Foundations of Mesmerizing Visual Experiences: Neurobiology of Motion Perception and Optic Flow

Full report

On this page

The Neurobiology of Motion Detection and Adaptation

The human visual system is exquisitely tuned to detect, process, and interpret motion. The foundational mechanisms of motion perception rely on a hierarchical cascade of neural processing that begins in the retina, projects through the lateral geniculate nucleus to the primary visual cortex (V1), and culminates in higher-order extrastriate areas such as the middle temporal (MT/V5) area and the medial superior temporal (MST) area. To construct mesmerizing, continuous visual experiences in graphics research, it is essential to leverage the biological architecture of these motion pathways, particularly the phenomena of motion energy, adaptation, and the resulting perceptual aftereffects.

The Motion Energy Model and V1 Processing

The processing of visual motion begins with the extraction of local spatiotemporal signals. The widely accepted "motion energy model" describes how direction-selective simple and complex cells in V1 compute local motion1. Simple cells can be modeled as linear spatiotemporal filters. According to the Adelson and Bergen formulation, these cells possess receptive fields tuned to specific spatial and temporal frequencies, operating as quadrature pairs that are phase-shifted by 90 degrees2. To achieve phase invariance—whereby the neuron responds to motion regardless of the precise spatial phase of the stimulus within its receptive field—the outputs of these linear quadrature filters are squared and summed, forming the characteristic response of a complex cell1.

Mathematically, the local motion energy [Figure omitted from source export] can be expressed by combining the squared outputs of even and odd symmetric filters:

[Figure omitted from source export]

where [Figure omitted from source export] and [Figure omitted from source export] are the responses of the quadrature pair. However, V1 neurons suffer profoundly from the "aperture problem"; because of their exceptionally small receptive fields (often less than one degree in diameter), they can only detect the component of motion that is perpendicular to the orientation of a moving edge1. They cannot discern the true global trajectory of an object moving through space.

Global Motion Pooling and Divisive Normalization in Area MT/V5

To resolve the aperture problem and determine the true two-dimensional velocity of an object, local motion signals from V1 are projected to and pooled in area MT (V5)5. MT contains specialized "pattern cells" that integrate the outputs of numerous V1 "component cells" that have differing spatial orientations and direction preferences6. This integration is not merely a linear addition; it involves highly nonlinear computations, specifically divisive normalization1. In divisive normalization, the excitatory response of a specific direction-selective neuron is divided by the summed activity of a broader pool of surrounding neurons. This neural architecture effectively provides contrast gain control and ensures that the neural response scales appropriately across widely varying stimulus intensities, maintaining the fidelity of the motion signal whether the stimulus is a high-contrast grating or a low-contrast fog1.

Area MT/V5 is also the primary locus of motion opponency in the human brain. Opponent motion channels act as critical filters that attenuate flicker or stationary noise while allowing true directional motion signals to propagate up the visual hierarchy5. When a visual stimulus contains perfectly balanced, counter-phase motion energy—such as two identical dot patterns moving in exactly opposite directions—MT neurons exhibit a muted response. In fact, functional magnetic resonance imaging (fMRI) studies utilizing multivariate pattern analysis demonstrate that the blood oxygen level-dependent (BOLD) signal in MT/V5 during counter-phase stimulation is significantly reduced, often indistinguishable from the cortical response to random visual flicker5. This opponent architecture is essential for isolating coherent motion layers.

The Waterfall Illusion, Opponent Channels, and Adaptation Dynamics

Continuous exposure to unidirectional motion results in a progressive, time-dependent reduction in the firing rate of the direction-selective neurons tuned to that specific motion vector. This neural fatigue, widely referred to as adaptation, profoundly alters the resting balance of activity within the opponent motion channels8. When the physical motion ceases and a stationary scene is subsequently viewed, the unadapted neurons—those tuned to the opposite direction of the initial stimulus—exhibit a higher baseline firing rate relative to the newly fatigued adapted neurons. The brain's downstream decision-making areas interpret this temporary firing rate imbalance as actual physical motion in the reverse direction. This phenomenon is classically known as the motion aftereffect (MAE) or the Waterfall Illusion8.

The dynamics of this adaptation are mathematically predictable. The decay of the MAE over time is generally modeled using an exponential decay function or a power law, where the strength of the illusory motion dissipates gradually as the adapted neurons recover their baseline metabolic state11. The psychometric function governing this shift demonstrates that past stimuli continuously alter the point of subjective equality in motion detection11. Extended adaptation periods, scaling from mere minutes to hours, lead to prolonged aftereffects, suggesting that adaptation mechanisms operate across multiple temporal scales to calibrate the visual system to the prevailing environmental statistics13. Furthermore, fMRI studies demonstrate that the perceived illusory motion during an MAE strongly correlates with persistent BOLD signal modulations in MT+ and V3A5.

Understanding this neurobiological decay rate allows graphics researchers and visual designers to pace mesmerizing visual experiences intelligently. By utilizing prolonged unidirectional flows, a designer can build up a substantial adaptation "charge" in the viewer's visual cortex. When this charge is subsequently released by transitioning the scene to a static or significantly slower-moving state, the stationary environment will appear to drift, melt, and breathe, creating a profound hypnotic effect without requiring any actual rendering of movement on the screen.

