Biology subjects

Sher, A.

Publications and source records attributed to Sher, A..

At least 19 recordsLinked to original sources

bioRxiv2026

Electrical stimulation precisely reproduces naturalistic spiking activity in complete and intermixed neural populations in the primate retina

Neurostimulation technologies have the potential to restore complex brain functions by evoking naturalistic patterns of neural activity. However, faithfully mimicking the brains natural neural code in diverse neuron types remains an unsolved challenge. Here we exploit focal electrical stimulation and recording to reproduce natural visual signals at cellular resolution in two intermixed and oppositely-tuned neural populations in the isolated macaque retina. We calibrated and delivered electrical stimulation sequences to precisely mimic recorded visually-evoked responses to natural images in complete populations of ~50 ON and OFF parasol cells uniformly sampling ~4{degrees} x 4{degrees} of visual angle. Distinct and opposing visually-evoked response patterns in the two intermixed populations were closely replicated spike-by-spike in 5 retinal preparations using electrical stimulation, reproducing spike trains with 51 {+/-} 17% lower variability than repeated presentations of the same visual stimulus. Focal electrical stimulation also replicated traveling waves of neural activity in response to visual motion, and highly structured artificial response patterns. These findings demonstrate for the first time that neurostimulation can match or exceed biological signaling fidelity in large, complete and intermixed neural populations.

bioengineering↗

bioRxiv2025

Cell-type specific repertoire of responses to natural scenes in primate retinal ganglion cells

At least 20 distinct retinal ganglion cells (RGC) types have been identified morphologically in the primate retina, but our understanding of the distinctive visual messages they send to various targets in the brain remains limited. Here, we use large-scale multi-electrode array recordings to examine how multiple functionally-distinct RGC types in the macaque retina respond to flashed natural images. Responses to white noise visual stimulation were used to functionally identify 936 RGCs of 12 types in three recordings. Each cell type was confirmed by the mosaic organization of receptive fields, and 7 cell types were cross-identified between recordings. The average kinetics of light response in each RGC type as well as the repertoire of distinct firing patterns that each type produces were examined across thousands of natural images. The kinetics of the average response across images were highly stereotyped among cells of each cell type and distinct for cells of different types. Moreover, the full repertoires of firing patterns produced by different cell types, assessed by their latency and duration, were generally quite distinct with only a few exceptions. Together these data provide an overview of the range of responses to natural images transmitted from the primate retina to the brain.

neuroscience↗

bioRxiv2025

A scalable framework for current steering at single-neuron resolution

Electrical stimulation at cellular resolution to restore the function of neural circuits is limited by the density of available electrode arrays. Although current steering with multi-electrode stimulation can be used to target cells between electrodes, it has not been proven for systematically targeting individual cells. We develop a framework for cellular-resolution current steering, leveraging the biophysics of electrically-evoked spike generation, and test its efficacy in isolated macaque and human retina. Currents were passed through three electrodes simultaneously using large-scale high-density microelectrode arrays, directly evoking single spikes in retinal ganglion cells. The currents combined either linearly or nonlinearly to drive spiking, depending on the geometry of the electrodes relative to the cell. These findings were captured by a biophysical model and by a simpler parametric model in which spikes can initiate at several sites on the cell membrane, and were leveraged to efficiently identify multi-electrode stimulation patterns that optimized cellular selectivity.

bioengineering↗

bioRxiv2024

Functional diversity in the output of the primate retina

The visual image transmitted by the retina to the brain has long been understood in terms of spatial filtering by the center-surround receptive fields of retinal ganglion cells (RGCs). Recently, this textbook view has been challenged by the stunning functional diversity and specificity observed in [~]40 distinct RGC types in the mouse retina. However, it is unclear whether the [~]20 morphologically and molecularly identified RGC types in primates exhibit similar functional diversity, or instead exhibit center-surround organization at different spatial scales. Here, we reveal striking and surprising functional diversity in macaque and human RGC types using large-scale multi-electrode recordings from isolated macaque and human retinas. In addition to the five well-known primate RGC types, 18-27 types were distinguished by their functional properties, likely revealing several previously unknown types. Surprisingly, many of these cell types exhibited striking non-classical receptive field structure, including irregular spatial and chromatic properties not previously reported in any species. Qualitatively similar results were observed in recordings from the human retina. The receptive fields of less-understood RGC types formed uniform mosaics covering visual space, confirming their classification, and the morphological counterparts of two types were established using single-cell recording. The striking receptive field diversity was paralleled by distinctive responses to natural movies and complexity of visual computation. These findings suggest that diverse RGC types, rather than merely filtering the scene at different spatial scales, instead play specialized roles in human vision.

