Wearable AR for Nystagmus Examination in Patients with Vertigo: Randomized Crossover Usability Study

CLINICAL QUESTION

Can a wearable augmented reality (AR)–based system feasibly perform standardized oculomotor/nystagmus testing in patients with vertigo compared with conventional video-oculography?

BOTTOM LINE

This small, randomized crossover usability study showed that a wearable AR-based nystagmus examination system was feasible and well tolerated, with moderate preliminary diagnostic agreement compared with conventional video-oculography. Its high negative predictive value suggests potential usefulness for screening, but the findings require larger multicenter validation before clinical implementation.

BACKGROUND: Nystagmus analysis is central to evaluating vertigo and differentiating benign vestibular dysfunction from central causes such as stroke. Conventional video-oculography can provide standardized oculomotor assessment, but it typically requires specialized laboratory equipment and trained personnel, limiting access. Portable AR-based systems may expand access to vestibular assessment in outpatient, emergency, or telemedicine settings.

STUDY DESIGN: Randomized crossover feasibility/usability study. Participants underwent both wearable AR-based testing and conventional video-oculography with a 30-minute washout period. Waveform outputs from both systems were pooled and interpreted blindly by a board-certified otologist.

SETTING: Hospital-based clinical setting at Kaohsiung Chang Gung Memorial Hospital, Taiwan

SYNOPSIS: Nine patients with vertigo were enrolled between October 2024 and January 2025; eight completed both AR and conventional video-oculography examinations. One participant with prior cataract surgery could not be calibrated with the AR glasses, highlighting a potential limitation related to altered ocular optics.

The wearable AR system consisted of J7EF Gaze smart glasses, an Android-based portable device, and a back-end platform. The glasses incorporated dual Si-OLED displays, a 30 Hz infrared eye-tracking sensor, and an optional magnetic light shield to simulate Frenzel goggles. The Unity 3D software generated a virtual 1-meter display, and in-house software delivered six standardized stimuli, including smooth pursuit in horizontal and vertical axes, fixation, and saccades. Real-time gaze data were transmitted via wi-fi for automated analysis and secure storage.

Across 48 oculomotor data points, agreement rates between AR and conventional video-oculography ranged from 62.5% to 87.5%. Overall diagnostic accuracy was 77.1%, with sensitivity of 81.8%, specificity of 75.7%, positive predictive value of 50.0%, and negative predictive value of 93.3%. For suspected central vestibular pathology, sensitivity was 83.3%, and specificity was 100%, although these estimates are based on a very small sample and should be interpreted cautiously.

Tolerability was favorable. Visual analog scale discomfort scores did not differ significantly between AR and conventional testing, and no significant discomfort or adverse effects were reported. The authors note that wearable AR may help address delays and access barriers associated with conventional vestibular testing. Limitations include the very small sample size, wide confidence intervals, single-center design, one interpreting clinician, use of percent agreement rather than Cohen kappa for diagnostic concordance, and moderate positive predictive value. Further studies should include larger cohorts, multiple raters, improved calibration algorithms, and multicenter validation.

CITATION: Wu CN, et al. Wearable augmented reality for nystagmus examination in patients with vertigo: randomized crossover usability study. J Med Internet Res. 2025;27:e75327. doi:10.2196/75327.