Abstract
Purpose :
The implementation of wavefront sensorless adaptive optics (WSAO) in an in vivo mouse retinal imager can enhance imaging quality. However, its practical application may encounter limitations if the WSAO optimization is domain-specific. This study aims to assess the feasibility of applying WSAO across various regions of the live mouse retina using a multimodal imager.
Methods :
A WSAO multimodal imaging system (Netra Systems, Inc., Pleasanton, CA) was custom-developed for in vivo mouse retinal imaging. The system comprises broadband spectral-domain optical coherence tomography (OCT) and multi-channel scanning laser ophthalmoscopy (SLO). A deformable mirror (DM) is integrated in a feedback control loop utilizing an image-based optimization algorithm. Cx3cr1+/GFP mice aged 2-4 months were used. Initially, we selected the region of interest (ROI) within the wide-field OCT en face image. Subsequently, the optimization algorithm, guided by a merit function designed to maximize image brightness, was applied. The DM's shape adjusted progressively, correcting lower-order modes first, then higher-order modes. The OCT-derived modal coefficients were also directly transferred for SLO image correction. Local patch-based image contrast and sharpness metrics were employed to evaluate the performance of WSAO across various regions.
Results :
We collected several ROI patches across four retinal quadrants per mouse. The Zernike modal optimization took less than a minute, allowing sufficient time to capture several ROIs during the ~45-minute anesthesia period. We observed that the overall brightness of the OCT en face images notably increased with WSAO in all instances. We also noted that application of the OCT-derived modal coefficients to the SLO imaging resulted in sharper images of microglial processes. In the image quality metric analysis of 13 SLO images, we found an average increase of approximately 15.5% in contrast and 9.0% in sharpness. However, one out of 13 images showed reduced quality metrics after implementing WSAO. Additionally, we noted that the image enhancement was frequently confined to a narrow field within the given ROI.
Conclusions :
We demonstrate that omni-regional WSAO is a feasible feature of the multi-modal mouse retinal imager. Future studies will focus on investigating the reproducibility of similar image quality at different time points, as well as the capability for quantitative analysis.
This abstract was presented at the 2024 ARVO Imaging in the Eye Conference, held in Seattle, WA, May 4, 2024.