neuroscience↗

bioRxiv2024

Understanding responses to multi-electrode epiretinal stimulation using a biophysical model

ObjectiveNeural interfaces are designed to evoke specific patterns of electrical activity in populations of neurons by stimulating with many electrodes. However, currents passed simultaneously through multiple electrodes often combine nonlinearly to drive neural responses, making evoked responses difficult to predict and control. This response nonlinearity could arise from the interaction of many excitable sites in each cell, any of which can produce a spike. However, this multi-site activation hypothesis is difficult to verify experimentally. ApproachWe developed a biophysical model to study retinal ganglion cell (RGC) responses to multi-electrode stimulation and validated it using data collected from ex vivo preparations of the macaque retina using a microelectrode array (512 electrodes; 30{micro}m pitch; 10{micro}m diameter). ResultsFirst, the model was validated by using it to reproduce essential empirical findings from single-electrode recording and stimulation, including recorded spike voltage waveforms at multiple locations and sigmoidal responses to injected current. Then, stimulation with two electrodes was modeled to test how the positioning of the electrodes relative to the cell affected the degree of response nonlinearity. Currents passed through pairs of electrodes positioned near the cell body or far from the axon (>40 {micro}m) exhibited approximately linear summation in evoking spikes. Currents passed through pairs of electrodes close to the axon summed linearly when their locations along the axon were similar, and nonlinearly otherwise. Over a range of electrode placements, several distinct, localized spike initiation sites were observed, and the number of these sites covaried with the degree of response nonlinearity. Similar trends were observed for three-electrode stimuli. All of these trends in the simulation were consistent with experimental observations. Significance. These findings support the multi-site activation hypothesis for nonlinear activation of neurons, providing a biophysical interpretation of previous experimental results and potentially enabling more efficient use of multi-electrode stimuli in future neural implants.

bioengineering↗

bioRxiv2024

Modeling responses of macaque and human retinal ganglion cells to natural images using a convolutional neural network

Linear-nonlinear (LN) cascade models provide a simple way to capture retinal ganglion cell (RGC) responses to artificial stimuli such as white noise, but their ability to model responses to natural images is limited. Recently, convolutional neural network (CNN) models have been shown to produce light response predictions that were substantially more accurate than those of a LN model. However, this modeling approach has not yet been applied to responses of macaque or human RGCs to natural images. Here, we train and test a CNN model on responses to natural images of the four numerically dominant RGC types in the macaque and human retina - ON parasol, OFF parasol, ON midget and OFF midget cells. Compared with the LN model, the CNN model provided substantially more accurate response predictions. Linear reconstructions of the visual stimulus were more accurate for CNN compared to LN model-generated responses, relative to reconstructions obtained from the recorded data. These findings demonstrate the effectiveness of a CNN model in capturing light responses of major RGC types in the macaque and human retinas in natural conditions.

neuroscience↗

bioRxiv2023

Decomposition of retinal ganglion cell electrical images for cell type and functional inference

ObjectiveIdentifying neuronal cell types and their biophysical properties based on their extracellular electrical features is a major challenge for experimental neuroscience and for the development of high-resolution brain-machine interfaces. One example is identification of retinal ganglion cell (RGC) types and their visual response properties, which is fundamental for developing future electronic implants that can restore vision. ApproachThe electrical image (EI) of a RGC, or the mean spatio-temporal voltage footprint of its recorded spikes on a high-density electrode array, contains substantial information about its anatomical, morphological, and functional properties. However, the analysis of these properties is complex because of the high-dimensional nature of the EI. We present a novel optimization-based algorithm to decompose electrical image into a low-dimensional, biophysically-based representation: the temporally-shifted superposition of three learned basis waveforms corresponding to spike waveforms produced in the somatic, dendritic and axonal cellular compartments. ResultsThe decomposition was evaluated using large-scale multi-electrode recordings from the macaque retina. The decomposition accurately localized the somatic and dendritic compartments of the cell. The imputed dendritic fields of RGCs correctly predicted the location and shape of their visual receptive fields. The inferred waveform amplitudes and shapes accurately identified the four major primate RGC types (ON and OFF midget and parasol cells) substantially more accurately than previous approaches. SignificanceThese findings contribute to more accurate inference of RGC types and their original light responses based purely on their electrical features, with potential implications for vision restoration technology.

