Non-centered spike-triggered covariance analysis reveals neurotrophin-3 as a developmental regulator of receptive field properties of ON-OFF retinal ganglion cells

Receptive Field Sizes of Nyxnob Mouse Retinal Ganglion Cells

Patients with congenital nystagmus, involuntary eye movements, often have a reduced visual acuity. Some of these patients have a retinal-specific mutation in the protein nyctalopin, which is also present in the Nyx^nob mouse. In these mice, retinal ganglion cells (RGCs) have oscillatory activity, which leads to expanded axonal projections towards the dLGN and consequently to a desegregation of retinal projections to the brain. In this study, we investigate whether the receptive fields of Nyx^nob RGCs have also expanded by measuring the size of their receptive fields using MEA recordings. Contrary to our expectation, relative to wild-type (WT) mice we found receptive field sizes in the Nyx^nob retina had not increased but instead had decreased for green-light preferring RGCs. Additionally, we also found the receptive fields of UV-light preferring RGCs are larger than green-light preferring RGCs in both WT and Nyx^nob mice.

The Face Module Emerged in a Deep Convolutional Neural Network Selectively Deprived of Face Experience

Can we recognize faces with zero experience on faces? This question is critical because it examines the role of experiences in the formation of domain-specific modules in the brain. Investigation with humans and non-human animals on this issue cannot easily dissociate the effect of the visual experience from that of the hardwired domain-specificity. Therefore, the present study built a model of selective deprivation of the experience on faces with a representative deep convolutional neural network, AlexNet, by removing all images containing faces from its training stimuli. This model did not show significant deficits in face categorization and discrimination, and face-selective modules automatically emerged. However, the deprivation reduced the domain-specificity of the face module. In sum, our study provides empirical evidence on the role of nature vs. nurture in developing the domain-specific modules that domain-specificity may evolve from non-specific experience without genetic predisposition, and is further fine-tuned by domain-specific experience.

Visual Deprivation Retards the Maturation of Dendritic Fields and Receptive Fields of Mouse Retinal Ganglion Cells

It was well documented that both the size of the dendritic field and receptive field of retinal ganglion cells (RGCs) are developmentally regulated in the mammalian retina, and visual stimulation is required for the maturation of the dendritic and receptive fields of mouse RGCs. However, it is not clear whether the developmental changes of the RGC receptive field correlate with the dendritic field and whether visual stimulation regulates the maturation of the dendritic field and receptive field of RGCs in a correlated manner. The present work demonstrated that both the dendritic and receptive fields of RGCs continuously develop after eye opening. However, the correlation between the developmental changes in the receptive field size and the dendritic field varies among different RGC types. These results suggest a continuous change of synaptic converging of RGC synaptic inputs in an RGC type-dependent manner. Besides, light deprivation impairs both the development of dendritic and receptive fields.

Spike-triggered Clustering for Retinal Ganglion Cell Classification

Retinal ganglion cells (RGCs), the retina's output neurons, encode visual information through spiking. The RGC receptive field (RF) represents the basic unit of visual information processing in the retina. RFs are commonly estimated using the spike-triggered average (STA), which is the average of the stimulus patterns to which a given RGC is sensitive. Whereas STA, based on the concept of the average, is simple and intuitive, it leaves more complex structures in the RFs undetected. Alternatively, spike-triggered covariance (STC) analysis provides information on second-order RF statistics. However, STC is computationally cumbersome and difficult to interpret. Thus, the objective of this study was to propose and validate a new computational method, called spike-triggered clustering (STCL), specific for multimodal RFs. Specifically, RFs were fit with a Gaussian mixture model, which provides the means and covariances of multiple RF clusters. The proposed method recovered bipolar stimulus patterns in the RFs of ON-OFF cells, while the STA identified only ON and OFF RGCs, and the remaining RGCs were labeled as unknown types. In contrast, our new STCL analysis distinguished ON-OFF RGCs from the ON, OFF, and unknown RGC types classified by STA. Thus, the proposed method enables us to include ON-OFF RGCs prior to retinal information analysis.

