|Home | About | Journals | Submit | Contact Us | Français|
Although single cone receptors have been imaged in the living human eye, there has been no observation of rods in vivo. Using an adaptive optics (AO) ophthalmoscope and post processing, evidence of rod mosaic was observed at 5° and 10° eccentricities in the horizontal, temporal retina. For 4 normal human subjects, small structures were observed in between the larger cones, and were observed repeatedly at the same locations on different days and with varying wavelengths. Image analysis gave spacings that agree well with rod measurements from histological data.
The earliest efforts to image retinal photoreceptors in vivo concentrated on snakes  and toads , which have good optical quality and large photoreceptors. In the living human eye, the requirement for pupil dilation and the associated aberration increase makes visualization of such cells challenging. To fully maximize the resolution and contrast of cellular structures in the human eye, it is necessary to correct all spatial and dynamic aberration variations. Adaptive optics (AO)  provides this capability, enabling resolution of single cones . Despite routine cone imaging, there is only one report of imaging of rods in vivo in normal human eyes  although Carroll et al.  report imaging of the photoreceptor mosaic in a rod monochromat. Imaging rods will have important applications in the study of retinitis pigmentosa  and age-related macular degeneration , which often affect these cells first. In vivo images of the foveal cones, which are similar in size to rods have been obtained , implying that the rod mosaic should also be observable. Foveal cones have center to center (c-c) spacing ranging from 1.9-3.4 μm, whereas for rods, the range is 2.2-3.0 μm [10,11].
Given that the eye has sufficient lateral resolution to image rods, a further challenge is their typical tuning properties (i.e. Stiles-Crawford effect): the directionality parameter ρ is 0.02 for rods compared to 0.05 for cones  implying they are less directional. The broader tuning and their much smaller size mean their photon return will be much less than surrounding cones.
Four normal subjects (denoted by N1, N2, N3 and N4) between the ages of 19 and 26 years were imaged using the Rochester AO ophthalmoscope . All subjects gave prior written consent in accordance with the Declaration of Helsinki. All had healthy retinas with either mild or no refractive error (3 emmetropes and 1 low myope). Axial lengths (IOL Master - Carl Zeiss Meditec, Dublin CA) and the scaled size of the 1° AO images using Bennett's adjustment  are given in Table 1. Subjects' pupils were dilated with one drop of 1% tropicamide followed by one drop of 2.5% phenylephrine prior to AO imaging. To ensure good fixation and to stabilize head movements, a bite bar was used. An image was taken once the AO system reached a rms. (root mean square) error value of <0.1 μm over the 6.8 mm exit pupil.
5° and 10° in the horizontal, temporal retina (TR) was chosen for several reasons: (i) the image quality degrades with increasing eccentricity due to increased scatter from the overlying retinal layers, (ii) the cones can be used as physical landmarks as the cone size and spacing increases with increasing eccentricity whereas the rod size and spacing stay constant over small excursions (<15°) [10,11], and (iii) there are fewer overlying nerve fibers.
The contrast of the cone mosaic is wavelength invariant  despite the increasing reflectance from the posterior layers at longer wavelengths . This increased reflectance coupled with a larger permitted input level, requires collection of a smaller number of frames, reducing the effect of camera read noise at the expense of slightly reduced resolution.
The imaging source was a krypton flashlamp, with pulse durations of 4 msec delivered through a 1.5 mm entrance pupil. Single pulse energies for the 650 and 750nm wavelengths were 0.44 and 0.27 μJ respectively, a factor of 40 below the safety limits . Wavelengths were chosen using interference filters (FWHM of 40 nm) with an estimated coherence length of 8.5-10.5μm, shorter than the thickness of the retina, reducing the speckle. Prior to imaging, subjects' retinas were bleached using a 10 sec exposure of 550nm light at 37 × 106 troland-sec, sufficient to bleach 98% of the photopigment . A set of 8 images was acquired in the 40 sec immediately afterwards, short enough to ensure that the photopigment had not regenerated . The process was then repeated at the other wavelength. Approximately 50 images were taken at each location for each wavelength and the best 5-7 images were selected for further processing. The imaging camera was mounted on an axial translation stage to correct for the chromatic aberration difference between the AO wavefront sensor (WFS) beacon at 820nm and the two imaging wavelengths.
