Gaps In Perception: How We See A Stable World Through Moving Eyes

We see the most detail in the center of our visual field. This region is called the fovea: it is most densely packed with receptors that turn light into electrical signals that the brain understands. To explore the visual field, our eyes move incessantly to focus on different objects in the scene. These eye movements are called “saccades.” They are very fast, very accurate, and most of the time we make them without any conscious effort at all. Saccades allow us to efficiently collect a lot of detailed information about our surroundings, improving the perceptual understanding we have of our environment.

Although saccades are necessary for understanding the visual scene around us, they also seem to pose a problem for visual perception: The movement of the eyes causes disruptions to the visual input, yet we seem to be almost completely unaware of these disruptions. The scope of the problem can perhaps be best demonstrated from one of the earliest examples: as Dodge (1900; Campbell and Wurtz, 1978) pointed out, if you look from one eye to the other in the mirror, you do not actually see your eyes move, but, if you look at someone else doing the same thing, the movement of their eyes is obvious.

The “problem” can be broken down into three parts. First, we lack awareness of the information our visual receptors receive when our eyes are in motion. Because of the motion of our eyes, our perception should be like a streak of the world zooming past, much like what you perceive when looking out the window of a fast train. Second, not only do we not perceive the streak, but there is also no apparent temporal gap during the saccade. Perception doesn’t seem to strobe but is continuous in time, similar to the immediate cuts in a film. Third, we do not perceive the world to move despite the fact that the position of objects relative to our eyes changes each time we move our eyes. Together, these factors mean that our perception is stable and continuous despite the interruptions caused by saccadic eye movements.

There are a variety of theories that seek to explain how this stable, continuous perception is achieved. One such theory is Backward Masking. Backward Masking is the mechanism behind “subliminal messaging,” which caused a stir in 1957 when it was claimed that the words, “Hungry? Eat popcorn,” were presented so briefly in a cinema commercial that the audience didn’t (consciously) perceive them but were still motivated to go and buy popcorn (Advertising Age, 1957; Key, 1973). The essential principle is that presenting a more prominent image (a “mask”) too quickly after an initial image will prevent proper processing of that initial image (a temporal bottleneck), and the observer will not perceive the initial image at all. This is an obvious mechanism for preventing intra-saccadic perception, as saccades themselves provide the right kind of input: the streaky image on the retina caused by the fast eye-movement is quickly followed by a static, high contrast, more detailed image of the world that would “mask” the feeble intra-saccadic streak.

Evidence for this theory is provided by studies showing that removing the visual input after the saccade causes the observer to see the intra-saccadic streaks: When Campbell and Wurtz (1978) constructed an experimental setup that briefly turned on a light only during their saccades, they could see the streaky image of their own eyes as they looked from one eye to the other in a mirror. Moreover, the spatial and temporal properties of this saccadic masking effect matched what we would expect from Backward Masking when the eye doesn’t move. It was, therefore, suggested that there might not be anything special about saccades; the fact that we are not disturbed by saccades arises naturally from the temporal bottleneck of visual processing. Backward Masking is both necessary and sufficient to eliminate the visual consequences of our own saccades from conscious perception.

But the story is not so simple. Backward Masking has been extensively tested when the eyes are not moving, and it has very specific spatial and temporal conditions: the masking stimulus must overlap spatially with the masked image in a brief time window of 50 milliseconds. To say that Backward Masking can explain the saccade problem, we need to directly compare what happens in saccades to what happens in Backward Masking when the eyes don’t move. We tested the observer’s ability to see intra-saccadic streaks with masks of varying similarity to what works when the eyes don’t move. Although spatially overlapping images shown within 50 milliseconds (most similar to Backward Masking) impaired the perception of intra-saccadic streaks the best, we also found that images that were not overlapping could impair the perception of the streak. Moreover, even removing these images at the end of the saccade (rather than making them appear) also negatively impacted intra-saccadic perception. Such manipulations would not impair perception of a streak when the eyes are still.

The results of our research show that perception during saccadic eye movements is incredibly vulnerable. The fact that the removal of images at the end of the saccade disrupted perception could be best explained by mechanisms of attention rather than masking effects in visual processing. We suggest that there might be a combination of processing bottlenecks. Sudden onsets and offsets in different parts of the visual field may lead to attentional distraction, and promote what appears to be a stable and continuous perception of the world: The visual information that reaches our eyes is neither stable nor continuous, we are just too busy to notice. Understanding how these different mechanisms operate on perception during saccades is essential for understanding our conscious experience of visual perception. We hope that further investigation into these phenomena will elucidate how exactly these effects combine to achieve a stable and continuous percept as our eyes explore the world around us.

These findings are described in the article entitled All is not lost: Post-saccadic contributions to the perceptual omission of intra-saccadic streaks, recently published in the journal Consciousness and Cognition.


  1. “‘Persuaders’ Get Deeply ‘Hidden’ Tool: Subliminal Projection,” Advertising Age 16 (Sept. 1957): 127.
    Key, B. W. (1973). Subliminal Seduction: Ad Media’s Manipulation of a Not So Innocent America. Prentice-Hall, Englewood Cliffs, N.J.
  2. Dodge, R. (1900). Visual perception during eye movement. Psychological Review, 7, 454-465.
  3. Campbell, F. W., & Wurtz, R. H. (1978). Saccadic omission: Why we do not see a grey-out during a saccadic eye movement. Vision Research, 18(10), 1297–1303.