Optic-Flow Geometries and the Perception of Heading

Optic flow is the complex, dynamic pattern of apparent motion across the visual field generated by the relative motion between an observer and the environment. Higher-order processing of optic flow occurs prominently in the dorsal subregion of the medial superior temporal area (MSTd)14. Neurons in MSTd have exceptionally large receptive fields—often spanning the entire visual field—and are tuned not to simple linear translation, but to complex geometrical transformations in a three-dimensional spiral space16. These specialized neurons provide the neurological foundation for "vection," which is the compelling psychological illusion of self-motion in a physically stationary observer19.

Forward Tunnel Travel, Looming, and Radial Expansion

Forward tunnel travel and looming visuals rely on radial expansion. This geometry simulates forward travel through an environment, creating a singular mathematical focus of expansion (FOE) from which all local motion vectors radiate outward across the retina. This geometry accurately mimics the visual input of walking, running, or driving forward18. In area MSTd, specific populations of neurons possess tuning curves that peak strongly when the FOE aligns with their preferred spatial coordinates15. As objects in a simulated tunnel expand and approach the periphery of the camera's field of view, their angular velocity increases exponentially, a cue the visual system heavily relies upon to estimate the time-to-collision for approaching objects17.

Backward Travel and Radial Contraction

Backward travel simulates the visual experience of retreating, characterized by radial contraction where all flow vectors converge upon a central focus of contraction. Interestingly, the psychophysical onset latency of vection differs dramatically between these two opposing radial geometries. Empirical evidence demonstrates that radial contraction consistently produces significantly shorter vection onset latencies and stronger perceived vection magnitudes than radial expansion across all age groups19. This pronounced asymmetry likely arises from an evolutionary or exposure-based neural expectancy: because forward locomotion is the default state of human movement, the visual-vestibular networks are highly habituated to radial expansion19. Consequently, radial contraction is treated as a highly salient, novel sensory event, triggering a rapid, urgent illusion of backward self-motion that easily overpowers the vestibular system's static signals.

Rotation, Spiral Flow, and Mixed Rotational/Radial Flow

Rotational flow fields, generated by rotating the visual field around the z-axis (the line of sight), stimulate the perception of roll or pitch self-rotation. Visual rotation can induce compelling roll vection, effectively tricking the brain into perceiving a dramatic shift in the gravitational vertical20. When radial expansion or contraction is mathematically combined with rotation, the result is spiral flow16.

The brain does not process spiral flow merely by linearly adding the outputs of separate expansion and rotation detectors. Instead, MSTd contains specific neuronal subpopulations tuned distinctly to intermediate spiral motions, representing a continuous mapping of optic flow space16. Graziano et al. discovered that the majority of neurons in MSTd span a continuum of preferred responses in this optic flow space, treating spiral motion as a singular, unified perceptual component rather than a composite of distinct orthogonal vectors24. Thus, mixed rotational and radial flows create an overwhelmingly immersive visual experience because they maximally stimulate these highly specialized complex motion detectors.

Helical Flow, Translation, Shear, and Oscillating Depth

Helical flow introduces a lateral translational drift to a spiraling tunnel. By adding translation to a radial-rotational mix, helical flow stimulates both MT linear detectors and MSTd complex detectors simultaneously. This geometry creates a highly disorienting but mesmerizing sense of traversing a continuously curving pathway, pushing the boundaries of the viewer's spatial mapping capabilities.

Translational flow involves uniform unidirectional movement, mimicking the experience of looking out the side window of a moving train. While it produces strong local motion energy in V1 and MT, its capacity to induce robust vection is highly dependent on covering the peripheral retina25. Shear flow occurs when adjacent spatial regions move at different speeds or in strictly opposite directions, creating sharp motion boundaries. This configuration strongly activates the center-surround opponent mechanisms in MT, often leading to motion transparency or localized directional repulsion26.

Oscillating depth represents an environment where the optic flow alternates smoothly and continuously between expansion and contraction. Because visual adaptation scales proportionately with the duration of unidirectional exposure, oscillating flows constantly reset the adaptation curve. This prevents the visual system from fully fatiguing to a single direction, sustaining continuous neural engagement and preventing the accumulation of a static motion aftereffect13.

Determinants of Perceived Intensity and Vection

The psychological and physiological intensity of a moving visual experience—specifically its ability to induce vection—is modulated by several core psychophysical parameters: visual speed, angular velocity, depth density, luminance contrast, peripheral coverage, the camera's field of view (FOV), and the focal point.

Peripheral Coverage and the Peripheral Dominance Hypothesis

Early theories of self-motion perception posited the "peripheral dominance hypothesis," which suggested that the peripheral retina (typically defined as vision outside a 30-degree central diameter) is exclusively specialized for driving vection and postural sway25. Contemporary research, however, supports "retinal invariance," indicating that central vision can also induce strong vection if the depth structure and optic flow patterns are appropriate25. Nevertheless, because radial flow vectors naturally achieve their highest angular velocities in the periphery as objects pass the observer, maximizing the peripheral FOV remains one of the most effective methods to increase perceived intensity and shorten the vection onset latency30. Immersive camera setups with an FOV extending beyond 100 degrees diagonal will exponentially increase the strength of the visual experience.

