TeleAutoHINTS: A Virtual/Augmented Reality System for Automated Tele-neurologic Evaluation of Acute Vertigo
Haochen Wei1, Preetham Bachina2, Peter Kazanzides1, Kemar Green2
1Computer Science, Johns Hopkins University, 2Neurology, Johns Hopkins University School of Medicine
Objective:

This study aims to enhance workup for dizziness/vertigo by translating advances in virtual reality (VR) and augmented reality (AR) eye tracking to clinical practice.

Background:
Dizziness/vertigo is a widespread issue in the US, with the challenge of distinguishing benign causes from severe conditions like stroke in acute cases. The head impulse, nystagmus, and test of skew (HINTS) exam is an effective diagnostic tool. However, a shortage of experts capable of conducting and interpreting the test hinders its widespread use. The development of an automated tele-assessment HINTS exam would reduce the need for skilled operators and make this diagnostic method more accessible.
Design/Methods:
This study introduces the TeleAutoHINTS system, composed of a head-mounted patient testing interface (PTI) on the Microsoft HoloLens 2. Gaze and head position data were collected with the HoloLens’ built-in eye tracker and gyroscope respectively. The PTI recorded gaze data while the following tests were performed: 1) Skew test via the PTI displaying a target to alternating eyes; 2) Nystagmus test via PTI instructions to either fixate on a target or look away; 3) Head impulse test via passive head impulses from an on-site care provider. The results were displayed on a desktop medical provider interface (MPI), showing bilateral pupil positions for skew and nystagmus tests and head and eye velocity-time graphs for the head impulse test. The system was evaluated on three adult volunteers to assess usability and data quality.
Results:
All subjects had normal head impulse test with an eye to head velocity ratio of ≥0.7 and head velocity exceeding 120 degrees/second. All three nystagmus traces exhibited a typical jerk waveform with right-beating movements. The three skew tests appeared normal, although blinks mimicked a vertical eye deviation.
Conclusions:
By harnessing the potential of AR/VR, TeleAutoHINTS provides a robust system capable of simulating the HINTS exam and collecting high quality data. 
10.1212/WNL.0000000000208168