Primate Retina Research Articles

PCDH15 Dual-AAV Gene Therapy for Deafness and Blindness in Usher Syndrome Type 1F Models.

Usher syndrome type 1F (USH1F), resulting from mutations in the protocadherin-15 (PCDH15) gene, is characterized by congenital lack of hearing and balance, and progressive blindness in the form of retinitis pigmentosa. In this study, we explore an approach for USH1F gene therapy, exceeding the single AAV packaging limit by employing a dual adeno-associated virus (AAV) strategy to deliver the full-length PCDH15 coding sequence. We demonstrate the efficacy of this strategy in mouse USH1F models, effectively restoring hearing and balance in these mice. Importantly, our approach also proves successful in expressing PCDH15 protein in clinically relevant retinal models, including human retinal organoids and non-human primate retina, showing efficient targeting of photoreceptors and proper protein expression in the calyceal processes. This research represents a major step toward advancing gene therapy for USH1F and the multiple challenges of hearing, balance, and vision impairment.

Regional tuning of photoreceptor adaptation in the primate retina

Adaptation in cone photoreceptors allows our visual system to effectively operate over an enormous range of light intensities. However, little is known about the properties of cone adaptation in the specialized region of the primate central retina called the fovea, which is densely packed with cones and mediates high-acuity central vision. Here we show that macaque foveal cones exhibit weaker and slower luminance adaptation compared to cones in the peripheral retina. We find that this difference in adaptive properties between foveal and peripheral cones is due to differences in the magnitude of a hyperpolarization-activated current, Ih. This Ih current regulates the strength and time course of luminance adaptation in peripheral cones where it is more prominent than in foveal cones. A weaker and slower adaptation in foveal cones helps maintain a higher sensitivity for a longer duration which may be well-suited for maximizing the collection of high-acuity information at the fovea during gaze fixation between rapid eye movements.

Development of Microcystoid Macular Degeneration in the Retina of Nonhuman Primates: Time-Course and Associated Pathologies

Purpose Microcystoid macular degeneration (MMD) is a condition where cystoid vacuoles develop within the inner nuclear layer of the retina in humans in a variety of disorders. Here we report the occurrence of MMD in non-human primates (NHPs) with various retinal ganglion cell (RGC) pathologies and evaluate the hypothesis that MMD does not precede RGC loss but follows it. Methods Morphological studies were performed of the retinas of NHPs, specifically both rhesus (Macaca mulatta) and cynomolgus macaques (Macaca fascicularis), in which MMD was identified after induction of experimental glaucoma (EG), hemiretinal endodiathermy axotomy (HEA), and spontaneous idiopathic bilateral optic atrophy. In vivo imaging analyses included fundus photography, fluorescein angiography (FA), optical coherence tomography (OCT), adaptive optics scanning laser ophthalmoscopy (AOSLO), light microscopy, and electron microscopy. Results MMD, like that seen on OCT scans of humans, was found in both rhesus and cynomolgus macaques with EG. Of 13 cynomolgus macaques with chronic EG imaged once with OCT six of 13 animals were noted to have MMD. MMD was also evident in a cynomolgus macaque with bilateral optic atrophy. Following HEA, MMD did not develop until at least 2 weeks following the RNFL loss. Conclusion These data suggest that MMD may be caused by a retrograde trans-synaptic process related to RGC loss. MMD is not associated with inflammation, nor would it be an independent indicator of drug toxicity per se in pre-clinical regulatory studies. Because of its inconsistent appearance and late development, MMD has limited use as a clinical biomarker.