Visual Speed, Angular Velocity, and Depth Density

Vection strength increases monotonically with the luminance contrast and spatial frequency of the moving pattern32. High-contrast edges provide sharp, unambiguous inputs to the spatiotemporal quadrature filters in V1, resulting in robust motion energy that easily cascades into MT33. Furthermore, depth density—the number of distinct moving elements per unit area—modulates both the perception of speed and the magnitude of visually induced postural sway. An increase in the density of elements perceived to be in the background relative to the foreground maximizes the strength of visually induced self-motion. The brain preferentially anchors its self-motion estimates to the densest, most distant available visual plane, treating it as the stable environmental reference frame34.

Focal Point, Smooth Pursuit, and Extraretinal Signals

The perceived intensity of an optic flow field is highly sensitive to the viewer's focal point and subsequent eye movements. When an observer tracks a moving object across the screen using smooth pursuit eye movements, the rotation of the eye induces a laminar motion across the entire retina. This extraretinal slip fundamentally disrupts the true focus of expansion21. However, the brain compensates for this via the vestibulo-ocular reflex (VOR) and efference copy signals. Neurons in MSTd receive an extraretinal signal indicating the speed and direction of the smooth pursuit, allowing them to shift their tuning curves mathematically to recover the true heading direction in world coordinates22. Notably, this internal compensation mechanism exhibits a latency; the eye movement signal has been found to lag the retinal signal by approximately 55 to 100 milliseconds35. In visual design, forcing the user to constantly shift their focal point (saccades) or track specific elements traversing an optic flow field reduces adaptation to unidirectional motion and actively engages these extraretinal compensation networks, significantly enhancing the perceived intensity and duration of vection38.

Coherence, Transparency, and Competing Flow Fields

When a visual scene contains multiple layers of motion—such as two overlapping dot fields or textures moving in different directions—the visual system is forced to solve a complex binding problem. It may either integrate the vectors into a single coherent motion or segregate them into transparent, distinct layers27.

Motion Transparency and Divisive Normalization

Motion transparency is heavily dependent on the spatial frequency of the stimuli, divisive normalization, and population coding in area MT33. When two distinct flow fields are superimposed, the resulting perception relies on how the neural population resolves the conflicting energy. If the spatial frequencies of the two layers are sufficiently distinct, or if they move at broad angular separations, MT neurons maintain a bimodal population response, allowing the viewer to perceive two transparent sheets of motion sliding past one another7.

Direction Repulsion and the Faster-Speed Bias

When transparent layers move at relatively narrow angular separations, a psychophysical phenomenon known as "direction repulsion" occurs27. Because the overlapping populations of direction-selective neurons in MT mutually inhibit each other through recurrent connections, the peak responses of the neural populations are pushed apart. Consequently, the viewer perceives an exaggerated angular difference between the two vectors, experiencing a profound sense of spatial separation41.

Furthermore, when multiple motion layers compete, they rarely receive equal perceptual weighting. If the speeds of the two competing layers differ, but both are moving at relatively slow speeds (under 20 degrees per second), the visual system exhibits a robust "faster-speed bias"39. Under divisive normalization models, MT neurons disproportionately weight their response toward the faster component, anchoring the perceived global motion to the higher-velocity stimuli33.

Single Dominant Flow vs. Competing Flows

Comparing a single dominant flow field against competing multiple layers reveals distinct perceptual outcomes. A single dominant flow field maximizes spatial coherence, generating rapid and intense vection by providing an unambiguous signal to MSTd. In contrast, introducing multiple competing motion layers reduces global coherence but vastly increases the perceived local kinetic energy. By leveraging the faster-speed bias, visual designers can add a subtle, fast-moving secondary layer of transparent noise or particulate matter to a primary, slow-moving dominant flow field. This technique dramatically increases the perceived overall kinetic energy and depth of the scene without destroying the underlying coherence necessary to induce vection.

Algorithmic Foundations of Endless Movement

To create mesmerizing, hypnotic graphics that simulate endless motion without breaking immersion, rendering architectures must meticulously avoid arbitrary geometry resets, visible textural seams, and floating-point precision degradation over extended viewing periods. In modern real-time graphics pipelines, this is optimally achieved using OpenGL Shading Language (GLSL) fragment shaders, which compute visual output mathematically on a per-pixel basis42.

Time-Based Functions and Modulo Arithmetic

Endless movement strictly relies on an absolute, monotonically increasing time variable (typically defined as iTime). If standard Cartesian geometries are continually translated along an axis using iTime, the spatial coordinates will eventually approach infinity. This causes severe floating-point precision loss, leading to visible visual stuttering and rendering artifacts. To resolve this, continuous modulo arithmetic or fractional wrapping is applied. In GLSL, the fract() function extracts only the fractional part of a floating-point number, effectively wrapping the infinite coordinate space within a strict [Figure omitted from source export] range42. By calculating UV coordinates using fract(uv \+ dir \* iTime), the rendered texture or geometry smoothly scrolls out of one side of the coordinate boundary and seamlessly wraps to the opposite side, maintaining the flawless illusion of infinite translation without mathematical degradation.

