A stationary image can appear to drift, rotate, shimmer, or flow even though
its pixels do not change. “Static-motion illusion” is an umbrella label, not a
single mechanism. This page centers on peripheral-drift and Fraser-Wilcox
patterns, especially Rotating Snakes, while treating Enigma and other effects
as adjacent families whose evidence cannot simply be transferred
(Fraser and Wilcox, 1979;
Faubert and Herbert, 1999;
Troncoso et al., 2008).
Three levels of explanation need to stay distinct. The image contains an
asymmetric order of dark, light, and intermediate values. A temporal driver—such
as onset, retinal slip, microsaccades, blinks, or pupil-linked changes in retinal
illuminance—prevents the pattern from remaining temporally static to the visual
system. Differences in response timing and nonlinear motion detection can then
turn those changing signals into a directional bias
(Backus and Oruç, 2005;
Fermüller et al., 2010;
Otero-Millan et al., 2012;
Bach and Atala-Gérard, 2020).
Current focus
- Peripheral drift and Rotating Snakes mechanism reviews
- Parameter-space work around luminance order, contrast, color, color temperature, and illuminance
- Eye-movement, blink, and qualified pupil-linked accounts
- Implementation bridge into filters and experimental visual tooling
Connected projects
- Brain Candy for induced-vision and pattern-design translation
- Deep Dream for a different class of altered-vision modeling
Synthesis
Static images with temporal consequences
The motion is illusory, but the signal that produces it is structured. The
account below separates the stimulus, its temporal drivers, and the motion
computations that respond to them.
A static image is not static to the visual system
Peripheral-drift and Rotating Snakes patterns can become time-varying retinal
signals even when their pixels do not change. Fixation instability,
microsaccades, saccades, and blinks alter retinal input or coincide with
perceptual episodes. Otero-Millan and colleagues directly related
microsaccades and blinks to illusory rotation in Rotating Snakes
(Otero-Millan et al., 2012).
Pupil dynamics provide another recent account. Mather and Cavanagh related
pupil-linked changes in retinal illuminance to the duration and direction of
peripheral drift. This is important evidence for the tested stimuli, not a
settled universal cause of every stationary-motion illusion
(Mather and Cavanagh, 2025).
Why the pattern matters
A canonical four-part luminance sequence is a useful starting point, but there
is no universal “motion palette.” Local order, intermediate luminance values,
contrast, edge width, spatial scale, eccentricity, chromatic arrangement,
illuminance, and display transfer can all matter. Atala-Gérard and Bach mapped
regions of luminance space that weakened or reversed perceived direction
(Atala-Gérard and Bach, 2017).
Chromatic findings should remain study-specific. Uesaki and colleagues found a
blue-yellow enhancement under their tested conditions, while Nishikawa and
Kitaoka separately tested color temperature and illuminance. Neither result
reduces to the claim that color in general “makes it move”
(Uesaki et al., 2024;
Nishikawa and Kitaoka, 2026).
Observer difference without overreading it
People differ in how strongly they experience these effects, but the evidence
does not justify a personality theory of illusion susceptibility. In one study,
stronger illusory motion was associated with contrast discrimination, not
generic motion sensitivity, self-reported visual discomfort, or migraine
status
(He et al., 2020).
Observer variation should therefore be measured rather than explained through
unsupported trait or personality claims. Fixation behavior,
viewing distance, stimulus size, display conditions, and contrast sensitivity
are concrete variables to record in future comparisons.
Design translation
Apparent motion can be treated as a controllable design variable, but a visual
implementation should preserve measured luminance order after color conversion
and treat edge geometry, carrier scale, eccentricity, brightness, illuminance,
gamma, and color management as experimental parameters. Resampling can destroy
the intended local sequence, and hue labels alone do not guarantee the required
luminance relations.
Direction and strength should not be promised across viewers or devices.
Adjustable intensity, an off switch, and a static preview are conservative
interface choices, not medical-safety thresholds. Canonical imagery should be
linked to its institutional source rather than reproduced without a separate
rights check
(Kitaoka, “Rotating Snakes”).