For each image frame, the WFS estimate of the 66 Zernike coefficients (ANSI format  - 10th order) was used to construct the corresponding wavefront aberration profile. Monochromatic light and constant amplitude across the unobstructed 6.8mm pupil diameter was assumed. By Fourier-transforming the retinal image and multiplying by the ideal optical transfer function (OTF) divided by the OTF computed from the residual Zernike coefficients the residual error and diffraction blur was removed. Where the OTF was much smaller than the ideal value, noise amplification can become an issue. To avoid this, the correction filter was clipped to a maximum complex amplitude while retaining its phase. The resulting filter was applied in the spatial frequency domain and Fourier-transformed back to the angular image space.
The resulting images were further enhanced to facilitate analysis. The lower angular-frequency albedo structure was estimated using a Gaussian high-pass filter, adjusted so as to pass the cone and rod spatial frequencies while suppressing the larger variations. Using this filter the deconvolved image was decomposed into two parts: (i) a low-angular frequency background image and (ii) a high-angular frequency image containing the cones and rods. Examination of the power spectral densities (PSDs) of the averaged images suggested that a power-law enhancement of angular frequency power would bring the rods' visibility in-line with the cones. The higher-spatial frequencies (beyond the rod PSD) had a lower signal-to-noise ratio and the power-law enhancement would boost this noise source. This was avoided by including a 2-D Chebyshev filter adjusted to pass the desired structures, while strongly suppressing the noise-dominated spatial frequencies. Once this was done, and the power normalized to preserve the cones' visibility, the resulting enhanced image was added back to the background, resulting in an enhanced retinal image. All image processing operations are linear in intensity, allowing us to use a correlation registration algorithm  which corrected for small shifts between images and removed any torsional eye motion. Approximately 5-7 images were registered and summed.
Fig 1 shows the effect of image processing for subject N1. The same structures can be seen in the single, registered and deconvolved images.
Fig 2 shows enhanced images for subject N2 at 5° and 10° TR at 650nm. The cones show a decrease in spatial frequency with increasing eccentricity having peaks at 30 and 24 c/deg at 5° and 10° respectively. The rods showed PSD peak at 95 c/deg for each location.
The mean values (15 rod or cone c-c measurements in each case) for all 4 subjects at each location are given in Table 2 along with histological comparisons [10,11]. The cone and rod spacings were determined by manual measurement from the 1° AO images. Additionally, PSDs were determined to confirm the cone and rod c-c spacing using the full 1° images.
Speckle  is a concern for light sources with high spatial and temporal coherence. The expected speckle size for a 6.8 mm pupil is 1.9 μm at 650 nm increasing to 2.35 μm at 750 nm. A flashlamp source like the one used here is temporally incoherent but to ensure that the rod structures were not due any spatial coherence, several parameters were varied. On varying the wavelength, the same structures were observed at the same retinal locations on the same subject, this was also true when the imaging procedure was repeated 7 days later. Moreover, the separation of the rods did not change with either wavelength or time. Finally, the diameter of the entrance pupil was varied (1.5 - 3 mm diameter in 0.5 mm steps), again the same structures were observed at the same locations with the same separation. As the entrance pupil was increased a decrease in the cones and rods contrast was observed. This is due to less efficient coupling of the light into the cone receptors and hence increased scatter from the underlying choroidal structures.
Rods were not observed throughout the entire image, however collectively; the results indicate that these are indeed rods. In order to image them more readily and routinely, further improvements in imaging technology are required, e.g., utilizing pupil plane obscurations centered at the peak of an individual's Stiles Crawford function to suppress the light from the brighter cones. Utilizing the difference in directional sensitivity between rod and cones and differential bleaching may also help. Other imaging approaches such as AOSLO  offers improved signal to noise and lateral resolution and the potential for averaging many more frames. Since rods are often first to be damaged in retinal disease, if one can image them as readily as we can with cones, in-vivo monitoring of rod viability will make an invaluable clinical tool from both diagnosis and therapeutic aspects.
The authors thank J. Lin and J. Plandowski for their help. This work has been supported in part by National Eye Institute grants 1RO1EY020901, EY04367 and EY01319 and through the National Science Foundation and Technology - Center for Adaptive Optics, managed by the University of California at Santa Cruz.