Assessing Saccadic Eye Movements With Head-Mounted Display Virtual Reality Technology

Abstract

As our society is ageing globally, neurodegenerative disorders are becoming a relevant issue. Assessment of saccadic eye movement could provide objective values to help to understand the symptoms of disorders. HTC Corporation launched a new virtual reality (VR) headset, VIVE Pro Eye, implementing an infrared-based eye tracking technique together with VR technology. The purpose of this study is to evaluate whether the device can be used as an assessment tool of saccadic eye movement and to investigate the technical features of eye tracking. We developed a measurement system of saccadic eye movement with a simple VR environment on Unity VR design platform, following an internationally proposed standard saccade measurement protocol. We then measured the saccadic eye movement of seven healthy young adults to analyze the oculo-metrics of latency, peak velocity, and error rate of pro- and anti-saccade tasks: 120 trials in each task. We calculated these parameters based on the saccade detection algorithm that we have developed following previous studies. Consequently, our results revealed latency of 220.40 ± 43.16 ms, peak velocity of 357.90 ± 111.99°/s, and error rate of 0.24 ± 0.41% for the pro-saccade task, and latency of 343.35 ± 76.42 ms, peak velocity of 318.79 ± 116.69°/s, and error rate of 0.66 ± 0.76% for the anti-saccade task. In addition, we observed pupil diameter of 4.30 ± 1.15 mm (left eye) and 4.29 ± 1.08 mm (right eye) for the pro-saccade task, and of 4.21 ± 1.04 mm (left eye) and 4.22 ± 0.97 mm (right eye) for the anti-saccade task. Comparing between the descriptive statistics of previous studies and our results suggests that VIVE Pro Eye can function as an assessment tool of saccadic eye movement since our results are in the range of or close to the results of previous studies. Nonetheless, we found technical limitations especially about time-related measurement parameters. Further improvements in software and hardware of the device and measurement protocol, and more measurements with diverse age-groups and people with different health conditions are warranted to enhance the whole assessment system of saccadic eye movement.

Keywords: ageing; dementia; head mounted display (HMD); neurological disorder; pupillary response; saccade; saccadic eye movement; virtual reality.

Figure 1

Coordinate system of HTC VIVE Pro Eye, based on the manual of SRanipal SDK. (A) Coordinate system of eye tracking on VIVE Pro Eye. (B) Coordinate system of pupil position data from user's view. (C) Coordinate system of gaze direction vector from user's view.

Figure 2

Experiment system of saccadic eye movement assessment with HTC VIVE Pro Eye.

Figure 3

Measurement protocol and VR design for the assessment of saccadic eye movement. (A) Protocol of pro- and anti-saccade tasks per trial. (B) Designed VR environment for saccade assessment. (C) Measurement flow of saccade assessment.

Figure 4

Data processing of eye movement data and detection algorithm of saccades.

Figure 5

Displacement of gaze direction in degrees and normalized pupil position on X axis from the origin; The data of all the saccade trials of all the participants are overlapped within the same time period from the time when the red target appears to the time 800 ms after the target appears (Blue dots: gaze toward the left; Orange dots: gaze toward the right).

Figure 6

Data distribution of parameters; latency, peak velocity, and error rate of saccadic eye movement of each eye in each pro- and anti-saccade task of all the participants (Pro, Pro-saccade; Anti, Anti-saccade; L, Left eye; R, Right eye).

Figure 7

Pupil diameter of each left and right eye in each pro- and anti-saccade task of all the participants.

Figure 8

Sampling interval calculated with Unix time on Unity and time stamp from SRanipal SDK of all the participants; Blue line shows the unfiltered data and yellow line shows the filtered data.

Figure 9

Evaluation of a while loop for 12,000 times on Unity and Visual Studio on Windows and Mac operating systems.

Figure 10

Detailed sampling interval based on the time stamp recorded with SRanipal SDK and visual explanation of relation between sampling of eye tracker and clock signal of computer.