Validation of a novel software‐based platform to extract oculometric measures
Israel Rosset, Eitan Raveh, Assaf Ben Shimon, Vova Anisimov, Edmund Ben-Ami, Rivka Riven Kreitman, Micha Yochanan Breakstone · Acta Ophthalmologica · 2022
Abstract Purpose: Eye‐tracking systems are a valuable tool in the assessment of disorders by detecting eye movement abnormalities. However, current systems require lab settings, extraction of oculometric measures entails dedicated long sessions, and operation requires qualified personnel. Recently, a novel software‐based platform was developed, using a webcam, enabling the extraction of oculometric measures in a short test with seamless operation. The aim of this study was to validate this platform as compared with a validated eye‐tracker. Methods: 82 participants (51 females, age range 21–88) were evaluated with a software‐based platform (NeuraLight, Israel) in a 15‐minute eye‐tracking test. Participants were presented with different stimuli for various tasks. During the test, data from a validated eye‐tracker (Tobii Pro, Sweden) were captured. Correlations between oculometric measures captured by both methods were calculated using Pearson correlation. Observable information on demographic data was measured using Fisher information matrix. Results: Eye and gaze positions captured by the NeuraLight platform were highly correlated with eye‐tracker (r = 0.9–0.99, p 0.9, p < 0.01). Relative root mean square error of all measures in both methods was <0.1. Saccadic latency and fixation time were found to be in relation with participants' age. Conclusions: NeuraLight platform was significantly correlated with a validated eye‐tracker, providing high accuracy for eye and gaze positions and oculometric measures. Relation between oculometric measures and age was established. The platform can be implemented for assessment of eye movements in practice. Future studies are needed to use the platform in various conditions.