neuroscience↗

bioRxiv2023

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 stimulusevoked correlations. Thus, fixational eye movements enhance the precision of the retinal representation.

neuroscience↗

bioRxiv2023

The heme oxygenase-1 metalloporphyrin inhibitor stannsoporfin enhances the bactericidal activity of a novel regimen for multidrug-resistant tuberculosis in a murine model

Multidrug-resistant (MDR) Mycobacterium tuberculosis (Mtb) poses significant challenges to global tuberculosis (TB) control efforts. Host-directed therapies (HDT) offer a novel approach for TB treatment by enhancing immune-mediated clearance of Mtb. Prior preclinical studies found that inhibition of heme oxygenase-1 (HO-1), an enzyme involved in heme metabolism, with tin-protoporphyrin IX (SnPP) significantly reduced mouse lung bacillary burden when co-administered with the first-line antitubercular regimen. Here we evaluated the adjunctive HDT activity of a novel HO-1 inhibitor, stannsoporfin (SnMP), in combination with a novel MDR-TB regimen comprising a next-generation diarylquinoline, TBAJ-876 (S), pretomanid (Pa), and a new oxazolidinone, TBI-223 (O) (collectively, SPaO) in Mtb-infected BALB/c mice. After 4 weeks of treatment, SPaO + SnMP 5 mg/kg reduced mean lung bacillary burden by an additional 0.69 log10 (P=0.01) relative to SPaO alone. As early as 2 weeks post-treatment initiation, SnMP adjunctive therapy differentially altered the expression of pro-inflammatory cytokine genes, and CD38, a marker of M1 macrophages. Next, we evaluated the sterilizing potential of SnMP adjunctive therapy in a mouse model of microbiological relapse. After 6 weeks of treatment, SPaO + SnMP 10 mg/kg reduced lung bacterial burdens to 0.71 {+/-} 0.23 log10 CFU, a 0.78 log-fold greater decrease in lung CFU compared to SpaO alone (P=0.005). However, adjunctive SnMP did not reduce microbiological relapse rates after 5 or 6 weeks of treatment. SnMP was well tolerated and did not significantly alter gross or histological lung pathology. SnMP is a promising HDT candidate requiring further study in combination with regimens for drug-resistant TB.

microbiology↗

bioRxiv2023

Pre-existing interferon gamma conditions the lung to mediate early control of SARS-CoV-2

Interferons (IFNs) are critical for anti-viral host defence. Type-1 and type-3 IFNs are typically associated with early control of viral replication and promotion of inflammatory immune responses; however, less is known about the role of IFN{gamma} in anti-viral immunity, particularly in the context of SARS-CoV-2. We have previously observed that lung infection with attenuated bacteria Mycobacterium bovis BCG achieved though intravenous (iv) administration provides strong protection against SARS-CoV-2 (SCV2) infection and disease in two mouse models. Assessment of the pulmonary cytokine milieu revealed that iv BCG induces a robust IFN{gamma} response and low levels of IFN{beta}. Here we examined the role of ongoing IFN{gamma} responses due to pre-established bacterial infection on SCV2 disease outcomes in two murine models. We report that IFN{gamma} is required for iv BCG induced reduction in pulmonary viral loads and that this outcome is dependent on IFN{gamma} receptor expression by non-hematopoietic cells. Further analysis revealed that BCG infection promotes the upregulation of interferon-stimulated genes (ISGs) with reported anti-viral activity by pneumocytes and bronchial epithelial cells in an IFN{gamma}-dependent manner, suggesting a possible mechanism for the observed protection. Finally, we confirmed the importance of IFN{gamma} in these anti-viral effects by demonstrating that the recombinant cytokine itself provides strong protection against SCV2 challenge when administered intranasally. Together, our data show that a pre-established IFN{gamma} response within the lung is protective against SCV2 infection, suggesting that concurrent or recent infections that drive IFN{gamma} may limit the pathogenesis of SCV2 and supporting possible prophylactic uses of IFN{gamma} in COVID-19 management.

