Donner, T. H., Sagi, D., Fried, M., Cooperman, A., Heeger, D. J., et al. (2010). Motion-induced blindness and microsaccades: cause and effect. J Vis, 10(14), 22.
• Carter, O. L., Presti, D. E., Callistemon, C., Ungerer, Y., Liu, G. B., & Pettigrew, J. D. (2005). Meditation alters perceptual rivalry in Tibetan Buddhist monks. Curr Biol, 15(11), R412-413.
• Ditchburn, R. W., & Ginsborg, B. L. (1952). Vision with a stabilized retinal image. Nature, 170, 36-37.
• Donner, T. H., Sagi, D., Bonneh, Y. S., & Heeger, D. J. (2008). Opposite neural signatures of motion-induced blindness in human dorsal and ventral visual cortex. J Neurosci, 28(41), 10298-10310.
• Gorea, A., & Caetta, F. (2009). Adaptation and prolonged inhibition as a main cause of motion-induced blindness. J Vis, 9(6), 16 11-17.
• Graf, E. W., Adams, W. J., & Lages, M. (2002). Modulating motion-induced blindness with depth ordering and surface completion. Vision Res, 42(25), 2731-2735.
• Grindley, G. C., & Townsend, V. (1967). Further experiments on movement masking. The quarterly journal of experimental psychology, 18, 319-326.
• Hofstoetter, C., Koch, C., & Kiper, D. C. (2004). Motion-induced blindness does not affect the formation of negative afterimages. Conscious Cogn, 13(4), 691-708.
• Hsieh, P. J., & Tse, P. U. (2009). Microsaccade rate varies with subjective visibility during motion-induced blindness. PLoS One, 4(4), e5163.
• Hsu, L. C., Yeh, S. L., & Kramer, P. (2006). A common mechanism for perceptual filling-in and motion-induced blindness. Vision Res, 46(12), 1973-1981.
• Kawabe, T., & Miura, K. (2007). Subjective disappearance of a target by flickering flankers. Vision Res, 47(7), 913-918.
• Keysers, C., & Perrett, D. I. (2002). Visual masking and RSVP reveal neural competition. Trends Cogn Sci, 6(3), 120-125.
• Kim, C. Y., & Blake, R. (2005). Psychophysical magic: rendering the visible 'invisible'. Trends Cogn Sci, 9(8), 381-388.
• Lee, S. H., Blake, R., & Heeger, D. J. (2007). Hierarchy of cortical responses underlying binocular rivalry. Nat Neurosci, 10(8), 1048-1054.
• Libedinsky, C., Savage, T., & Livingstone, M. (2009). Perceptual and physiological evidence for a role for early visual areas in motion-induced blindness. J Vis, 9(1), 14 11-10.
• Livingstone, M. S., & Hubel, D. H. (1987). Psychophysical evidence for separate channels for the perception of form, color, movement, and depth. Journal of Neuroscience, 7(11), 3416-3468.
• Maier, A., Wilke, M., Aura, C., Zhu, C., Ye, F. Q., & Leopold, D. A. (2008). Divergence of fMRI and neural signals in V1 during perceptual suppression in the awake monkey. Nat Neurosci, 11(10), 1193-1200.
• Mitroff, S. R., & Scholl, B. J. (2005). Forming and updating object representations without awareness: evidence from motion-induced blindness. Vision Res, 45(8), 961-967.
• Montaser-Kouhsari, L., Moradi, F., Zandvakili, A., & Esteky, H. (2004). Orientation-selective adaptation during motion-induced blindness. Perception, 33(2), 249-254.
• Pastukhov, A., & Braun, J. (2007). Perceptual reversals need no prompting by attention. J Vis, 7(10), 5 1-17.
• Ramachandran, V. S., & Gregory, R. L. (1991). Perceptual filling in of artificially induced scotomas in human vision. Nature, 350(6320), 699-702.
• Scholvinck, M. L., & Rees, G. (2009). Attentional influences on the dynamics of motion-induced blindness. J Vis, 9(1), 38 31-39.
• Scholvinck, M. L., & Rees, G. (2010). Neural Correlates of Motion-induced Blindness in the Human Brain. J Cogn Neurosci.
• Shimojo, S. (2008). Self and world: large scale installations at science museums. Spat Vis, 21(3-5), 337-346.
• Wallis, T. S., & Arnold, D. H. (2009). Motion-induced blindness and motion streak suppression. Curr Biol, 19(4), 325-329.
• Wilke, M., Logothetis, N. K., & Leopold, D. A. (2006). Local field potential reflects perceptual suppression in monkey visual cortex. Proc Natl Acad Sci U S A, 103(46), 17507-17512.
• Wilke, M., Mueller, K. M., & Leopold, D. A. (2009). Neural activity in the visual thalamus reflects perceptual suppression. Proc Natl Acad Sci U S A, 106(23), 9465-9470.

Internal references

• Lawrence M. Ward (2008) Attention. Scholarpedia, 3(10):1538.
• Naotsugu Tsuchiya and Christof Koch (2008) Attention and consciousness. Scholarpedia, 3(5):4173.
• Keith Rayner and Monica Castelhano (2007) Eye movements. Scholarpedia, 2(10):3649.
• Naotsugu Tsuchiya (2008) Flash suppression. Scholarpedia, 3(2):5640.
• Seiji Ogawa and Yul-Wan Sung (2007) Functional magnetic resonance imaging. Scholarpedia, 2(10):3105.
• Florian Mormann and Christof Koch (2007) Neural correlates of consciousness. Scholarpedia, 2(12):1740.
• Peter Jonas and Gyorgy Buzsaki (2007) Neural inhibition. Scholarpedia, 2(9):3286.
• Bruno G. Breitmeyer and Haluk Ogmen (2007) Visual masking. Scholarpedia, 2(7):3330.

Recommended reading

• Bonneh, Y. S., Cooperman, A., & Sagi, D. (2001). Motion-induced blindness in normal observers. Nature, 411(6839), 798-801.
• Kim, C.Y. & Blake, R. (2005) Psychophysical magic: rendering the visible 'invisible'. Trends in Cognitive Science 9: 381-388
• Donner, T. H., Sagi, D., Bonneh, Y. S., & Heeger, D. J. (2008). Opposite neural signatures of motion-induced blindness in human dorsal and ventral visual cortex. J Neurosci, 28(41), 10298-10310.

External links

• Dr. Yoram Bonneh's home page
• Dr. Tobias Donner's home page

See also

Attention and consciousness, Binocular Rivalry, Flash suppression, Neural correlates of consciousness, Self-organization of brain function, Visual illusions: An empirical explanation, Visual masking

Sponsored by: Eugene M. Izhikevich, Editor-in-Chief of Scholarpedia, the peer-reviewed open-access encyclopedia

Sponsored by: Naotsugu Tsuchiya, California Institute of Technology, Pasadena, CA, USA

Reviewed by: Anonymous

Reviewed by: Su-Ling Yeh, National Taiwan University

Accepted on: 2011-06-08 22:34:51 GMT