Abstract
Purpose :
Patient cooperation to fix gaze and suppress eye motion is a nearly universal strategy for avoiding motion artifacts during ophthalmic imaging. Imaging patients with nystagmus is consequently challenging as the constant involuntary motion of the eye corrupts the acquired data.
Methods :
Using our previously reported robotic OCT system that integrates an eye tracking system, we developed a predictive approach to optically attenuate motion artifacts during periodic eye motion. Our algorithm extracts the first and second movements over multiple nystagmus cycles, combines these observations to estimate the profile of each movement individually, and determines the fundamental frequency of oscillation. We then predict the future eye position using this model and aim the OCT beam to optically cancel the nystagmus motion. All calculations operate at 250 Hz whereas the estimation process operates in real-time at 25 Hz to tolerate variations in the nystagmus profile, frequency, or phase delay. This approach is sufficiently general to operate without explicitly classifying the type of nystagmus. We evaluated the effectiveness of this approach by imaging a model eye mounted on a motorized translation stage (Fig. 1) which simulated a logarithmic jerk nystagmus profile with 0.87 mm amplitude at 3.5 Hz. A second motorized stage simulated slow head drift, which the robotic OCT corrected by moving the scanner. We performed imaging and quantified motion artifacts with no correction, our previously reported non-predictive aiming, and predictive aiming.
Results :
Our experiments revealed oscillatory motion artifacts with amplitudes of 23, 13.5, and 6 pixels for no correction, non-predictive aiming, and predictive aiming, respectively, averaged over approximately 30 nystagmus cycles (Fig. 2). Predictive aiming achieved 73.9% and 55.6% reductions in motion artifact over no correction and non-predictive aiming, respectively.
Conclusions :
Predictive aiming enables a substantial and meaningful reduction in motion artifacts during OCT imaging of eyes undergoing periodic movements, such as those seen during nystagmus.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.