A circuit motif for color in the human foveal retina

The neural pathways that start human color vision begin in the complex synaptic network of the foveal retina where signals originating in long (L), middle (M), and short (S) wavelength-sensitive cone photoreceptor types are compared through antagonistic interactions, referred to as opponency. In nonhuman primates, two cone opponent pathways are well established: an L vs. M cone circuit linked to the midget ganglion cell type, often called the red-green pathway, and an S vs. L + M cone circuit linked to the small bistratified ganglion cell type, often called the blue-yellow pathway. These pathways have been taken to correspond in human vision to cardinal directions in a trichromatic color space, providing the parallel inputs to higher-level color processing. Yet linking cone opponency in the nonhuman primate retina to color mechanisms in human vision has proven particularly difficult. Here, we apply connectomic reconstruction to the human foveal retina to trace parallel excitatory synaptic outputs from the S-ON (or "blue-cone") bipolar cell to the small bistratified cell and two additional ganglion cell types: a large bistratified ganglion cell and a subpopulation of ON-midget ganglion cells, whose synaptic connections suggest a significant and unique role in color vision. These two ganglion cell types are postsynaptic to both S-ON and L vs. M opponent midget bipolar cells and thus define excitatory pathways in the foveal retina that merge the cardinal red-green and blue-yellow circuits, with the potential for trichromatic cone opponency at the first stage of human vision.

Fixational Eye Movements Enhance the Precision of Visual Information Transmitted by the Primate Retina.

Fixational eye movements alter the number and timing of spikes transmitted from the retina to the brain, but whether these changes enhance or degrade the retinal signal is unclear. To quantify this, we developed a Bayesian method for reconstructing natural images from the recorded spikes of hundreds of retinal ganglion cells (RGCs) in the macaque retina (male), combining a likelihood model for RGC light responses with the natural image prior implicitly embedded in an artificial neural network optimized for denoising. The method matched or surpassed the performance of previous reconstruction algorithms, and provides an interpretable framework for characterizing the retinal signal. Reconstructions were improved with artificial stimulus jitter that emulated fixational eye movements, even when the eye movement trajectory was assumed to be unknown and had to be inferred from retinal spikes. Reconstructions were degraded by small artificial perturbations of spike times, revealing more precise temporal encoding than suggested by previous studies. Finally, reconstructions were substantially degraded when derived from a model that ignored cell-to-cell interactions, indicating the importance of stimulus-evoked correlations. Thus, fixational eye movements enhance the precision of the retinal representation.

Species–specific circuitry of double cone photoreceptors in two avian retinas

In most avian retinas, double cones (consisting of a principal and accessory member) outnumber other photoreceptor types and have been associated with various functions, such as encoding luminance, sensing polarized light, and magnetoreception. However, their down-stream circuitry is poorly understood, particularly across bird species. Analysing species differences is important to understand changes in circuitry driven by ecological adaptations. We compare the ultrastructure of double cones and their postsynaptic bipolar cells between a night-migratory European robin and non-migratory chicken. We discover four previously unidentified bipolar cell types in the European robin retina, including midget-like bipolar cells mainly connected to one principal member. A downstream ganglion cell reveals a complete midget-like circuit similar to a circuit in the peripheral primate retina. Additionally, we identify a selective circuit transmitting information from a specific subset of accessory members. Our data highlight species-specific differences in double cone to bipolar cell connectivity, potentially reflecting ecological adaptations.

Toward an Indoor Lighting Solution for Social Jet Lag.

There is growing interest in developing artificial lighting that stimulates intrinsically photosensitive retinal ganglion cells (ipRGCs) to entrain circadian rhythms to improve mood, sleep, and health. Efforts have focused on stimulating the intrinsic photopigment, melanopsin; however, specialized color vision circuits have been elucidated in the primate retina that transmit blue-yellow cone-opponent signals to ipRGCs. We designed a light that stimulates color-opponent inputs to ipRGCs by temporally alternating short- and long-wavelength components that strongly modulate short-wavelength sensitive (S) cones. Two-hour exposure to this S-cone modulating light produced an average circadian phase advance of 1 h and 20 min in 6 subjects (mean age = 30 years) compared to no phase advance for the subjects after exposure to a 500 lux white light equated for melanopsin effectiveness. These results are promising for developing artificial lighting that is highly effective in controlling circadian rhythms by invisibly modulating cone-opponent circuits.

