Holographic optogenetic control of retinal neural activity

Sensory information is typically represented in distributed patterns of activity across large populations of neurons. Therefore, many emerging neuro-photonic applications, such as optogenetic retinal prostheses require systems capable of delivering intense, parallel and dynamic light patterns. Such a system will ideally allow photo-control with single-cell selectivity of large neural populations expressing optogenetic probes, rather than nonspecific flashed illumination of the whole population (as provided by many current optogenetic light delivery systems). In a recent Nature Communications report, Reutsky-Gefen, Shoham and colleagues demonstrate holographic optogenetic control of retinal neural activity which is shown to provide rapid cellular-resolution, massively parallel excitation across macroscopic (millimeter-scale) coverage areas. The study illustrates that diffractive wavefront shaping (holographic) tools offer a powerful modality for dynamic patterned photo-stimulation as they naturally combine the high intensity, efficiency and resolution that are characteristic of sequential laser deflectors (like acousto-optical deflectors) with the simultaneous scan-less parallel illumination of multiple locations of microdisplay array projectors, but without their respective limitations. Holographic tools were previously shown to allow structured excitation of dendritic arbors and neurons using neurotransmitter photolysis, as well as two-photon optogenetic stimulation of proximal neurons in brain slices.

Figure 1. Cellular-resolution optogenetic ‘retinal stimulation fields’ (SFs).  (a) Fluorescent image of retina and multielectrode array (black dots) with superposed distribution of SFs (estimated from spike-triggered averaging of pseudo-random holographic patterns with 20 spots each). Scale bar, 200 µm. (b) Retinal SFs matching somata of visualized ChR2-eYFP-expressing RGCs. Scale bar, 20 µm. (c) Mean SF spatial distribution (n = 202 units from 11 retinas). Scale bar, 50 µm.

Figure 1. Cellular-resolution optogenetic ‘retinal stimulation fields’ (SFs).
(a) Fluorescent image of retina and multielectrode array (black dots) with superposed distribution of SFs (estimated from spike-triggered averaging of pseudo-random holographic patterns with 20 spots each). Scale bar, 200 µm. (b) Retinal SFs matching somata of visualized ChR2-eYFP-expressing RGCs. Scale bar, 20 µm. (c) Mean SF spatial distribution (n = 202 units from 11 retinas). Scale bar, 50 µm.

The study’s main goal was to develop a prosthetic system that would enable minimally invasive stimulation of visually impaired retinas lacking their natural photoreceptors – bypassing the degenerated photoreceptors by direct optogenetic stimulation of the surviving ‘downstream’ retinal ganglion cells (RGCs). The optical system’s ability to stimulate ChR2-expressing RGCs in isolated retinas was characterized, demonstrating an ability to reliably excite RGCs (>80%) with very high temporal precision (half a millisecond), and sub-cellular spatial resolution. The RGCs’ “stimulation fields” (analogous to receptive fields in natural vision) were mapped using the holographic system. They were found to accurately overlap in both size and shape with the underlying retinal ganglion cells (Figure 1), establishing the system’s high spatial resolution and ability to target single cells.

Finally, the response characteristics of ChR2-expressing RGCs were utilized in order to emulate normal visual activity in pharmacologically-blinded retinas. The artificial activity of RGCs driven with the holographic system was highly correlated (R=0.8) with the RGCs’ normal activity obtained for visual stimuli prior to the addition of pharmacological blockers (Figure 2), validating a first step towards cellular-scale vision restoration in blind retinas using optogenetic stimulation.

Figure 2. Optogenetic emulation of visual responses. Raster graphs of visually evoked neural activity in response to running cat (blue) and optogenetically emulated neural activity (red). PSTH graphs were calculated and overlaid to illustrate the resemblance between the two point processes.

Figure 2. Optogenetic emulation of visual responses. Raster graphs of visually evoked neural activity in response to running cat (blue) and optogenetically emulated neural activity (red). PSTH
graphs were calculated and overlaid to illustrate the resemblance between the two point processes.

The next step towards implementing the system as an optogenetic retinal prosthesis would be to test the holographic system in vivo. Towards this goal the authors developed a modified optical system with the ability to project and image the optogenetic mouse retina in vivo at a cellular resolution. In vivo translation of the in vitro results could lead to the development of a portable prosthetic device for restoring high acuity vision in humans (Figure 3).

Figure 3. Holographic vision restoration scheme and translation to in vivo setup. a) Camera video stream is processed and fed to holographic projection system: a spatial light modulator (top left), phase-modulates a laser beam wavefront and excitation patterns are projected onto the photo-sensitized retina. b) Fundus fluorescence image of ChR2-eYFP expressing mouse retinal neurons (green) superimposed with fluorescence image of holographic illumination pattern on the retina (blue). Pattern spots were smaller than 40 µm. Scale bar, 250 µm.

Figure 3. Holographic vision restoration scheme and translation to in vivo setup. a) Camera video stream is processed and fed to holographic projection system: a spatial light modulator (top left), phase-modulates a laser beam wavefront and excitation patterns are projected onto the photo-sensitized retina. b) Fundus fluorescence image of ChR2-eYFP expressing mouse retinal neurons (green) superimposed with fluorescence image of holographic illumination pattern on the retina (blue). Pattern spots were smaller than 40 µm. Scale bar, 250 µm.

Applications of rapid, parallel holographic stimulation are not limited to vision restoration: this strategy can also permit flexible and efficient control of activity in large, optogenetic cellular networks. Therefore, in addition to this stimulation modality’s natural application as a compact neural interface medical device, it can be implemented as a powerful open-loop or closed-loop research tool for many neuroscience applications, to assist in deciphering the neural code.

This entry was posted in Journal Club and tagged , , , , . Bookmark the permalink.
Add Comment Register



Leave a Reply