Fig 1.
High speed adaptive optics line-scan confocal retinal imaging system.
LS1, LS2: Light sources. SLD: Superluminescent diode. CL: Cylindrical lens. DF: Dichroic filter. L1-L5: Lenses. BS1, BS2: Beam splitter. S1-S5: Spherical mirrors. GS: Galvanometric scanner. FM1, FM2: Flat mirrors. WS: Wavefront sensor. DM: Deformable mirror. The imaging path and the wavefront sensing path are indicted by dark red and yellow lines, respectively.
Fig 2.
Schematic of the image acquisition timing diagram.
The galvanometric scanner is driven by the sawtooth signal generated by a function generator. One period scanning of the galvanometric scanner forms a frame including 512 lines. Each line includes a time for exposure to collect imaging photons (the line exposure time) and a time for reading out the image data (readout time). The line control signal (L-Sync) has two programmable parameters: “Line Period” and “Line Exposure Time.” Given a fixed line numbers per frame, the line period is determined by the frame rate or the time for acquiring a frame, e.g., for a frame consisting of 512 lines, the line period for the frame rate 200 Hz is 9.7 μs, within which a maximum time of 8.4 μs can be used for camera exposure to collect imaging photons, and a minimum 1.3 μs must be reserved for data readout. The table listed the line period time and the maximum exposure time during each line under different frame rates. The L-Sync is sent to the frame grabber. When the frame grabber receives the frame synchronization signal (F-Sync) from the function generator, it starts to record a frame and a counter triggered by the L-Sync will record the number of lines acquired. When the counter reaches 512, the frame grabber will finish acquisition and wait for the next F-Sync. The F-Sync was purposely set with a delay to avoid image distortion at the beginning of each frame.
Fig 3.
Imaging SNR measured under different frame rates.
(A) The testing image for assessing SNR. (B) Intensity profile of 5 bright lines acquired under different frame rates. (C) The SNRs measured under different line exposure time corresponding to different frame rates, as indicated by 5 symbols. Red line is the average SNR of 5 measurements. The mean SNR at 200 FPS, 100 FPS, and 30 FPS are indicated by arrows.
Fig 4.
Photoreceptors imaged at different locations.
(A) The montage of the central macula, yellow boxes indicate retina areas shown in the following panels. (B) The foveal center. All cells are cone photoreceptors. Except for the very center area (< 0.2°), most cones were resolved. (C) 1.0° eccentricity temporal to the foveal center. All cells are cone photoreceptors. (D) 1.5° eccentricity superior to the foveal center. (E) 2.0° eccentricity inferior to the foveal center. The dark band is a blood vessel shadow. Yellow arrowheads point to rods. (F) 3° eccentricity nasal-inferior to the foveal center. Rods are smaller dots (indicated by yellow arrowheads) surrounding cones which are bigger and brighter. All images were acquired at 200 FPS, and then registered and averaged a set of 50 successive frames using custom software [35]. The scale bar in panel (B) also applies to (C)—(F).
Fig 5.
Retinal images acquired at different layers and system axial resolution.
The images were taken at approximately 3° eccentricity from the foveal center nasally. (A) The imaging light focused on the photoreceptor layer. The dark bands are shadows of retinal blood vessels. (B) The imaging light focused on the retinal blood vessels between the photoreceptor layer and the optic nerve fiber layer. (C) The focal plane is at the surface of the optic nerve fibers. (D) System axial resolution. All images were acquired at 200 FPS, and then registered and averaged a set of 50 successive frames using custom software [35]. The scale bar in (A) also applies to (B) and (C).
Fig 6.
Retinal motion estimated with retinal images acquired at different frame rates.
(A) Retinal motion trace calculated from images acquired at 30 FPS. (B) Power spectra of motion amplitude of (A). (C) Retinal motion trace calculated from images acquired at 100 FPS. (D) Power spectra of (C). Retinal motion trace calculated from images acquired at 200 FPS. (F) Power spectra of motion amplitude of (E). Red arrows indicate exemplary drifting motion periods.
Fig 7.
Image intra-frame distortions under different frame rates.
Fig 8.
Intra-frame distortions of 3 subjects assessed with images acquired with different frame rates.