Log-Polar Transformations for Infinite Zoom

Simulating infinite forward or backward travel—endless radial expansion or contraction—without relying on discrete, memory-intensive 3D geometry requires sophisticated mathematical manipulation of the 2D coordinate space. The log-polar transformation converts standard Cartesian coordinates [Figure omitted from source export] into a space defined by the logarithm of the radius ([Figure omitted from source export]) and the angle ([Figure omitted from source export])42.

The forward mapping is given by:

[Figure omitted from source export]

[Figure omitted from source export]

In the context of the inverse log-polar map used within a fragment shader, continuous linear translation along the [Figure omitted from source export] axis naturally results in exponential scaling (infinite zoom) when transformed back into Cartesian space. Simultaneously, linear translation along the [Figure omitted from source export] axis results in continuous rotation. By applying fractional wrapping to both the [Figure omitted from source export] and [Figure omitted from source export] coordinates prior to the inverse transformation, the shader produces an infinitely repeating, scalable, and rotatable tunnel effect (classically known as the "Droste effect"). This mathematical approach ensures perfect phase-continuous animation that can run indefinitely without seams42.

Vector Flow Fields and Curl Noise

To simulate chaotic but fluidly continuous motion—such as swirling smoke, atmospheric liquid dynamics, or hypnotic magnetic fields—procedural noise functions are heavily utilized. However, standard Perlin or Simplex noise algorithms often produce visual artifacts, localized sinkholes, or source points where particles congregate or vanish unnaturally.

To circumvent this, graphics researchers utilize "curl noise"45. By taking the mathematical curl of a procedural scalar or vector potential field, the resulting velocity field is physically guaranteed to be divergence-free ([Figure omitted from source export]). When pixels or particles are advected through a curl noise flow field, they will swirl continuously along the vector paths without ever converging into singular points or expanding into empty voids. This ensures a mesmerizing, unbroken kinetic texture that mimics perfect fluid dynamics45.

Circular Splines and Helical Extrusion

For endless three-dimensional travel that is not strictly linear, such as helical or spiral tunnels, developers rely on circular splines combined with raymarching and domain repetition. By defining a mathematical spline that curves through 3D space, the camera's position is evaluated continuously along this path using iTime. To avoid geometry limits, the 3D space is folded using a modulo operator, meaning the camera is effectively looping through the same bounded physical space, but the continuous modification of the local viewing matrix creates the persistent illusion of traveling down an endless, twisting corridor.

Vestibular Mismatch, Simulator Sickness, and Accessibility

While mesmerizing visual experiences that fill the peripheral FOV and induce strong vection are highly sought after, they inherently carry a significant risk of causing visually induced motion sickness (VIMS), cybersickness, or simulator sickness. Addressing this requires a nuanced physiological understanding of the underlying causes and rigorous adherence to accessibility guidelines.

Sensory Conflict Theory vs. Postural Instability Theory

The traditional framework for understanding VIMS is the "Sensory Conflict Theory," initially formalized by Reason in 1978\. This theory posits that motion sickness arises when the sensory signals regarding self-motion and orientation disagree20. In a visually mesmerizing experience, the visual system detects acceleration through optic flow, but the vestibular system (the semicircular canals and otolith organs in the inner ear) accurately reports that the physical body is stationary20. This profound neural mismatch triggers an error signal in the brain's autonomic networks, leading to symptoms such as nausea, sweating, fullness of the head, and disorientation20.

Conversely, the competing "Postural Instability Theory" formulated by Riccio and Stoffregen argues that sickness is not caused by conflict per se, but occurs when an observer lacks the necessary strategies to maintain postural control in a novel or challenging environment49. When exposed to immersive optic flow, observers exhibit subconscious postural sway as they attempt to orient themselves. This sway is measurable as micro-shifts in their Center of Pressure (CoP) on a force plate31. A Fourier spectral analysis of this postural sway often reveals distinct low-frequency oscillations (below 0.1 Hz) as the body's musculature attempts to synchronize with the visual motion52. If the visual motion is too chaotic or unpredictable, postural control significantly degrades, leading directly to sickness.

The Simulated Viewpoint Jitter Paradox

A fascinating and counterintuitive exception to classical sensory conflict theory is the phenomenon of simulated viewpoint jitter. Adding high-frequency, random, or oscillating camera shake to a smooth optic flow field should theoretically increase visual-vestibular conflict, as the vestibular system expects to feel the violent shaking but does not. However, empirical psychophysical research demonstrates the exact opposite: adding simulated viewpoint jitter actually decreases vection onset latency, increases vection duration, and creates a much more compelling illusion of self-motion compared to perfectly smooth displays20.

This jitter advantage suggests that the central nervous system may treat perfectly smooth optic flow as an artificial anomaly. In contrast, the presence of high-frequency visual noise validates the self-motion signal, closely mimicking the ecological realities of natural locomotion (such as walking or running, which inherently generate complex six-degree-of-freedom head movements)57. Consequently, adding low-amplitude, high-frequency camera oscillations to an animated scene can dramatically heighten the mesmerizing impact while anchoring the perceptual realism of the flow, potentially mitigating postural instability.

