Optic flow at the view edge drives vection; the centre contributes less

What it is

At the same forward speed, high-contrast repeating texture at the view edge drags people more than the same texture dead ahead. Peripheral optic flow is the main channel for vection; flow near the fovea contributes far less to “I am moving.” This is the spatial distribution of flow in the view, not whether continuous locomotion exists, and not painting the whole field black.

The centre is for the path, the aim, the text. The edge is what tells the body whether the world is streaming.

Why it happens

The peripheral retina covers a large area and is tuned to large-scale motion; visual evidence of self-motion is mostly integrated there. The fovea is for identifying objects and is less sensitive to whole-field flow. The same texture at the edge therefore drives vection; at the centre it is more often “something is moving.” Faster speed increases the tangential component at the edge and hardens the illusion.

This is a different road from narrowing the field of view. Narrowing removes the edge stimulus. Here the edge is still present; the question is what is drawn on it. High-contrast stripes, nearby repeating rails, particles skimming the cheeks, all flood that channel. Pull those elements into the centre and make the edge a low-contrast far field, and vection drops while field width can stay.

Studying it

Use partitioned stimuli: the same flow pattern in the centre only, in an annular periphery only, or both. Take vection onset and intensity, and FMS alongside. Classic rotating drums and random-dot flow already show a peripheral advantage; in a headset the independent variable should be edge-texture density of a real scene, not only dots on a blank field.

Keep “feels like moving” and “feels sick” apart. Vection can precede discomfort, and for some people it is a pleasant sense of going forward. Treating a vection scale and FMS as one number mixes what the edge is doing.

Where it stops holding

A very narrow headset field has almost no usable periphery, and the distribution effect is cropped by hardware. If the centre is filled with radial expansion (a tunnel mouth, an acceleration effect over the reticle), the centre will drive vection too and the peripheral advantage shrinks. Eyes closed, very dark, or an edge bricked in by solid UI, and the channel is absent. The peripheral effect is purer in passive viewing than in active movement, because there is no “I am pushing the stick” expectation to explain the stream. Under optical see-through, real-room edge flow is genuine; virtual edge texture stacked on it can make two streams fight.

Applying it

• During continuous translation, do not pave the view edge with high-contrast, near, repeating texture or particles. Move near rails and fragments that fly past the ears into a central band, or into a low-contrast far field.
• When a speed sensation is wanted, put speed lines in a narrow band beside the centre, not as a ring spinning against the lens.
• If UI and HUD live at the edge, check whether they become extra flow during movement — head-locked edge chrome is manufactured peripheral flow.
• How to check: straight translation at one speed, high-contrast edge texture versus low-contrast edge. Vection reports and FMS should rise with the high-contrast edge. If only centre-texture changes matter, the distribution claim does not hold in that scene.

Related

• Same group: N4.07.2 Snap turns can replace smooth rotation, trading discrete jumps for less discomfort · N4.07.3 Forced first-person camera shake is more nauseating than the locomotion itself · N4.07.4 Real physical cues such as airflow and vibration can lower discomfort from virtual motion · N4.07.5 Susceptibility to a locomotion technique varies by content type, not only by person
• Nearby: N4.02 Continuous Locomotion · N1.13 Causes of Cybersickness and Individual Differences
• Search terms: peripheral optic flow · vection · retinal periphery