Pre-corneal tear film thickness in humans measured with a novel technique.

PURPOSE: The purpose of this work was to gather preliminary data in normals and dry eye subjects, using a new, non-invasive imaging platform to measure the thickness of pre-corneal tear film. METHODS: Human subjects were screened for dry eye and classified as dry or normal. Tear film thickness over the inferior paracentral cornea was measured using laser illumination and a complementary metal-oxide-semiconductor (CMOS) camera. A previously developed mathematical model was used to calculate the thickness of the tear film by applying the principle of spatial auto-correlation function (ACF). RESULTS: Mean tear film thickness values (±SD) were 3.05 µm (0.20) and 2.48 µm (0.32) on the initial visit for normals (n=18) and dry eye subjects (n=22), respectively, and were significantly different (p<0.001, 2-sample t-test). Repeatability was good between visit 1 and 2 for normals (intraclass correlation coefficient [ICC]=0.935) and dry eye subjects (ICC=0.950). Tear film thickness increased above baseline for the dry eye subjects following viscous drop instillation and remained significantly elevated for up to approximately 32 min (n=20; p<0.05 until 32 min; general linear mixed model and Dunnett's tests). CONCLUSIONS: This technique for imaging the ocular surface appears to provide tear thickness values in agreement with other non-invasive methods. Moreover, the technique can differentiate between normal and dry eye patient types.

Non-invasive in vivo measurement of the tear film using spatial autocorrelation in a live mammal model.

Tear film stability and its interaction with the corneal surface play an important role in maintaining ocular surface integrity and quality of vision. We present a non-invasive technique to quantify the pre-corneal tear film thickness. A cMOS camera is used to record the interference pattern produced by the reflections from multiple layers of the tear film Principles of spatial autocorrelation are applied to extract the frequency of the periodic patterns in the images. A mathematical model is developed to obtain the thickness of the tear film from the spatial autocorrelation image. The technique is validated using micro-fabricated thin parylene films. We obtained repeatable and precise measurement on a live rabbit model (N = 6). We obtained an average value of 10.2µm and standard deviation of, SD = 0.3 (N = 4). We measured one rabbit infected with HSV-1 virus that had a baseline tear film thickness of 4.7µm.