immunology↗

bioRxiv2022

Helminth exposure protects against murine SARS-CoV-2 infection through macrophage dependent T cell activation

Helminth endemic regions report lower COVID-19 morbidity and mortality. Here, we show that lung remodeling from a prior infection with a lung migrating helminth, Nippostrongylus brasiliensis, enhances viral clearance and survival of human-ACE2 transgenic mice challenged with SARS-CoV-2 (SCV2). This protection is associated with a lymphocytic infiltrate including an increased accumulation of pulmonary SCV2-specific CD8+ T cells and anti-CD8 antibody depletion abrogated the N. brasiliensis-mediated reduction in viral loads. Pulmonary macrophages with a type-2 transcriptional signature persist in the lungs of N. brasiliensis exposed mice after clearance of the parasite and establish a primed environment for increased antigen presentation. Accordingly, depletion of macrophages ablated the augmented viral clearance and accumulation of CD8+ T cells driven by prior N. brasiliensis infection. Together, these findings support the concept that lung migrating helminths can limit disease severity during SCV2 infection through macrophage-dependent enhancement of anti-viral CD8+ T cell responses. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=139 SRC="FIGDIR/small/515832v1_ufig1.gif" ALT="Figure 1"> View larger version (55K): org.highwire.dtl.DTLVardef@1c17a22org.highwire.dtl.DTLVardef@13967b3org.highwire.dtl.DTLVardef@1c723a7org.highwire.dtl.DTLVardef@f5e038_HPS_FORMAT_FIGEXP M_FIG C_FIG

immunology↗

bioRxiv2022

Blood and site of disease inflammatory profiles differ in HIV-1-infected pericardial tuberculosis patients

ObjectivesTo better understand the pathogenesis of pericardial tuberculosis (PCTB), we sought to characterize the systemic inflammatory profile in HIV-1-infected participants with latent TB infection (LTBI), pulmonary TB (PTB) and PCTB. MethodsUsing Luminex, we measured 39 analytes in pericardial fluid (PCF) and paired plasma from 18 PCTB participants, and plasma from 16 LTBI and 20 PTB. Follow-up plasma samples were also obtained from PTB and PCTB participants. HLA-DR expression on Mtb-specific CD4 T cells was measured in baseline samples using flow cytometry. ResultsAssessment of the overall systemic inflammatory profile by principal component analysis showed that the inflammatory profile of active TB participants was distinct from the LTBI group, while PTB patients could not be distinguished from those with PCTB. In the LTBI group, 12 analytes showed a positive association with plasma HIV-1 viral load, and most of these associations were lost in the diseased groups. When comparing the inflammatory profile between PCF and paired blood, we found that the concentrations of most analytes (24/39) were elevated at site of disease. However, the inflammatory profile in PCF partially mirrored inflammatory events in the blood. After TB treatment completion, the overall plasma inflammatory profile reverted to those observed in the LTBI group. Lastly, HLA-DR expression showed the best performance for TB diagnosis compared to previously described biosignatures built from soluble markers. ConclusionOur results describe the inflammatory profile associated with PTB and PCTB and emphasize the potential role of HLA-DR as a promising biomarker for TB diagnosis.

immunology↗

bioRxiv2022

Glucuronoxylomannan intranasal challenge prior to Cryptococcus neoformans pulmonary infection enhances cerebral cryptococcosis in rodents

