Rapid Measurement of Cone Photoreceptor Pointing using Focus Diversity

Abstract

 

Purpose

 

We present a method to rapidly measure the pointing direction of individual cone photoreceptors with adaptive optics (AO). Roorda and Williams (JOV, 2002) adjusted the entrance pupil to alter the direction of the light incident on the cones and showed that in normal eyes cones point towards the same pupil location. Choi et al. (IOVS 2008) used the same technique in a diseased eye showing a higher degree of variability. Here, for a fixed entrance pupil we rapidly change the focus to image the waveguided light in various axial planes. This defocus causes, in successive images, a relative shift in the apparent position of cones with different pointing directions. The relative shift with known defocus allows the relative tilt to be measured.

 

Methods

 

The AO flood illuminated fundus camera at The Ohio State University was used to acquire AO retinal images on normal subjects at 2 and 4° in the temporal retina (680nm light, 0.75° field of view, 6 mm pupil). Once the best focus was achieved, focus plane was rapidly changed both anterior and posterior to this position and the corresponding images recorded. Defocus was introduced using the AO DM, in sweeps of up to 0.5D in 0.025D steps lasting 1 second. Images were registered and cone positions measured in each image.

 

Results

 

The pointing variability of cones can be clearly seen, the majority of cones appearing essentially parallel. We explored using open and closed-loop images, seeing the effect in both. We collected data on subjects using both this technique and the method of Roorda and Williams, 2002. Tilt measurements can be compared on a per-cone basis.

 

Conclusions

 

Focus diversity allows the relative tilt of cones to be measured in less than a second. The degree of uniformity in the cone pointing may be an indicator of retinal health. The coming months will see the imaging of further normal subjects and refinement of the analysis procedure.

 

 

Montage (a) of non-AO retinal images (2° temporal) with focus sweep. Images were registered and added. Tilted cones appear to move laterally as focus shifts, causing them to appear blurred in the montage. Image (b) is a variance-based enhancement of (a) to better show the cones' motion with focus.

 

 

3 sample cuts through a registered image set. Defocus allows us to measure the irradiance in different planes. Most cones are seen as parallel beams, while some exhibit notable tilt. 2 cones are highlighted with red lines for each sample.

 

Keywords: 688 retina • 648 photoreceptors • 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)