Quantifying Simulator Sickness

To evaluate the physiological impact of these graphics objectively, researchers utilize the Simulator Sickness Questionnaire (SSQ) developed by Kennedy et al. (1993). The SSQ requires participants to rate 16 distinct symptoms, which are then mathematically grouped into three specific factors: Nausea, Oculomotor disturbances (such as eyestrain and difficulty focusing), and Disorientation (such as vertigo and dizziness)58.

The calculations for these specific subscales rely on precise weighting multipliers:

• Nausea \= Sum of specific symptom scores [Figure omitted from source export]
• Oculomotor \= Sum of specific symptom scores [Figure omitted from source export]
• Disorientation \= Sum of specific symptom scores [Figure omitted from source export]
• Total Severity \= (Nausea \+ Oculomotor \+ Disorientation) [Figure omitted from source export]58.

Scenes containing highly unpredictable rotational components or sudden shifts in spatial frequency yield extremely heavy loadings on the Disorientation and Oculomotor axes. Managing these factors through design choices is critical for broad audience accessibility.

Accessibility Thresholds and Controllable Intensity Bands

To prevent severe adverse effects such as photosensitive epileptic seizures, vestibular distress, and extreme cybersickness, visual experiences must adhere to rigorous medical and technical accessibility guidelines. According to the Web Content Accessibility Guidelines (WCAG 2.1), a flash or rapidly alternating visual pattern becomes a potential seizure hazard if it exceeds a frequency of 3 Hz, provided it occupies a solid visual angle of [Figure omitted from source export] steradians and possesses a luminance of [Figure omitted from source export] cd/m²60.

Furthermore, to minimize "motion-to-photon latency" sickness—especially in interactive or VR contexts—systems must process and render updates in under 20 milliseconds. Latencies exceeding this threshold introduce a noticeable visual lag that fundamentally breaks the synchronicity of the visual-vestibular loop, acting as a primary driver of cybersickness61.

To safely accommodate users with varying tolerances for optic flow, visual experiences should be categorized into distinct, controllable intensity bands:

1. Gentle Mode: Eliminates all rotational optic flow and viewpoint jitter to protect the vestibular system. Utilizes slow, low-contrast radial contraction, which induces vection more comfortably than expansion. Keeps temporal frequencies strictly below 1 Hz and avoids high-density focal points.

2. Standard Mode: Incorporates smooth radial expansion, mild translational flow, and moderate depth density. Employs soft, phase-continuous color shifts and limits high-contrast edges to the central FOV to reduce harsh peripheral flicker, ensuring the 3 Hz threshold is never approached.

3. High Mode: Introduces highly complex optic flow geometries, including spiral space combinations and helical paths. Incorporates multiple transparent motion layers, direction repulsion, high-contrast boundaries, and controlled viewpoint jitter to absolutely maximize vection. Operates near the upper bounds of comfortable spatiotemporal energy while strictly obeying accessibility safety thresholds.

Visual-Design Decision Matrix

The following decision matrix provides a rigorous computational and physiological framework for selecting specific optic-flow parameters based on their intended perceptual outcome and neurobiological consequence.

Scene Concepts: Twelve Optic-Flow Environments

Based on the intricate intersection of neurobiology, optic-flow geometry, and procedural shader mathematics, the following twelve scene concepts are explicitly designed to evoke specific, mesmerizing perceptual states.

1. The Breathing Monolith (Oscillating Depth)

This environment centers on a massive, abstract monolithic structure located at the exact center of the FOV. The visual field smoothly oscillates between slow radial expansion and contraction. By utilizing a low-frequency sine wave to modulate the radial velocity over absolute time, the scene actively prevents the accumulation of a static motion aftereffect13. The constant, rhythmic reversal of the motion direction continuously re-engages the opponent motion channels in area MT, causing the geometry to appear as if it is organically drawing breath, providing a deep, hypnotic rhythm.

2. The Endless Descent (Log-Polar Spiral)

This scene simulates an infinitely zooming, slowly rotating tunnel that imparts the visceral sensation of falling down a bottomless well. It is rendered entirely via a 2D fragment shader utilizing the inverse log-polar map42. The radial coordinate ([Figure omitted from source export]) and angular coordinate ([Figure omitted from source export]) are continuously incremented by iTime and seamlessly wrapped using the GLSL fract() function. The resulting spiral flow maximally stimulates the specialized subpopulations of MSTd neurons tuned to intermediate complex motions in spiral space16, creating an unbroken, phase-continuous plunge.

3. The Phantom Drift (Direction Repulsion and Transparency)

A hazy, ethereal field of luminous particles where the high-frequency foreground drifts slowly to the left while a dense, low-frequency background drifts to the right. This scene directly exploits the neurobiology of motion transparency and direction repulsion. The mutual inhibition of V1 and MT neural populations tuned to these opposing directions drastically exaggerates the perceived angular difference and spatial separation between the layers27. A faster-speed bias is applied to the foreground, ensuring the scene feels highly energetic despite the mathematically low absolute pixel velocities39.

