Motion sickness

Why does the passenger reading a map feel sick while the driver steering the same car feels fine?

Two people sit in one car. The same corners, the same braking, the same accelerations arrive at both bodies. Twenty minutes in, the passenger with the map is grey and quiet, and the driver — who is doing far more work — is perfectly well.

Whatever motion sickness is, then, it cannot be caused by the motion. The physical input was identical. So the difference must lie in something each brain brought to the ride rather than something the road delivered. What could a driver possess that a passenger does not?

Start with what the body actually measures. The semicircular canals in each inner ear respond to angular acceleration — turning. The otolith organs respond to linear acceleration and to the constant pull of gravity, which is why they cannot distinguish tilting your head from accelerating forward without other information. Vision, meanwhile, reports how the world slides across the retina. In ordinary walking these sources agree, and the agreement is so reliable that the brain is built to expect it.

Now the passenger with the map. The page, the seat, the door pillar are all fixed relative to her eyes, so vision reports a stationary world with considerable confidence. The inner ear reports cornering and braking. The two disagree. That is the classic sensory-conflict account, and it is where most explanations stop.

But look at what it has actually done. It has named a discrepancy. It has not said why a discrepancy should produce nausea rather than, say, a mild recalibration you never notice — and it does not yet explain the driver, who is receiving the same vestibular signals and who is also, much of the time, looking at instruments or mirrors rather than at flowing scenery. So the conflict story is descriptive. We need something with more content.

Here is the candidate. The driver is not merely receiving the turn; she is causing it. Motor commands are accompanied by an internal copy that lets the brain predict the sensory consequences of its own actions before they arrive. The driver's model says "in half a second the world will swing left" — and it does. The prediction is met. The passenger has no such copy. Her model has nothing to go on, so every swerve arrives as an unmodelled surprise.

That reformulation earns its keep, because it predicts things the raw conflict story does not. Passengers who can see the road ahead get less sick than passengers who cannot, because the visual scene restores the ability to anticipate. Being told the manoeuvre in advance helps for the same reason. And habituation — getting your sea legs — makes sense only on this version: nothing about the ship's motion changes over three days, so if raw mismatch were the cause, nothing should improve. What changes is the model.

Then a sharp test. This account puts the vestibular organs at the centre of the story, so people whose vestibular organs do not function should be largely immune. They are: bilateral loss of labyrinthine function confers near-immunity to seasickness. Had such people become sick at ordinary rates, the whole vestibular framing would be finished, and no amount of talk about conflict would save it.

One honesty warning before we go further. Nothing above explains why the outcome is nausea. The popular answer — that the brain reads a mismatch between motion sensed and motion predicted as the signature of a swallowed neurotoxin, and so triggers vomiting — is a hypothesis, proposed in the 1970s and still unproven. It is plausible and it is not established, and you should hold it loosely. There is also a live rival to the whole framework, which holds that sickness follows prolonged failure of postural control rather than sensory mismatch as such, and which reports that instability appears before symptoms do. This is not settled ground.

The refinement worth keeping is that the conflict is not really between the eye and the ear. It is between what the brain expected and what arrived. Sense-versus-sense is one way that expectation can be violated; being moved by forces you did not command is another, and it is the one that separates driver from passenger.

Two qualifications. First, "conflict" is a slippery notion, because after the fact one can always locate some mismatch and declare it the cause. The internal-model version is preferable mainly because predictability can be manipulated on purpose — give a passenger the horizon, or the wheel, and watch symptoms fall. Second, individual susceptibility varies enormously, and the reasons are poorly understood. I am deliberately not quoting prevalence figures or sex differences: the published values move around with how the question is asked, and a number you cannot defend is worse than no number.

• Why simulator and VR sickness sometimes reverses the conflict — vision moving while the body is still — and whether the same account covers both directions.
• How the vestibular system recalibrates during adaptation, and why the adaptation reverses on return to land.

Every term the collection defines is gathered in the glossary.