How to Reduce VR Motion Sickness in Patients: The Design Rules We Ship
A practical, evidence-based guide to reducing VR motion sickness in patients: framerate floors, vignetting, teleport locomotion, seated design and the extra care frail users need.
To reduce VR motion sickness in patients, the single most effective thing you can do is hold a rock-solid framerate at the headset's native refresh rate and remove player-driven acceleration from the camera. Everything else — vignetting, teleport locomotion, seated experiences, narrowed field of view — is a refinement on those two foundations. We build accessible and medical VR for Class I devices under EU MDR 2017/745, and the difference between a tolerable session and a patient pulling the headset off in distress almost always comes down to engineering discipline, not clever content.
The short version: lock to 90fps (never let it drop), avoid smooth acceleration, prefer teleport over joystick locomotion, design seated-first, apply a dynamic vignette during any movement, and give frail or elderly patients shorter sessions with a supervised exit at all times.
What causes VR motion sickness (cybersickness)?
Cybersickness is the nausea, cold sweats, pallor, disorientation and headache some people feel in virtual reality. It is a specific form of visually induced motion sickness. The dominant explanation is sensory conflict: your eyes report that you are moving — the world is sweeping past you — while your vestibular system, the balance organs in your inner ear, reports that you are sitting perfectly still. The brain has no clean way to reconcile "I'm moving" with "I'm not moving," and it responds the way it responds to suspected neurotoxins: by trying to make you stop and, failing that, vomit.
Three engineering failures feed that conflict, and they are the ones you actually control:
- Latency and dropped frames. If the rendered image lags behind the user's head movement, the world appears to "swim." Motion-to-photon latency above roughly 20ms is felt; above 40ms it is sickening.
- Visually-implied acceleration the body never feels. Accelerating, braking, strafing, or rotating the camera with a thumbstick are the worst offenders, because acceleration (not constant velocity) is precisely what the vestibular system is built to detect.
- Vection at the visual periphery. The sense of self-motion is driven heavily by movement in your peripheral vision. A wide, fully-rendered field of view full of moving texture is a vection machine.
Susceptibility varies enormously between people, and it is higher in exactly the populations that medical VR most wants to reach: elderly patients, people with vestibular disorders, those on certain medications, and anyone already nauseated or anxious. That is why the rules below are not optional polish — for a patient cohort they are a safety requirement.
Rule 1: Treat framerate as a hard floor, not a target
The most important number in comfortable VR is the one you must never miss. Modern standalone and PC headsets refresh at 72, 90, or 120Hz. Pick the device's native rate and treat it as a floor you are forbidden to drop below — for most clinical hardware that means a hard 90fps. A steady 72fps is more comfortable than a 90fps experience that stutters to 45; consistency beats peak.
Concretely, that changes how we engineer the scene rather than just how we wish it ran:
- We budget the entire frame to the refresh interval up front — at 90fps that is 11.1ms per frame, and we profile against it, not against an average.
- We lean on the runtime's reprojection (asynchronous spacewarp / timewarp) as a safety net for the occasional missed frame, never as a crutch for a scene that can't hit budget.
- We cap geometric and shader complexity so the worst-case frame — the patient turning to face the busiest part of the room — still lands inside budget.
If you only enforce one rule, enforce this one. A flawless comfort design running at an unstable framerate will still make people sick.
Rule 2: Kill player-driven acceleration
The body tolerates constant velocity far better than it tolerates change in velocity. So the design hierarchy we follow, from most to least comfortable, is:
- No virtual locomotion at all — the patient's real movement (head, hands, a step or two) maps 1:1 to the virtual world. Zero conflict.
- Teleport locomotion — the patient points, and the view cuts or quickly dashes to the new spot. There is no sustained visual acceleration to conflict with the still inner ear.
- Smooth (continuous) locomotion — joystick gliding. Comfortable for seasoned gamers, a reliable way to nauseate a first-time elderly patient. We avoid it in clinical builds, and when a client insists, it is off by default and heavily mitigated.
Rotation deserves its own warning. Smooth thumbstick turning is one of the fastest routes to nausea we know of. Replace it with snap turning — discrete 30° to 45° jumps that skip the in-between frames entirely. For seated patients who can physically turn their chair or torso, prefer real rotation and disable virtual turning altogether.