4. The Jittering Expanse (Viewpoint Jitter Vection)

The observer experiences a high-speed race across a flat, high-contrast geometric grid extending infinitely to the horizon, strictly simulating forward linear translation. To overcome initial sensory conflict and trigger immediate linear vection, a high-frequency (e.g., 5 Hz), low-amplitude vertical camera oscillation is applied20. This simulated viewpoint jitter validates the visual motion signal for the brain, overriding the vestibular mismatch and heavily anchoring the user in the illusion of high-speed travel, dramatically reducing the vection onset latency56.

5. The Waterfall's Echo (Sustained MAE Induction)

This environment presents a cascading curtain of flowing digital liquid moving strictly downward at a constant velocity for a prolonged period (e.g., 60 seconds), which then suddenly freezes in place. This is specifically engineered to trigger a profound static MAE (the classical Waterfall Illusion)8. The prolonged unidirectional motion deeply fatigues the downward-tuned component cells in V1 and MT. When the texture stops, the unopposed upward-tuned cells fire at their baseline rate, causing the stationary image to appear to flow upwards in a powerful, hallucinatory manner8.

6. The Inverted Universe (Radial Contraction)

An enveloping visual environment where all stars, textures, and geometric elements flow continuously from the extreme periphery toward a central vanishing point. This utilizes pure radial contraction optic flow. Because the human visual system processes visual contraction with a significantly shorter latency and greater overall vection intensity than expansion, this scene immediately induces a profound sense of backward locomotion19. The intense peripheral dominance ensures the observer's posture control systems interpret the flow as an immediate spatial retreat25.

7. Magnetic Swarm (Curl Noise Flow Fields)

Millions of glowing, high-contrast particles swirl in chaotic, liquid-like eddies without ever clustering into masses or dissipating into empty space. The velocity of every individual particle is driven in real-time by a curl noise algorithm executed within a compute shader45. Because the mathematical curl of a potential field is inherently divergence-free ([Figure omitted from source export]), the particles exhibit seamless, organic turbulence that perfectly mimics fluid dynamics, maintaining constant depth density and spatial frequency across the entire FOV indefinitely.

8. The Chromatic Shear (Opponent Motion Fatigue)

The visual field is divided into alternating horizontal bands of high-contrast texture moving in strictly opposite directions (shear flow). This configuration powerfully activates the motion opponent mechanisms in MT/V55. The perfectly balanced counter-phase motion energy cancels out any global directional pooling, making it impossible for the brain to extract a singular heading. After prolonged exposure, transitioning to a static gray screen will result in localized, band-specific aftereffects, where different horizontal strata of the user's vision appear to aggressively shear against one another.

9. The Orbiting Labyrinth (Roll Vection)

The camera is locked inside a massive, highly detailed cylinder that slowly, relentlessly rotates around the z-axis (the viewer's direct line of sight). This employs pure rotational optic flow. The continuous rotation of the peripheral visual field creates severe sensory conflict with the otolith organs, quickly inducing powerful roll vection20. To mitigate extreme SSQ Disorientation scores58, the angular velocity is kept deliberately low, and the central focal point remains entirely static, providing a much-needed anchor for the oculomotor system and suppressing optokinetic nystagmus.

10. The Moiré Singularity (High Spatial Frequency Drift)

Overlapping grids of exceptionally fine, high-contrast lines occupy the visual field, moving at slightly different translational velocities. This scene leverages high spatial frequency and luminance contrast to absolutely maximize local motion energy extraction in V132. The subtle, continuous translation of the grids creates vast, sweeping interference patterns (Moiré effects) that appear to move much faster than the individual lines themselves. This creates a mesmerizing second-order motion illusion that effectively hacks the MT pattern cells, simulating vast macroscopic movements from microscopic shifts.

11. The Phase-Continuous Helix (Helical Flow)

The observer is propelled down an intricate tunnel that is continuously twisting and turning like a coiled spring. This uses circular splines and 3D raymarching with domain repetition to render an infinite, seamless helical tube. The complex geometry perfectly stimulates the MSTd neurons tuned to both translation and rotation simultaneously17. To completely avoid the frame-count reset stutter that plagues endless runners, the camera's forward progression is wrapped mathematically modulo the length of one complete helical coil, creating a flawless, infinite loop.

12. Gentle Awakening (Accessibility Mode)

Designed specifically for the "Gentle" intensity band to accommodate users highly sensitive to cybersickness, this scene features a slow, soft-focus outward expansion of low-contrast pastel gradients, entirely devoid of sharp edges or rotational vectors. By purposefully eliminating high spatial frequencies, keeping temporal frequencies strictly below 1 Hz, and utilizing soft radial expansion rather than high-velocity rotation, this environment completely avoids stimulating the SSQ Disorientation and Oculomotor factors58. It provides a soothing, meditative sense of forward floating without triggering the severe vestibular mismatch inherent in sharp-edged, high-velocity graphics.

