A slow peripheral change is easily missed entirely after prolonged fixation on one point

What it is

When gaze holds steady on one point for a long time, a peripheral object that would normally be visible in the corner of the eye gradually "disappears" from awareness — even though it is still physically there, it subjectively feels like empty space. This is Troxler fading. The more practically relevant consequence is: under this same sustained-fixation condition, if the peripheral region undergoes a slow, gradual change — a colour deepening bit by bit, a progress value creeping upward — the change process itself goes completely unnoticed, and users often only "discover" the result has changed after the change has accumulated substantially.

This looks similar to change blindness, but the trigger condition is different: change blindness needs an interruption (blink, saccade, occlusion) to mask the motion signal a change produces. Here no interruption is needed at all — simply holding fixation steady while the change itself is slow enough is sufficient to cause the same failure to see the change. Two independent causes converge on the same outcome: a change that goes unseen.

Why it happens

The visual system gradually reduces its response to constant, unchanging input — neural adaptation, a normal mechanism for improving signal-to-noise ratio: a stimulus that has been present without any new information does not need continued processing resources. Under normal conditions, the eye continually produces tiny, unconscious microsaccades that keep generating small positional shifts on the retina, preventing an image from fully fading away simply because it is perfectly still — which is why peripheral objects are rarely noticed to actually vanish in everyday life. But under deliberately sustained fixation, combined with peripheral stimuli that are low-contrast and poorly bounded, this natural microsaccade compensation is not enough to fully block adaptation, and the peripheral region fades from subjective awareness.

A gradual change is missed more easily because it never produces a motion transient capable of triggering detection in the first place — its rate of change falls below the visual system's detection threshold for motion/change, and the peripheral region is simultaneously undergoing adaptive fading. The two factors compound, so that the entire ramp, from start to finish, never produces a moment with a clear "detection signal."

Studying it

• Classic Troxler fading demonstration: participants fixate a point at the centre of a screen while a low-contrast, poorly bounded pattern (commonly a Gaussian-blurred coloured blob) is presented peripherally; the time needed for the pattern to subjectively "disappear," ranging from a few seconds to tens of seconds, serves as the measure of fading speed.
• Gradual-change detection experiments: for the same total magnitude of change, detection rates are compared between an "abrupt" presentation (a one-step jump) and a "gradual" one (spread across many tiny steps); the gradual presentation typically yields a significantly lower detection rate.
• Typical independent variables: duration of sustained fixation, contrast and boundary sharpness of the peripheral stimulus, total ramp duration and step size, stimulus eccentricity.
• Typical dependent variables: subjectively reported time to fading, detection rate for a gradual change, the proportion of cases where a change is only noticed after it has already completed.
• Methodological caution: the fading effect is highly sensitive to a stimulus's contrast and spatial frequency. A strong effect obtained in the lab with specially designed low-contrast patterns should not be assumed to occur to the same degree with the higher-contrast, sharper-edged elements typical of real interfaces — it should be read as evidence that this direction of risk exists, not as a precise transferable magnitude.

Where it stops holding

• This mainly occurs during tasks that force sustained fixation. Normal browsing involves continual eye movement, and saccades themselves keep refreshing the sampling of peripheral regions, sharply reducing the risk of fading and missed gradual changes. This entry applies most to scenarios that force sustained fixation — video calls, prolonged reading, staring continuously at a monitoring feed.
• The effect is weaker for high-contrast, sharply bounded stimuli. Troxler fading is easiest to observe with low-contrast, blurry-edged targets; a sharply outlined, vividly coloured shape resists fading. The risk concentrates on gradual cues that are already low-salience to begin with.
• This entry describes what happens peripherally while the fovea holds steady fixation; it does not apply to the fovea itself — under normal viewing conditions, the foveal region does not undergo the same kind of fading.
• Its boundary differs from change blindness in an important way: change blindness's remedy is "avoid unmarked abrupt jumps, give changes a transition" — but here the opposite is true: the slow transition itself is the problem. The two entries' design advice points in opposite directions and should not be conflated when applying them.

Applying it

• In scenarios that force sustained fixation (video calls, prolonged reading, continuous monitoring), do not rely solely on a slow gradual change for any status change that needs to be noticed — a gradual transition is itself high-risk in these scenarios, not a "gentler, friendlier" choice.
• For values or states that must change slowly (battery drain, cumulative progress), layer on a periodic discrete marker (a step jump at each threshold, a periodic brief flash) that breaks the continuous ramp into discrete events capable of triggering detection, rather than letting it advance at the same slow rate throughout.
• Do not place a gradual change in a peripheral region and hope the user "will notice eventually." If the information matters, either place it at the expected fixation point or change how it's presented into something that can trigger detection (a discrete jump, accompanied by a brief salient cue).
• How to check: have participants hold fixation on a central task (continuous reading or a video call) while a gradual stimulus at varying rates is embedded peripherally, and measure the actual detection rate and detection delay. If the detection rate at the intended rate of change is clearly below what the use case can tolerate, the current presentation needs to be made discrete or moved to a central position.

Related

• Same group: A1.02.1 The fovea delivers high resolution but covers a tiny angle · A1.02.4 Peripheral flicker forcibly captures attention · A1.02.6 Detail-reading information belongs at the expected fixation point, change cues belong where periphery can reach them
• Nearby: A5.04 Change blindness
• Search terms: Troxler fading · peripheral fading · gradual change blindness · microsaccade