Abstract
Purpose :
Full-field swept-source optical coherence tomography (FF-SS-OCT) acquires 3D images of the living human retina carrying meaningful phase information on the backscattered light. Here, we show that these phases can be used to enhance imaging by providing additional contrast, e.g., for imaging the function of photoreceptors and ganglion cells. We present an algorithm to extract functional data even in adverse conditions when noise and statistical sample fluctuations impede a direct phase evaluation.
Methods :
For FF-SS-OCT imaging we used a high speed area camera (60,000 frames/s) which acquired typically 512 interferograms during a wavelength sweep from which we reconstructed volumetric OCT data. This imaging is equivalent to an A-scan rates of about 40 MHz, largely exceeding the speed of most conventional OCT systems. We acquired series with 70 volumes during which we stimulated the retina with white light to provoke responses from photoreceptors and ganglion cells. After data reconstruction and residual motion correction, we computationally corrected for ocular aberrations if needed. Finally, the extended Knox-Thompson (KT) method, an algorithm that originated in astronomic speckle interferometry to image through the turbulent atmosphere, was adapted to achieve functional contrast.
Results :
The applied processing technique to visualize retinal function showed clear signals even in demanding situations, where fluctuating scatterers or uncertainties in image registration affect the imaging; situations, in which phase signals degraded over time. The KT algorithm had the biggest impact on functional ganglion cell imaging for which it significantly increased the signal-to-noise ratio. Even in image series, where standard phase differences hardly extracted meaningful functional contrast, robust signals were obtained.
Conclusions :
Combining phase stable data acquisition of FF-SS-OCT with suitable processing techniques is a powerful tool for structural and functional imaging in the living retina. Despite the considerably lower signal-to-noise ratio compared to conventional OCT, our phase processing techniques can extract functional signals from layers which currently cannot be imaged by scanning OCT. The combination of phase stable imaging with advanced algorithms thus brings us closer to functional imaging of the entire neural retina on a cellular level.
This abstract was presented at the 2019 ARVO Imaging in the Eye Conference, held in Vancouver, Canada, April 26-27, 2019.