Works cited

1. A Model of Neuronal Responses in Visual Area MT, https://www.cns.nyu.edu/pub/lcv/simoncelli96.pdf

2. A spatiotemporal energy model based on spiking neurons for ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC11297886/

3. Nonlinear circuits for naturalistic visual motion estimation \- eLife, https://elifesciences.org/articles/09123

4. Understanding Vision: Theory, Models, and Data \1, [https://dokumen.pub/understanding-vision-theory-models-and-data-1nbsped-0199564663-9780199564668.html

5. (PDF) Motion opponency examined throughout visual cortex with, https://www.researchgate.net/publication/344130420\_Motion\_opponency\_examined\_throughout\_visual\_cortex\_with\_multivariate\_pattern\_analysis\_of\_fMRI\_data

6. How MT cells analyze the motion of visual patterns, http://courses.washington.edu/psy448/pdf/Rust-nn06.pdf

7. Advancement of motion psychophysics: Review 2001–2010 | JOV, https://jov.arvojournals.org/article.aspx?articleid=2191813

8. Decoding the motion aftereffect in human visual cortex \- Ovid, https://www.ovid.com/journals/neuima/pdf/10.1016/j.neuroimage.2013.06.034\~decoding-the-motion-aftereffect-in-human-visual-cortex

9. The motion after-effect reloaded \- PMC \- NIH, https://pmc.ncbi.nlm.nih.gov/articles/PMC3087115/

10. (PDF) Motion adaptation and attention: A critical review and meta, https://www.researchgate.net/publication/328383418\_Motion\_adaptation\_and\_attention\_A\_critical\_review\_and\_meta-analysis

11. Disambiguating serial effects of multiple timescales \- Journal of Vision, https://jov.arvojournals.org/article.aspx?articleid=2737369

12. the induction of the optic flow motion aftereffect in the central, https://rex.libraries.wsu.edu/view/pdfCoverPage?instCode=01ALLIANCE\_WSU\&filePid=13338233100001842\&download=true

13. Distinct mechanism for long-term contrast adaptation \- PMC \- NIH, https://pmc.ncbi.nlm.nih.gov/articles/PMC3326457/

14. Clustering of Self-Motion Selectivity and Visual Response Properties, https://pmc.ncbi.nlm.nih.gov/articles/PMC2652227/

15. neural mechanisms for self-motion \- perception in area mst, https://www.vis.caltech.edu/documents/9402/perception\_in\_area\_mst\_1999.pdf

16. A space-variant model for motion interpretation across the visual field, https://jov.arvojournals.org/article.aspx?articleid=2546929

17. Hierarchical processing of complex motion along the primate dorsal, https://www.pnas.org/doi/10.1073/pnas.1115685109

18. Visual-vestibular cue integration for heading perception \- PMC \- NIH, https://pmc.ncbi.nlm.nih.gov/articles/PMC3108057/

19. Expanding and Contracting Optic-Flow Patterns and Vection, https://www.researchgate.net/publication/5246152\_Expanding\_and\_Contracting\_Optic-Flow\_Patterns\_and\_Vection

20. Review Simulated Viewpoint Jitter Shakes Sensory Conflict, https://www.eecs.yorku.ca/\~rallison/papers/palm%20jitter%20review.pdf

21. The Contributions of Vestibular Signals to the Representations of, https://www.vis.caltech.edu/documents/9404/the\_contribution\_of\_vestibular\_99\_1.pdf

22. Area MSTd Neurons Encode Visual Stimuli in Eye Coordinates, https://pmc.ncbi.nlm.nih.gov/articles/PMC3023369/

23. Orientation-defined visual rotation significantly affects observer's, https://www.researchgate.net/publication/347922160\_Orientation-defined\_visual\_rotation\_significantly\_affects\_observer's\_perceived\_self-motion

24. Computational modelling of optic flow selectivity in MSTd neurons, https://www.bu.edu/bravi/files/2018/11/beardsley1998MSTd\_model.pdf

25. (PDF) The role of central and peripheral vision in postural control, https://www.researchgate.net/publication/12729704\_The\_role\_of\_central\_and\_peripheral\_vision\_in\_postural\_control\_duringwalking

26. Computational Mechanisms for Perceptual Stability using Disparity, https://pmc.ncbi.nlm.nih.gov/articles/PMC6989005/

27. Directional performance in motion transparency. \- Semantic Scholar, https://www.semanticscholar.org/paper/bad70d023a0c6c0621b7f7ae8f4044f73f5df8d3

28. The Effect of Retinal Eccentricity on Optic Flow Deficits \- PMC \- NIH, https://pmc.ncbi.nlm.nih.gov/articles/PMC4658432/

29. (PDF) Investigating the Crucial Role of Optic Flow in Postural Control, https://www.researchgate.net/publication/331602868\_Investigating\_the\_Crucial\_Role\_of\_Optic\_Flow\_in\_Postural\_Control\_Central\_vs\_Peripheral\_Visual\_Field

30. (PDF) Comparing the effectiveness of different displays in, https://www.researchgate.net/publication/277885485\_Comparing\_the\_effectiveness\_of\_different\_displays\_in\_enhancing\_illusions\_of\_self-movement\_vection

31. Postural sway in the moving room scenario: New evidence for ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC8432855/

32. Effects of luminance contrast, averaged luminance and spatial, https://www.researchgate.net/publication/354646511\_Effects\_of\_luminance\_contrast\_averaged\_luminance\_and\_spatial\_frequency\_on\_vection