Comparison of modulation efficiency between normal and degenerated primate retina.

With electrical stimulation, retinal prostheses bypass dysfunctional photoreceptors and activate the surviving bipolar or retinal ganglion cells (RGCs). Therefore, the effective modulation of RGCs is crucial for developing retinal prostheses. Substantial research has been performed on the ability of an electrical stimulus to generate a reliable RGC response. However, different experimental conditions show varying levels of how well the electrical stimulation evokes RGC spikes. Therefore, in this study, we attempted to extract an indicator to understand how the electrical stimulation effectively evokes RGC spikes. Six cynomolgus monkeys were used: three as controls and three as an N-methyl-N-nitrosourea (MNU)-induced retinal degeneration model. The retinal recordings were performed using 8 × 8 multi-electrode arrays (MEAs). Electrical stimulation consisted of symmetrical biphasic pulses of varying amplitudes and durations. The number of stimulation conditions that resulted in significantly higher post-stimulation firing rates than pre-stimulus firing rates was defined as the modulation efficiency ratio (MER). The MER was significantly lower in degenerated retinas than in normal retinas. We investigated the relationship between the variables and the MER in normal and degenerated primate RGCs. External variables, such as duration and inter-electrode distance, and internal variables, such as average firing rates and statistics (mean, standard deviation, and coefficient of variation [CV]) of inter-spike intervals (ISIs) of spontaneous spikes, were used. External variables had similar effects on MER in normal and degenerated RGCs. In contrast, internal variables affected MER differently in normal and degenerated RGCs. While in normal RGCs, they were not related to MER, in degenerated RGCs, the mean ISIs were positively correlated with MER, and the CV of ISIs was negatively correlated with MER. The most important variable affecting MER was the mean ISI. A shorter ISI indicates hyperactive firing in the degenerated retina, which prevents electrical stimulation from evoking more RGCs. We believe that this hyperactivity in degenerated retinas results in a lower MER than that in the normal retina. Our findings can be used to optimize the selection of stimulation channels for in vitro MEA experiments and practical calibration methods to achieve higher efficiency when testing retinal prostheses.

Primate retina trades single-photon detection for high-fidelity contrast encoding.

How the spike output of the retina enables human visual perception is not fully understood. Here, we address this at the sensitivity limit of vision by correlating human visual perception with the spike outputs of primate ON and OFF parasol (magnocellular) retinal ganglion cells in tightly matching stimulus conditions. We show that human vision at its ultimate sensitivity limit depends on the spike output of the ON but not the OFF retinal pathway. Consequently, nonlinear signal processing in the retinal ON pathway precludes perceptual detection of single photons in darkness but enables quantal-resolution discrimination of differences in light intensity.

The retina's neurovascular unit: Müller glial sheaths and neuronal contacts.

The neurovascular unit (NVU), comprising vascular, glial and neural elements, supports the energetic demands of neural computation, but this aspect of the retina's trilaminar vessel network is poorly understood. Only the innermost vessel layer - the superficial vascular plexus (SVP) - is ensheathed by astrocytes, like brain capillaries, whereas glial ensheathment in other layers derives from radial Müller glia. Using serial electron microscopy reconstructions from mouse and primate retina, we find that Müller processes cover capillaries in a tessellating pattern, mirroring the tiled astrocytic endfeet wrapping brain capillaries. However, gaps in the Müller sheath, found mainly in the intermediate vascular plexus (IVP), permit different neuron types to contact pericytes and the endothelial cells directly. Pericyte somata are a favored target, often at spine-like structures with a reduced or absent vascular basement lamina. Focal application of adenosine triphosphate (ATP) to the vitreal surface evoked Ca2+ signals in Müller sheaths in all three vascular layers. Pharmacological experiments confirmed that Müller sheaths express purinergic receptors that, when activated, trigger intracellular Ca2+ signals that are amplified by IP3-controlled intracellular Ca2+ stores. When rod photoreceptors die in a mouse model of retinitis pigmentosa (rd10), Müller sheaths dissociate from the deep vascular plexus (DVP) but are largely unchanged within the IVP or SVP. Thus, Müller glia interact with retinal vessels in a laminar, compartmentalized manner: glial sheathes are virtually complete in the SVP but fenestrated in the IVP, permitting direct neural-to-vascular contacts. In the DVP, the glial sheath is only modestly fenestrated and is vulnerable to photoreceptor degeneration.

