Analysis of the mechanical response of in vivo human accommodation from dynamic OCT recording

Abstract

Purpose : Advancements in anterior eye OCT imaging has opened the opportunity to analyse in greater precision the mechanical response of the accommodation system. We applied an existing mechanical model to analyse in vivo human accommodation using images captured with a synchronous OCT system.

Methods : The study was approved by the IRB of the University of Miami and followed the tenets of the Declaration of Helsinki. Three subjects (age 22, 39 and 45 yo) were enrolled. The anterior eye of each subject was imaged at 13 fps with an OCT during accommodation tasks. Full details of the synchronous OCT imaging system have been described in Ruggeri et al (2016). Test eyes were corrected for their spherical equivalent refractive error. Following stabilisation at zero object vergence for at least 1.5 s, accommodation was initiated monocularly (left eye) by a step change to near object vergence. The younger subjects were recorded for 2 D and 4 D response. The 45 yo was tested only on 2 D vergence. Dynamic recording of the accommodation response then continued for approximately 4.5 s. Image frames were digitally analyzed to extract the geometry and positions of the lens and ciliary body. These data were fitted to the model of Beers et al (1994 & 1996) using ordinary least squares technique to obtain values for lens initial thickness (X), thickness change (A), time-constant for response (τ) and initial time of response (t₀).

Results : Figure 1 shows the result for the 2 yo during 2 D accommodation. Results for X, A, τ, t₀ and RMS of residuals of model fitting for both subjects and object vergences are given in Table 1. Analysis of the method using Monte Carlo techniques (1000 trials per scenario; Gaussian noise matching the fitted RMS) estimated precisions of ±0.001 (0.03%), ±0.0016 (1%), ±0.02 (6%) and ±0.013 (0.7%) for X, A, τ and t₀ respectively. Thus the results are in agreement with those of Beers et al (1994 & 1996). Examination of residual errors revealed the rapid rise portion of A around t₀ departs noticeably from the single exponential model. This suggests the need for either a higher-order dynamic component or additional viscous components to be added to achieve a mechanical model with greater fidelity.

Conclusions : Data from dynamic OCT recording on in vivo human accommodation were fitted to a mechanical model. Examination of residual errors suggest the need for an enhanced model with additional components.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.