33. Normalization of neuronal responses in cortical area MT across, https://pmc.ncbi.nlm.nih.gov/articles/PMC4137245/

34. Vection increases the magnitude and accuracy of visually evoked, https://www.researchgate.net/publication/11023430\_Vection\_increases\_the\_magnitude\_and\_accuracy\_of\_visually\_evoked\_postural\_responses

35. Motion perception during sinusoidal smooth pursuit eye movements, https://jov.arvojournals.org/article.aspx?articleid=2193172

36. The neural basis of smooth-pursuit eye movements, http://wexler.free.fr/library/files/thier%20(2005)%20the%20neural%20basis%20of%20smooth-pursuit%20eye%20movements.pdf

37. Cognitive processes involved in smooth pursuit eye movements, https://www.frontiersin.org/journals/systems-neuroscience/articles/10.3389/fnsys.2013.00004/full

38. A new spin on vection in depth | JOV \- Journal of Vision, https://jov.arvojournals.org/article.aspx?articleid=2121810

39. Neural coding of multiple motion speeds in visual cortical area MT, http://biorxiv.org/cgi/content/full/2023.04.08.532456v4

40. Spikes, Decisions, And Actions: The Dynamical Foundations Of, https://vdoc.pub/documents/spikes-decisions-and-actions-the-dynamical-foundations-of-neurosciences-4dn74d7q0pb0

41. Determinants of the direction illusion: Motion speed and dichoptic, https://www.researchgate.net/publication/264054213\_Determinants\_of\_the\_direction\_illusion\_Motion\_speed\_and\_dichoptic\_presentation\_interact\_to\_reveal\_systematic\_individual\_differences\_in\_sign

42. Log-spherical Mapping in SDF Raymarching \- osar.fr, https://www.osar.fr/notes/logspherical/

43. hyprsaver — Rust application // Lib.rs, https://lib.rs/crates/hyprsaver

44. Fractals/Computer graphic techniques/2D/exp \- Wikibooks, https://en.wikibooks.org/wiki/Fractals/Computer\_graphic\_techniques/2D/exp

45. Creating Chaotic Flow Fields with GPGPU in React Three Fiber, https://medium.com/@midnightdemise123/creating-chaotic-flow-fields-with-gpgpu-in-react-three-fiber-f9aad608c534

46. Curl Noise \- Emil Dziewanowski, https://emildziewanowski.com/curl-noise/

47. Directable, High-Resolution Simulation of Fire on the GPU, https://www.researchgate.net/publication/220184274\_Directable\_High-Resolution\_Simulation\_of\_Fire\_on\_the\_GPU

48. Cybersickness in Head-Mounted Displays Is Caused by Differences, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2020.587698/full

49. A review of two theories of motion sickness and their implications for, https://www.techno-press.org/content/?page=article\&journal=was\&volume=14\&num=6\&ordernum=2

50. Motion Sickness: Recent Advances and Emerging Insights, https://www.ijpsjournal.com/article/motion-sickness-recent-advances-and-emerging-insights

51. An ecological Theory of Motion Sickness and Postural Instability, https://www.researchgate.net/publication/239393320\_An\_ecological\_Theory\_of\_Motion\_Sickness\_and\_Postural\_Instability

52. Postural control modifications induced by synchronized ice skating, https://www.researchgate.net/publication/226058037\_Postural\_control\_modifications\_induced\_by\_synchronized\_ice\_skating

53. Effects of Ocular Dominance and Visual Input on Body Sway, https://www.researchgate.net/publication/5919216\_Effects\_of\_Ocular\_Dominance\_and\_Visual\_Input\_on\_Body\_Sway

54. (PDF) Simulated Viewpoint Jitter Shakes Sensory Conflict Accounts, https://www.researchgate.net/publication/51592876\_Simulated\_Viewpoint\_Jitter\_Shakes\_Sensory\_Conflict\_Accounts\_of\_Vection

55. Sensory conflict account of vection and its time course. This figure is, https://www.researchgate.net/figure/Sensory-conflict-account-of-vection-and-its-time-course-This-figure-is-published-in\_fig1\_51592876

56. An oscillating visual prime reduces vection onset latency | PLOS One, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0195886

57. Evidence against an ecological explanation of the jitter advantage, https://pmc.ncbi.nlm.nih.gov/articles/PMC4227477/

58. Simulator Sickness Questionnaire Scoring: How those "conversion, https://www.researchgate.net/post/Simulator-Sickness-Questionnaire-Scoring-How-those-conversion-formulas-were-derived

59. simulator sickness questionnaire: twenty years later, https://pubs.lib.uiowa.edu/driving/article/28474/galley/136766/view/

60. Web accessibility for seizures and physical reactions \- MDN Web Docs, https://developer.mozilla.org/en-US/docs/Web/Accessibility/Guides/Seizure\_disorders

61. Beating Cybersickness: The Complete VR/AR Comfort Playbook, https://medium.com/antaeus-ar/beating-cybersickness-the-complete-vr-ar-comfort-playbook-2025-59ea4e083b9f

62. Immersive Games | IEEE Technology Navigator, https://technav.ieee.org/topic/immersive-games/