Cone-Opponent Ganglion Cells in the Primate Fovea Tuned to Noncardinal Color Directions.

A long-standing question in vision science is how the three cone photoreceptor types-long (L), medium (M), and short (S) wavelength sensitive-combine to generate our perception of color. Hue perception can be described along two opponent axes: red-green and blue-yellow. Psychophysical measurements of color appearance indicate that the cone inputs to the red-green and blue-yellow opponent axes are M vs. L + S and L vs. M + S, respectively. However, the "cardinal directions of color space" revealed by psychophysical measurements of color detection thresholds following adaptation are L vs. M and S vs. L + M. These cardinal directions match the most common cone-opponent retinal ganglion cells (RGCs) in the primate retina. Accordingly, the cone opponency necessary for color appearance is thought to be established in the cortex. While neurons with the appropriate M vs. L + S and L vs. M + S opponency have been reported in the retina and lateral geniculate nucleus, their existence continues to be debated. Resolving this long-standing debate is necessary because a complete account of the cone opponency in the retinal output is critical for understanding how downstream neural circuits process color. Here, we performed adaptive optics calcium imaging to noninvasively measure foveal RGC light responses in the living Macaca fascicularis eye. We confirm the presence of L vs. M + S and M vs. L + S neurons with noncardinal cone opponency and demonstrate that cone-opponent signals in the retinal output are more diverse than classically thought.

Foveal Retinal Ganglion Cells Develop Altered Calcium Dynamics Weeks After Photoreceptor Ablation

Objective or purposePhysiological changes in retinal ganglion cells (RGCs) have been reported in rodent models of photoreceptor (PR) loss but this has not been investigated in primates. By expressing both a calcium indicator (GCaMP6s) and an optogenetic actuator (ChrimsonR) in foveal RGCs of the macaque, we reactivated RGCs in vivo and assessed their response in the weeks and years following PR loss. DesignWe used an in vivo calcium imaging approach to record optogenetically evoked activity in deafferented RGCs in primate fovea. Cellular scale recordings were made longitudinally over a 10 week period following photoreceptor ablation and compared to responses from RGCs that had lost photoreceptor input more than two years prior. ParticipantsThree eyes received photoreceptor ablation, OD of a male Macaca mulatta (M1), OS of a female Macaca fascicularis (M2) and OD of a male Macaca fascicularis (M3). Two animals were used for in vivo recording, one for histological assessment.Methods: Cones were ablated with an ultrafast laser delivered through an adaptive optics scanning light ophthalmoscope (AOSLO). A 0.5 s pulse of 25Hz 660nm light optogenetically stimulated RGCs, and the resulting GCaMP fluorescence signal was recorded using AOSLO. Measurements were repeated over 10 weeks immediately after PR ablation, at 2.3 years and in control RGCs. Main Outcome measuresThe calcium rise time, decay constant and sensitivity index of optogenetic mediated RGC were derived from GCaMP fluorescence recordings from 221 RGCs (Animal M1) and 218 RGCs (Animal M2) in vivo. ResultsFollowing photoreceptor ablation, the mean decay constant of the calcium response in RGCs decreased 1.5 fold (1.6±0.5 s to 0.6±0.3 s SD) over the 10 week observation period in subject 1 and 2.1 fold (2.5±0.5 s to 1.2±0.2 s SD) within 8 weeks in subject 2. Calcium rise time and sensitivity index were stable. Optogenetic reactivation remained possible 2.3 years after PR ablation. ConclusionsAltered calcium dynamics developed in primate foveal RGCs in the weeks after photoreceptor ablation. The mean decay constant of optogenetic mediated calcium responses decreased 1.5-2-fold. This is the first report of this phenomenon in primate retina and further work is required to understand the role these changes play in cell survival and activity.

