Abstract
Purpose :
To evaluate the reliability of pupillometer measurements obtained from a novel software application incorporated into a commercially available virtual reality (VR) headset.
Methods :
A total of 25 control subjects participated in the study (13 females, 12 males, 18-58 age range). Each subject was assessed using Heru Prime (Heru, Inc., Miami, FL) software application installed on a wearable virtual reality head mounted device (Pico Neo 3 Pro Eye, ByteDance, Beijing, China), equipped with infrared eye tracking hardware. Pupillometer measurements were performed on three separate occasions within a single day. During these measurements, we recorded pupil size in dim and bright light, pupil size before and after accommodation, and velocity of pupillary constriction. The data obtained was then analyzed using Intraclass Correlation Coefficient (ICC), Cronbach’s alpha, Standard Deviation of the Residues (SD Residues) and the Coefficient of Variance (CoV) to assess the repeatability of the measurements.
Results :
Out of the 25 control subjects, 24 completed all three tests. High repeatability was noted for pupil size in dim (Cronbach’s alpha = 0.987, ICC = 0.981, 95% CI: 0.959-0.991) and bright light (Cronbach’s alpha = 0.985, ICC = 0.973, 95% CI: 0.909-0.988) with SD Residues at 0.111 mm and 0.108 mm, and CoV of 4.169% and 5.334% respectively. Moderate repeatability was observed for reactive velocity (Cronbach’s alpha = 0.716, ICC = 0.717, 95% CI: 0.544-0.832), with a SD Residues at 0.324 mm/s and CoV of 19.407%. Accommodation showed good consistency (Cronbach’s alpha = 0.837, ICC = 0.829, 95% CI: 0.724-0.898), with a SD Residues at 0.141 mm and CoV of 17.885%. Pupillary reactivity was categorized into three groups based on constriction velocity: non-reactive (velocity < 0.4 mm/s), sluggish (0.4 mm/s ≤ velocity < 0.8 mm/s), and brisk (velocity ≥ 0.8 mm/s). The measured repeatability among these categories was 100%.
Conclusions :
The preliminary results indicate that a pupillometer based on a wearable VR device can provide consistent and reproducible pupillometry results, demonstrating its potential as a reliable tool for pupillary assessment.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.