When noise becomes signal: a study of blink rate using an eye tracker

Abstract

PurposeAlthough blink rate is considered important for digital eye strain and dry eye conditions, its assessment has been mostly manual. This study aimed at quantifying blink rate automatically using the Eyelink 1000 Plus eye tracker and comparing it with manual counting. Additionally, blink rate was correlated with reading eye movements and reading speed. MethodsThirty participants (mean age 26.8±3.5 years; 19 females) were enrolled. An on-screen reading task for 3 minutes duration, was repeated 3 times with a 2-minute break between the repetitions. During this task, eye movements and blinks were measured through the Eyelink 1000 Plus eye tracker. The blink was also counted manually by the examiner in real time from a video feed. ResultsThe blink rate obtained by the eye tracker (median [Q1, Q3]:11.7 [6.8,15.5] blinks/min) and manual counting (11.6 [7.1, 15.4] blinks/min) were comparable (p>0.5). No significant correlations were observed between the blink rate and any of the reading eye movement parameters except saccadic amplitude (Spearman's rho, rs =0.506). No statistically significant correlation was found between the blink rate and reading speed. ConclusionThe Eyelink 1000 Plus eye tracker gives a comparable blink rate compared to that obtained through manual counting. A positive correlation between the blink rate and saccade amplitude could indicate a coupled interaction between the two mechanisms to minimize visual loss during reading. The option of using this tracker to investigate blinks in dry eye and binocular vision disorders can be further explored.