A computational approach to high-resolution imaging of the living human retina without hardware adaptive optics

We demonstrate high-resolution imaging of the living human retina by computationally correcting highorder ocular aberrations. These corrections are performed post-acquisition and without the need for a deformable mirror or wavefront sensor that are commonly employed in hardware adaptive optics (HAO) systems. With the introduction of HAO to ophthalmic imaging, high-resolution near diffraction-limited imaging of the living human retina has become possible. The combination of a deformable mirror, wavefront sensor, and supporting hardware/software, though, can more than double the cost of the underlying imaging modality, in addition to significantly increasing the system complexity and sensitivity to misalignment. Optical coherence tomography (OCT) allows 3-D imaging in addition to naturally providing the complex optical field of backscattered light. This is unlike a scanning laser ophthalmoscope which measures only the intensity of the backscattered light. Previously, our group has demonstrated the utility of a technique called computational adaptive optics (CAO) which utilizes the complex field measured with OCT to computationally correct for optical aberrations in a manner similar to HAO. Until now, CAO has been applied to ex vivo imaging and in vivo skin imaging. Here, we demonstrate in vivo imaging of cone photoreceptors using CAO. Additional practical considerations such as imaging speed, and stability are discussed.