The encapsulated fungus Cryptococcus neoformans is the most common cause of fungal meningitis, with the highest rate of disease in patients with AIDS or immunosuppression. This microbe enters the human body via inhalation of infectious particles. C. neoformans capsular polysaccharide, in which the major component is glucuronoxylomannan (GXM), extensively accumulates in tissues and compromises host immune responses. C. neoformans travels from the lungs to the bloodstream and crosses to the brain via transcytosis, paracytosis, or inside of phagocytes using a "Trojan horse" mechanism. The fungus causes life-threatening meningoencephalitis with high mortality rates. Hence, we investigated the impact of intranasal exogenous GXM administration on C. neoformans infection in C57BL/6 mice. GXM enhances cryptococcal pulmonary infection and facilitates fungal systemic dissemination and brain invasion. Pre-challenge of GXM results in detection of the polysaccharide in lungs, serum, and surprisingly brain, the latter likely reached through the nasal cavity. GXM significantly alters endothelial cell tight junction protein expression in vivo, suggesting significant implications for the C. neoformans mechanisms of brain invasion. Using a microtiter transwell system, we showed that GXM disrupts the trans-endothelial electrical resistance, weakening the human brain endothelial cell monolayers co-cultured with pericytes, supportive cells of blood vessels/capillaries found in the blood-brain barrier (BBB), and promotes C. neoformans BBB penetration. Our findings should be considered in the development of therapeutics to combat the devastating complications of cryptococcosis that results in an estimated [~]200,000 deaths worldwide each year. AUTHOR SUMMARYCryptococcus neoformans infection of the central nervous system (CNS) typically begins by inhalation of fungal spores and results in devastating mortality rates worldwide. Over 200,000 deaths have been reported annually, with cryptococcal meningoencephalitis being the most severe form of the disease. This study investigates the ability of the fungus to invade, colonize, and cause damage to the host through properties of the fungal polysaccharide capsule, which allows the microbe a variety of both protective and offensive abilities. This capsule, made primarily of the polysaccharide glucuronoxylomannan (GXM), has been implicated in the progression and severity of cryptococcal infection in the CNS. We determined that GXM increases the fungal burden in the lungs of mice and enhances fungal migration to the brain. Interaction of GXM with the blood-brain barrier, which is a protective structure that regulates movement of particles into the CNS, demonstrated that GXM can disrupt the integrity of this barrier, compromising the delicate balances of fluids, immune cells, and other factors vital to the maintenance of the CNS. The findings of this study reveal the substantial role of GXM in establishing C. neoformans infection in the brain and necessitate future studies to further understand these interactions.

microbiology↗

bioRxiv2022

Large scale interrogation of retinal cell functions by 1-photon light-sheet microscopy

Visual processing in the retina depends on the collective activity of large ensembles of neurons organized in different layers. Current techniques for measuring activity of layer-specific neural ensembles rely on expensive pulsed infrared lasers to drive 2-photon activation of calcium-dependent fluorescent reporters. Here, we present a 1-photon light-sheet imaging system that can measure the activity in hundreds of ex vivo retinal neurons over a large field of view while simultaneously presenting visual stimuli. This allowed for a reliable functional classification of different retinal ganglion cell types. We also demonstrate that the system has sufficient resolution to image calcium entry at individual synaptic release sites across the axon terminals of dozens of simultaneously imaged bipolar cells. The simple design, a large field of view, and fast image acquisition, make this a powerful system for high-throughput and high-resolution measurements of retinal processing at a fraction of the cost of alternative approaches.

neuroscience↗

bioRxiv2022

Precise control of neural activity using temporally dithered and spatially multiplexed electrical stimulation

Neural implants have the potential to restore lost sensory function by electrically evoking the complex naturalistic activity patterns of neural populations. However, it can be difficult to predict and control evoked neural responses to simultaneous multi-electrode stimulation due to nonlinearity of the responses. We present a solution to this problem and demonstrate its utility in the context of a bi-directional retinal implant for restoring vision. A dynamically optimized stimulation approach encodes incoming visual stimuli into a rapid, greedily chosen, temporally dithered and spatially multiplexed sequence of simple stimulation patterns. Stimuli are selected to optimize the reconstruction of the visual stimulus from the evoked responses. Temporal dithering exploits the slow time scales of downstream neural processing, and spatial multiplexing exploits the independence of responses generated by distant electrodes. The approach was evaluated using an experimental laboratory prototype of a retinal implant: large-scale, high-resolution multi-electrode stimulation and recording of macaque and rat retinal ganglion cells ex vivo. The dynamically optimized stimulation approach substantially enhanced performance compared to existing approaches based on static mapping between visual stimulus intensity and current amplitude. The modular framework enabled parallel extensions to naturalistic viewing conditions, incorporation of perceptual similarity measures, and efficient implementation for an implantable device. A direct closed-loop test of the approach supported its potential use in vision restoration. Impact StatementA novel method of electrical stimulation to precisely control neural activity for sensory restoration exhibits improvements in visual stimulus reconstruction, enables efficient hardware design, and extends to naturalistic conditions.