Spike sorting in the presence of stimulation artifacts: a dynamical control systems approach

Objective. Bi-directional electronic neural interfaces, capable of both electrical recording and stimulation, communicate with the nervous system to permit precise calibration of electrical inputs by capturing the evoked neural responses. However, one significant challenge is that stimulation artifacts often mask the actual neural signals. To address this issue, we introduce a novel approach that employs dynamical control systems to detect and decipher electrically evoked neural activity despite the presence of electrical artifacts. Approach. Our proposed method leverages the unique spatiotemporal patterns of neural activity and electrical artifacts to distinguish and identify individual neural spikes. We designed distinctive dynamical models for both the stimulation artifact and each neuron observed during spontaneous neural activity. We can estimate which neurons were active by analyzing the recorded voltage responses across multiple electrodes post-stimulation. This technique also allows us to exclude signals from electrodes heavily affected by stimulation artifacts, such as the stimulating electrode itself, yet still accurately differentiate between evoked spikes and electrical artifacts. Main results. We applied our method to high-density multi-electrode recordings from the primate retina in an ex vivo setup, using a grid of 512 electrodes. Through repeated electrical stimulations at varying amplitudes, we were able to construct activation curves for each neuron. The curves obtained with our method closely resembled those derived from manual spike sorting. Additionally, the stimulation thresholds we estimated strongly agreed with those determined through manual analysis, demonstrating high reliability ( R2=0.951 for human 1 and R2=0.944 for human 2). Significance. Our method can effectively separate evoked neural spikes from stimulation artifacts by exploiting the distinct spatiotemporal propagation patterns captured by a dense, large-scale multi-electrode array. This technique holds promise for future applications in real-time closed-loop stimulation systems and for managing multi-channel stimulation strategies.

Mannose-coupled AAV2: A second-generation AAV vector for increased retinal gene therapy efficiency

Inherited retinal diseases are a leading and untreatable cause of blindness and are therefore candidate diseases for gene therapy. Recombinant vectors derived from adeno-associated virus (rAAV) are currently the most promising vehicles for in vivo therapeutic gene delivery to the retina. However, there is a need for novel AAV-based vectors with greater efficacy for ophthalmic applications, as underscored by recent reports of dose-related inflammatory responses in clinical trials of rAAV-based ocular gene therapies. Improved therapeutic efficacy of vectors would allow for decreases in the dose delivered, with consequent reductions in inflammatory reactions. Here, we describe the development of new rAAV vectors using bioconjugation chemistry to modify the rAAV capsid, thereby improving the therapeutic index. Covalent coupling of a mannose ligand, via the formation of a thiourea bond, to the amino groups of the rAAV capsid significantly increases vector transduction efficiency of both rat and nonhuman primate retinas. These optimized rAAV vectors have important implications for the treatment of a wide range of retinal diseases.

Single-Cell Transcriptomic Sequencing Reveals Tissue Architecture and Deciphers Pathological Reprogramming During Retinal Ischemia in Macaca fascicularis.

