The Key to Smooth Self-Driving Journeys May Be Hidden Deep Inside Inner Ear

The research paper 'Vestibular Time Constant and Individual Susceptibility to Motion Sickness in Real-World Driving,' published in the journal nature Scientific Reports, explores whether a measurable feature of the inner-ear balance system can identify passengers who experience motion sickness during real road journeys.

The results suggest that the vestibular time constant, or VTC, may offer useful physiological information that a conventional motion-sickness questionnaire cannot capture, especially when an electric vehicle slows through regenerative braking.

Why automated travel could make motion sickness more common

Autonomous vehicles may allow passengers to work, watch videos, read, or rest instead of concentrating on the road, but these activities can prevent the brain from anticipating changes in speed and direction. A passenger looking at a stationary screen receives visual information suggesting little movement, while the inner ear detects acceleration, braking, and cornering, creating a sensory mismatch that can cause dizziness, sweating, stomach discomfort, nausea, or vomiting.

Susceptibility differs from person to person: Age, sex, previous exposure, expectations, physiology, and control over movement can all shape the response. Researchers often use the Motion Sickness Susceptibility Questionnaire, known as the MSSQ, to estimate risk from a person's recalled experiences in cars, buses, aircraft, boats, swings, and amusement rides during childhood and adulthood. This method is accessible and practical, though memories may be incomplete and people have not necessarily encountered the same motion environments.

The VTC offers a different kind of information: It measures how quickly the vestibulo-ocular response fades after rotational movement and reflects a brain process called velocity storage, which briefly preserves signals from the inner ear. A longer VTC means that movement information remains active for more time, potentially prolonging a disagreement between vestibular, visual, and bodily signals.

Researchers recreated demanding passenger journeys on public roads

A total of 617 women completed the long-form MSSQ during screening. Thirty-six women with scores of at least 50 were invited, a threshold chosen to recruit people with average or greater motion-sickness susceptibility. Twelve withdrew before testing, leaving 24 participants with an average age of 33.2 years. Recruitment was limited to women because earlier research has generally found higher average susceptibility among women, increasing the likelihood that the experiment would produce measurable symptoms.

Each participant completed a clinical rotary-chair examination in darkness. The chair accelerated to 100 degrees per second, maintained that speed for 60 seconds, and stopped, while electro-oculography recorded involuntary horizontal eye movements. Researchers calculated each participant's VTC from the decline in slow-phase eye velocity during clockwise and counterclockwise rotations.

Motion sickness was tested using two identically configured Kia EV6 electric vehicles on a seven-kilometre public-road loop in South Korea. Curtains blocked the outside view, and participants watched videos on a tablet while sitting in the rear seat, reproducing a situation in which a future automated-vehicle passenger might be occupied with a screen. Professional drivers followed predefined motion profiles, and participants rated their discomfort every two minutes during each 30-minute session.

Twenty-one participants completed four acceleration-focused sessions involving rapid acceleration, steady driving, and gradual slowing. All 24 completed nine deceleration-focused sessions built around regenerative braking. Each profile was tested on a separate day. Symptoms were measured with the 11-point Misery Scale, or MISC, where zero means no discomfort, scores from two to five represent non-nausea symptoms of increasing intensity, six marks the onset of nausea, and ten represents vomiting.

Longer vestibular responses were connected with stronger symptoms

Average symptoms rose during the first 18 minutes of both driving conditions and then changed relatively little. By the end of the 30-minute sessions, mean MISC ratings had reached about 3.4 during acceleration and 3.7 during deceleration. Symptoms appeared to rise more quickly between four and ten minutes during the braking condition and remained slightly higher near the end.

The repeated measurements proved highly consistent across sessions, indicating that participants who became comparatively sick under one profile tended to respond similarly during other profiles. Reliability values for the session-averaged symptom measures ranged from 0.917 to 0.949 during acceleration and from 0.957 to 0.971 during deceleration.

MSSQ scores showed weak positive relationships with symptoms, none of which reached statistical significance. The questionnaire was also almost unrelated to VTC itself, with a correlation of just 0.04, suggesting that the two assessments capture different aspects of susceptibility.

VTC produced clearer results: During acceleration, longer time constants were moderately associated with higher symptom measures, although most findings did not reach the study's statistical threshold. Among participants who developed symptoms, VTC had its strongest acceleration-related association with discomfort reported at the end of a session.

The braking condition revealed a more consistent pattern: Across all 24 participants, VTC was significantly associated with peak symptoms and final-session symptoms. Among the 20 participants who became symptomatic, the relationship between VTC and maximum symptom severity reached a correlation of 0.58. Associations with mean, cumulative, and final symptoms also remained significant after correction for multiple comparisons.

Researchers found that participants with minimal symptoms tended to have VTC values of 12.7 seconds or less, while symptomatic participants included most of those with values of at least 13 seconds. These observations came from a small sample and were not validated as clinical cut-offs.

A possible tool for passenger-specific vehicle design

VTC was better than the MSSQ at distinguishing participants who reached the nausea threshold. During deceleration, its area under the diagnostic curve was 0.89, compared with 0.60 for the questionnaire, a statistically significant difference. At the more severe MISC threshold of seven, the figures were 0.96 for VTC and 0.63 for MSSQ. During acceleration, VTC also performed better, with an area of 0.75 compared with 0.64.

Larger changes in acceleration cannot simply explain the stronger braking result, since the deceleration sessions used lower nominal jerk levels than the acceleration sessions. Braking direction and timing may create a distinctive combination of vestibular and visual signals. A passenger with a longer-lasting inner-ear response could retain movement information after other sensory signals have changed, increasing conflict inside the brain.

The study involved only 24 adult women already selected for high questionnaire scores, so the findings cannot yet be generalised to men, broader populations, or people with low susceptibility. Menstrual-cycle phase was not fully controlled, symptoms depended on self-report, and electro-oculography was used instead of video-based eye tracking. Larger studies should include diverse participants, additional physiological measures such as heart-rate variability and brain activity, and independent testing of acceleration direction and timing.