neuroscience↗

bioRxiv2022

Inferring light responses of primate retinal ganglion cells using intrinsic electrical signatures

Reproducing high-acuity vision with an epiretinal implant will likely require inferring the natural light responses of diverse RGC types in the implanted retina, without measuring them directly. Here we demonstrate an approach that exploits intrinsic electrical features of primate RGCs. First, ON-parasol and OFF-parasol RGCs were identified with 95% accuracy using electrical features. Then, the somatic electrical footprint, predicted cell type, and average linear-nonlinear-Poisson model parameters of each cell type were used to infer a light response model for each cell. Across five retinas, these models achieved an average correlation with measured firing rates of 0.49 for white noise visual stimuli and 0.50 for natural scenes stimuli, compared to 0.65 and 0.58 respectively for models fitted to recorded light responses, an upper bound. This finding, and linear decoding of images from predicted RGC activity, suggested that the inference approach may be useful for high-fidelity sight restoration.

neuroscience↗

bioRxiv2022

High-fidelity restoration of visual signals by electrical stimulation in the central primate retina

Electrical stimulation of retinal ganglion cells (RGCs) with electronic implants provides rudimentary artificial vision to people blinded by retinal degeneration. However, current devices stimulate indiscriminately and therefore cannot reproduce the intricate neural code of the retina. Recent work has demonstrated more precise activation of RGCs using focal electrical stimulation with multi-electrode arrays in the peripheral macaque retina, but it is unclear how effective this can be in the central retina, which is required for high-resolution vision. This work probes the neural code and effectiveness of focal epiretinal stimulation in the central macaque retina, using large-scale electrical recording and stimulation ex vivo. The functional organization, light response properties, and electrical properties of the major RGC types in the central retina were mostly similar to the peripheral retina, with some notable differences in density, kinetics, linearity, spiking statistics and correlations. The major RGC types could be distinguished by their intrinsic electrical properties. Electrical stimulation targeting parasol cells revealed similar activation thresholds and reduced axon bundle activation in the central retina, but lower stimulation selectivity. Quantitative evaluation of the potential for image reconstruction from electrically-evoked parasol cell signals revealed higher overall expected image quality in the central retina. An exploration of inadvertent midget cell activation suggested that it could contribute high spatial frequency noise to the visual signal carried by parasol cells. These results support the possibility of reproducing high-acuity visual signals in the central retina with an epiretinal implant. Significance StatementArtificial restoration of vision with retinal implants is a major treatment for blindness. However, present-day implants do not provide high-resolution visual perception, in part because they do not reproduce the natural neural code of the retina. Here we demonstrate the level of visual signal reproduction that is possible with a future implant by examining how accurately responses to electrical stimulation of parasol retinal ganglion cells (RGCs) can convey visual signals. Although the precision of electrical stimulation in the central retina was diminished relative to the peripheral retina, the quality of expected visual signal reconstruction in parasol cells was greater. These findings suggest that visual signals could be restored with high fidelity in the central retina using a future retinal implant.

neuroscience↗

bioRxiv2022

Maximum a posteriori natural scene reconstruction from retinal ganglion cells with deep denoiser priors

Visual information arriving at the retina is transmitted to the brain by signals in the optic nerve, and the brain must rely solely on these signals to make inferences about the visual world. Previous work has probed the content of these signals by directly reconstructing images from retinal activity using linear regression or nonlinear regression with neural networks. Maximum a posteriori (MAP) reconstruction using retinal encoding models and separately-trained natural image priors offers a more general and principled approach. We develop a novel method for approximate MAP reconstruction that combines a generalized linear model for retinal responses to light, including their dependence on spike history and spikes of neighboring cells, with the image prior implicitly embedded in a deep convolutional neural network trained for image denoising. We use this method to reconstruct natural images from ex vivo simultaneously-recorded spikes of hundreds of retinal ganglion cells uniformly sampling a region of the retina. The method produces reconstructions that match or exceed the state-of-the-art in perceptual similarity and exhibit additional fine detail, while using substantially fewer model parameters than previous approaches. The use of more rudimentary encoding models (a linear-nonlinear-Poisson cascade) or image priors (a 1/f spectral model) significantly reduces reconstruction performance, indicating the essential role of both components in achieving high-quality reconstructed images from the retinal signal.

neuroscience↗