Acute retinal arterial ischemia diseases (ARAIDs) are ocular emergencies that require immediate intervention within a restricted therapeutic window to prevent blindness. However, the underlying molecular mechanisms contributing to the pathogenesis of ARAIDs remain enigmatic. Herein, we present the single-cell RNA sequencing (scRNA-seq) alterations during ischemia in the primate retina as a preliminary endeavor in understanding the molecular complexities of ARAIDs. An ophthalmic artery occlusion model was established through ophthalmic artery ligation in two Macaca fascicularis. scRNA-seq and bioinformatics analyses were used to detect retinal changes during ischemia, which are further validated by immunofluorescence analysis. Western blot and flow cytometry assays were performed to measure the microglia polarization status. The findings of this study reveal notable changes in the retina under acute ischemic conditions. Particularly, retinal ischemia compromised mitochondrial functions of rod photoreceptors, partly leading to the rapid loss of healthy rods. Furthermore, we observed a noteworthy transcriptional alteration in the activation of microglia induced by ischemia. The targeted correction of the proinflammatory cytokine CXCL8 effectively suppresses microglia M1 polarization in retinal ischemia, ultimately reducing the proinflammatory transformation in vitro. In addition, retina ischemia induced the apoptotic inclination of endothelial cells and the heightened interaction with microglia, which signifies the influence of microglia in disrupting the retinal-blood barrier. Our research has successfully identified and described the pathologic alterations occurring in several cell types during a short period of ischemia. These observations provide valuable insights for ameliorating retinal damage and promoting the restoration of vision.

Expression patterns of CYP26A1, FGF8, CDKN1A, and NPVF in the developing rhesus monkey retina

The fovea centralis (fovea) is a specialized region of the primate retina that plays crucial roles in high-resolution visual acuity and color perception. The fovea is characterized by a high density of cone photoreceptors and no rods, and unique anatomical properties that contribute to its remarkable visual capabilities. Early histological analyses identified some of the key events that contribute to foveal development, but the mechanisms that direct the specification of this area are not understood. Recently, the expression of the retinoic acid-metabolizing enzyme CYP26A1 has become a hallmark of some of the retinal specializations found in vertebrates, including the primate fovea and the high-acuity area in avian species. In chickens, the retinoic acid pathway regulates the expression of FGF8 to then direct the development of a rod-free area. Similarly, high levels of CYP26A1, CDKN1A, and NPVF expression have been observed in the primate macula using transcriptomic approaches. However, which retinal cells express these genes and their expression dynamics in the developing primate eye remain unknown. Here, we systematically characterize the expression patterns of CYP26A1, FGF8, CDKN1A, and NPVF during the development of the rhesus monkey retina, from early stages of development in the first trimester until the third trimester (near term). Our data suggest that some of the markers previously proposed to be fovea-specific are not enriched in the progenitors of the rhesus monkey fovea. In contrast, CYP26A1 is expressed at high levels in the progenitors of the fovea, while it localizes in a subpopulation of macular Müller glia cells later in development. Together these data provide invaluable insights into the expression dynamics of several molecules in the nonhuman primate retina and highlight the developmental advancement of the foveal region.

Contributed Session I: Inner limiting membrane peel extends vivo calcium imaging of ganglion cells (RGC) beyond the fovea in non-human primate.

Viral expression of the calcium indicator GCaMP in primate RGCs has enabled optical readout of retinal function at a cellular scale in vivo. To date, functional recording has been limited to transduced RGCs close to the foveal pit. In this study we evaluate ILM peel as a strategy to expand the area of transduced RGCs and allow functional recording beyond the fovea in the living eye. 4 eyes of 3 immunosuppressed macaca fascicularis received a 9-12° ILM peel centered on the fovea, followed by intravitreal injection of GCaMP8s 4-8 weeks post-peel. A 660nm flickering visual stimulus drove RGC GCaMP responses which were recorded with fluorescence adaptive optics scanning laser ophthalmoscopy. In all eyes GCaMP was expressed throughout the peeled area, representing a mean 8-fold enlargement in the area of expression relative to a control eye with no peel. Functional responses were obtained from RGCs at max eccentricities of 11.7 o, 8.0 o, 9.7 o, and 13.7 o and could be classified as ON or OFF types up to the edge of the peel. Mean RGC responses in ILM peeled and control eyes of the same animal were comparable at 3.5 o and longitudinal tracking of individual RGCs showed stable responses up to 6 months post-peel. ILM peel substantially expands the region of primate retina accessible for in vivo GCaMP beyond the foveal ring of RGCs. This presents new opportunities for physiological study of the retina and pre-clinical testing of novel therapies in retinal degeneration models.