Rule 3: Vignette during motion (the "tunnelling" trick)
Because vection comes from the periphery, the most reliable mitigation for any unavoidable movement is a dynamic vignette: when the camera moves, we smoothly fade a soft black mask in from the edges, narrowing the visible field; when motion stops, we fade it back out. The patient keeps a clear, stable centre to fixate on while the provocative peripheral flow is simply hidden.
Done well it is nearly invisible — users report feeling fine without consciously noticing the effect. The parameters we tune: vignette aperture (how much we narrow it), and the fade timing so the mask itself never becomes a jarring acceleration. A stable reference frame in the periphery — a virtual cockpit, a fixed nose or visor element, a grounded horizon — helps for the same reason.
Rule 4: Design seated-first for patient cohorts
For frail, elderly, post-operative, or anxious patients, seated experiences are the default, not a fallback. A patient who cannot fall cannot be injured by a balance failure, and removing the standing-balance load itself reduces nausea and fatigue. We design the entire interaction so everything important sits within a comfortable seated reach and gaze cone — no content that demands standing, large steps, or floor-level reaching.
Around that we add the operational guardrails clinical use requires:
- Short sessions. We start frail patients at well under ten minutes and lengthen only as tolerance is demonstrated, rather than running a fixed program to completion.
- An always-available, one-action exit to a calm neutral scene — a fade to a static, horizon-locked environment, reachable by the patient or the supervising clinician instantly.
- A clinician in the loop. The supervisor watches for pallor, sweating and restlessness — early signs the patient may not report in time — and can end the session without navigating a menu.
- No flashing or high-frequency flicker, both for comfort and because of photosensitive epilepsy risk in any patient population.
Rule 5: Be deliberate about field of view
A narrower field of view reduces peripheral vection and therefore sickness — but it also reduces immersion and presence, which is often the therapeutic point. So we do not crop the headset's FOV permanently. Instead we narrow it dynamically (the Rule 3 vignette) so the patient gets the full, immersive field while still and a protected, narrowed field only while moving. This is the right trade: comfort exactly when it is needed, immersion the rest of the time.
Other things that move the needle
- Eliminate camera moves the patient didn't ask for. No cinematic camera, no screen shake, no auto-tilting horizon. The patient's head is the only thing allowed to move the viewpoint.
- Keep the horizon stable and upright. A tilting or rolling horizon is strongly nauseogenic; lock it level.
- Get the IPD and headset fit right. A wrong inter-pupillary distance or a poorly seated headset causes eye strain and headaches that get blamed on "VR sickness." Fitting is part of the protocol, not an afterthought.
- Avoid dense, high-contrast moving textures across large surfaces — they amplify vection even without locomotion.
- Acclimatise. Tolerance builds with exposure. Short, repeated, comfortable sessions beat one long ambitious one, especially for first-time older patients.
FAQ
What framerate stops VR motion sickness?
There is no magic number that "stops" it, but a stable framerate at the headset's native refresh — commonly 90fps — with motion-to-photon latency under about 20ms is the baseline for comfort. Stability matters more than the peak figure: a consistent 72fps beats a 90fps experience that frequently drops frames.
Is teleport or smooth locomotion better for motion sickness?
Teleport is markedly more comfortable because it removes the sustained visual acceleration that conflicts with the inner ear. Smooth (joystick) locomotion is more immersive but far more provocative, especially for first-time and older users. For patients, default to teleport or no virtual locomotion at all.
Why are seated VR experiences recommended for patients?
Seated design removes fall risk, lowers physical fatigue, and reduces the balance load that contributes to nausea. For frail, elderly, or post-operative patients it is the safe default — paired with short sessions and an instantly available exit.
The takeaway
Reducing VR motion sickness in patients is mostly unglamorous engineering: a framerate you never miss, a camera the patient alone controls, teleport instead of glide, a vignette during motion, and a seated, supervised, short-session protocol for the people most at risk. Get those right and the content can finally do its job. We build this discipline into accessible and medical VR from the first prototype, because for a patient cohort, comfort is not a feature — it is the safety floor everything else stands on.