Abstract
Purpose :
To identify and demonstrate novel contrast mechanisms available in simultaneous multi-aperture AOSLO imaging of human retina, and to devise new methods of combining both conventional and unique new modalities to reveal previously unseen structures in oriented tissues (e.g. vessels and neural layers).
Methods :
In conventional offset-aperture AOSLO imaging, the contrast mechanism is based on refractive deflection of the scanning beam as it passes through inner retinal structures that produce phase gradients. Two-channel split-detection schemes employ paired detectors to sense uniaxial scatter of reflected photons, whereas our quad-channel split-detection scheme enables biaxial scatter detection (four anti-symmetric modes among four detectors). This arrangement enables additional information to be gathered from many tissues, especially cylindrical ones (i.e., “aligned structures”), that scatter light differently relative to the alignment, thus creating a dipole effect. Using up to 200 µW of 760 nm light, we have imaged a number of structures throughout the eyes of healthy adult subjects with a five-channel AOSLO system (confocal plus four offsets) with various aperture offset parameters (4-14 Airy disk diameters, to date).
Results :
Our dipole (quad) detection scheme has provided enhanced contrast imaging and, thus, visualization of non-axisymmetric structures such as vessels and neural axons bundles (see figure).
Conclusions :
The symmetric and anti-symmetric quad-channel AOSLO offset imaging modalities are complementary; dipole-detection is insensitive to split-detection signals, and split-detection cannot sense dipole scattering. The addition of more detectors in new patterns or additional beam properties (e.g., shorter wavelengths, polarization) enables new contrast mechanisms in AOSLO retinal imaging. We expect these modalities to provide new insights into disease processes often characterized by defects in the normal structure of oriented cellular matrices.
This is a 2020 ARVO Annual Meeting abstract.