Dye coupling of horizontal cells in the primate retina.

In the monkey retina, there are two distinct types of axon-bearing horizontal cells, known as H1 and H2 horizontal cells (HCs). In this study, cell bodies were prelabled using 4',6-diamidino-2-phenylindole (DAPI), and both H1 and H2 horizontal cells were filled with Neurobiotin™ to reveal their coupling, cellular details, and photoreceptor contacts. The confocal analysis of H1 and H2 HCs was used to assess the colocalization of terminal dendrites with glutamate receptors at cone pedicles. After filling H1 somas, a large coupled mosaic of H1 cells was labeled. The dendritic terminals of H1 cells contacted red/green cone pedicles, with the occasional sparse contact with blue cone pedicles observed. The H2 cells were also dye-coupled. They had larger dendritic fields and lower densities. The dendritic terminals of H2 cells preferentially contacted blue cone pedicles, but additional contacts with nearly all cones within the dendritic field were still observed. The red/green cones constitute 99% of the input to H1 HCs, whereas H2 HCs receive a more balanced input, which is composed of 58% red/green cones and 42% blue cones. These observations confirm those made in earlier studies on primate horizontal cells by Dacey and Goodchild in 1996. Both H1 and H2 HCs were axon-bearing. H1 axon terminals (H1 ATs) were independently coupled and contacted rod spherules exclusively. In contrast, the H2 axon terminals contacted cones, with some preference for blue cone pedicles, as reported by Chan and Grünert in 1998. The primate retina contains three independently coupled HC networks in the outer plexiform layer (OPL), identified as H1 and H2 somatic dendrites, and H1 ATs. At each cone pedicle, the colocalization of both H1 and H2 dendritic tips with GluA4 subunits close to the cone synaptic ribbons indicates that glutamate signaling from the cones to H1 and H2 horizontal cells is mediated by α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors.

An ON-type direction-selective ganglion cell in primate retina

To maintain a stable and clear image of the world, our eyes reflexively follow the direction in which a visual scene is moving. Such gaze-stabilization mechanisms reduce image blur as we move in the environment. In non-primate mammals, this behaviour is initiated by retinal output neurons called ON-type direction-selective ganglion cells (ON-DSGCs), which detect the direction of image motion and transmit signals to brainstem nuclei that drive compensatory eye movements1. However, ON-DSGCs have not yet been identified in the retina of primates, raising the possibility that this reflex is mediated by cortical visual areas. Here we mined single-cell RNA transcriptomic data from primate retina to identify a candidate ON-DSGC. We then combined two-photon calcium imaging, molecular identification and morphological analysis to reveal a population of ON-DSGCs in the macaque retina. The morphology, molecular signature and GABA (γ-aminobutyric acid)-dependent mechanisms that underlie direction selectivity in primate ON-DSGCs are highly conserved with those in other mammals. We further identify a candidate ON-DSGC in human retina. The presence of ON-DSGCs in primates highlights the need to examine the contribution of subcortical retinal mechanisms to normal and aberrant gaze stabilization in the developing and mature visual system.

Regional variation in the organization and connectivity of the first synapse in the primate night vision pathway

Sensitivity of primate daylight vision varies across the visual field. This is attributed to regional variations in cone photoreceptor density and synaptic connectivity of the underlying circuitry. In contrast, we have limited understanding of how synapse organization of the primate night vision pathway changes across space. Using serial electron microscopy, we reconstructed the first synapse of the night vision pathway between rod photoreceptors and second-order neurons, at multiple locations from the central part of the primate retina, fovea, to the periphery. We find that most facets of the rod synapse connectivity vary across retinal regions. However, rod synaptic divergence and convergence patterns do not change in the same manner across locations. Moreover, patterns of rod synapse organization are tightly correlated with photoreceptor density. Such regional heterogeneities revise the connectivity diagram of the primate rod synapse which will shape synapse function and sensitivity of the night vision pathway across visual space.