Colleen Rhoades

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.

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.

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.

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.

Variation in the neural code contributes to making each individual unique. We probed neural code variation using ∼100 neural population recordings from major ganglion cell types in the macaque retina, combined with an interpretable computational representation of individual variability using machine learning. This representation captured individual variation and covariation in properties such as nonlinearity, temporal dynamics, and spatial receptive field size, while preserving invariances, such as asymmetries between ON and OFF cells. The covariation of response properties in different cell types was associated with the proximity of lamination of their synaptic inputs. Surprisingly, male retinas exhibited higher firing rates and faster temporal integration than female retinas. Exploiting data from previously recorded macaque retinas enabled efficient characterization of a new macaque retina, and of a human retina. Simulations indicated that combining a vast dataset of healthy macaque recordings with behavioral feedback could be used to identify the neural code and thus improve retinal implants for vision restoration.

The visual message conveyed by a retinal ganglion cell (RGC) is often summarized by its spatial receptive field, but in principle also depends on the responses of other RGCs and natural image statistics. This possibility was explored by linear reconstruction of natural images from responses of the four numerically-dominant macaque RGC types. Reconstructions were highly consistent across retinas. The optimal reconstruction filter for each RGC – its visual message – reflected natural image statistics, and resembled the receptive field only when nearby, same-type cells were included. ON and OFF cells conveyed largely independent, complementary representations, and parasol and midget cells conveyed distinct features. Correlated activity and nonlinearities had statistically significant but minor effects on reconstruction. Simulated reconstructions, using linear-nonlinear cascade models of RGC light responses that incorporated measured spatial properties and nonlinearities, produced similar results. Spatiotemporal reconstructions exhibited similar spatial properties, suggesting that the results are relevant for natural vision.

The visual message conveyed by a retinal ganglion cell (RGC) is often summarized by its spatial receptive field, but in principle also depends on the responses of other RGCs and natural image statistics. This possibility was explored by linear reconstruction of natural images from responses of the four numerically-dominant macaque RGC types. Reconstructions were highly consistent across retinas. The optimal reconstruction filter for each RGC – its visual message – reflected natural image statistics, and resembled the receptive field only when nearby, same-type cells were included. ON and OFF cells conveyed largely independent, complementary representations, and parasol and midget cells conveyed distinct and expected features. Correlated activity and nonlinearities had statistically significant but minor effects on reconstruction. Simulated reconstructions, using linear-nonlinear cascade models of RGC light responses that incorporated measured spatial properties and nonlinearities, produced similar results. Spatiotemporal reconstructions exhibited similar spatial properties, suggesting that the results are relevant for natural vision.

Summary: The visual functions of the diverse retinal ganglion cell types in the primate retina, and the parallel visual pathways they initiate, remain poorly understood. Here, the unusual physiological and computational properties of the ON and OFF smooth monostratified (SM) ganglion cells are explored. Large-scale multi-electrode recordings from 48 macaque retinas revealed that these cells exhibited strikingly irregular receptive field structure composed of spatially segregated hotspots, quite different from the receptive fields of previously described retinal ganglion cell types. The ON and OFF SM cells are paired cell types, but OFF SM cells exhibited stronger hotspot structure than ON cells. Targeted visual stimulation and computational inference demonstrate strong nonlinear subunit properties of each hotspot that contributed to the signaling properties of SM cells. Analysis of shared inputs to neighboring SM cells indicated that each hotspot could not be explained by an individual presynaptic input. Surprisingly, visual stimulation of different hotspots produced subtly different spatiotemporal spike waveforms in the same SM cell, consistent with a dendritic contribution to hotspot structure. These findings point to a previously unreported nonlinear mechanism in the output of the primate retina that contributes to signaling spatial information.

Integration of rectified synaptic inputs is a widespread nonlinear motif in sensory neuroscience. We present a novel method for maximum likelihood estimation of nonlinear subunits by soft-clustering spike-triggered stimuli. Subunits estimated from parasol ganglion cells recorded in macaque retina partitioned the receptive field into compact regions, likely representing bipolar cell inputs. Joint clustering with multiple RGCs revealed shared subunits in neighboring cells, producing a parsimonious population model. Closed-loop subunit validation was then performed by projecting white noise into the null space of the linear receptive field. Responses to these null stimuli were more accurately explained by a model with multiple subunits, and were stronger in OFF cells than ON cells. Presentation of natural stimuli containing jittering edges and textures also revealed greater response prediction accuracy with the subunit model. Finally, the generality of the approach was demonstrated by application to V1 data.