Optimized nonionic emulsifier for the efficient delivery of astaxanthin nanodispersions to retina: in vivo and ex vivo evaluations

Astaxanthin (AST) is a naturally occurring carotenoid with potent anti-oxidative and anti-inflammatory potency against chronic diseases. In this study, we suspended AST in different nonionic emulsifiers to produce nanodispersions. The basic physicochemical properties of the produced AST nanodispersions were verified to select the optimized nonionic emulsifier. Among the tested emulsifiers, Polysorbate 20 produced the AST nanoemulsions with smaller particle diameters, narrower size distributions, and higher AST contents among these emulsifiers. The N-methyl-N-nitrosourea (MNU) administered mouse is a chemically induced retinal degeneration (RD) model with rapid progress rate. AST suspended in Polysorbate 20 was demonstrated to ameliorate the dramatic consequences of MNU on retina architectures and function in several different tests encompassing from electrophysiology to histology and molecular tests. Furthermore, the multi-electrodes array (MEA) was used to detect the firing activities of retinal ganglion cells within the inner retinal circuits. We found that AST nanodispersions could restrain the spontaneous firing response, enhance the light induced firing response, and preserve the basic configurations of visual signal pathway in degenerative retinas. The MEA assay provided an appropriate example to evaluate the potency of pharmacological compounds on retinal plasticity. In summary, emulsifier type affects the basic physicochemical characteristic of AST nanodispersions. Polysorbate 20 acts as an optimized nonionic emulsifier for the efficient delivery of AST nanodispersions to retina. AST nanodispersions can alleviate the photoreceptor loss and rectify the abnormities in visual signal transmission.

New Features of Receptive Fields in Mouse Retina through Spike-triggered Covariance

Retinal ganglion cells (RGCs) encode various spatiotemporal features of visual information into spiking patterns. The receptive field (RF) of each RGC is usually calculated by spike-triggered average (STA), which is fast and easy to understand, but limited to simple and unimodal RFs. As an alternative, spike-triggered covariance (STC) has been proposed to characterize more complex patterns in RFs. This study compares STA and STC for the characterization of RFs and demonstrates that STC has an advantage over STA for identifying novel spatiotemporal features of RFs in mouse RGCs. We first classified mouse RGCs into ON, OFF, and ON/OFF cells according to their response to full-field light stimulus, and then investigated the spatiotemporal patterns of RFs with random checkerboard stimulation, using both STA and STC analysis. We propose five sub-types (T1–T5) in the STC of mouse RGCs together with their physiological implications. In particular, the relatively slow biphasic pattern (T1) could be related to excitatory inputs from bipolar cells. The transient biphasic pattern (T2) allows one to characterize complex patterns in RFs of ON/OFF cells. The other patterns (T3–T5), which are contrasting, alternating, and monophasic patterns, could be related to inhibitory inputs from amacrine cells. Thus, combining STA and STC and considering the proposed sub-types unveil novel characteristics of RFs in the mouse retina and offer a more holistic understanding of the neural coding mechanisms of mouse RGCs.

Functional characterization of retinal ganglion cells using tailored nonlinear modeling

The mammalian retina encodes the visual world in action potentials generated by 20-50 functionally and anatomically-distinct types of retinal ganglion cell (RGC). Individual RGC types receive synaptic input from distinct presynaptic circuits; therefore, their responsiveness to specific features in the visual scene arises from the information encoded in synaptic input and shaped by postsynaptic signal integration and spike generation. Unfortunately, there is a dearth of tools for characterizing the computations reflected in RGC spike output. Therefore, we developed a statistical model, the separable Nonlinear Input Model, to characterize the excitatory and suppressive components of RGC receptive fields. We recorded RGC responses to a correlated noise ("cloud") stimulus in an in vitro preparation of mouse retina and found that our model accurately predicted RGC responses at high spatiotemporal resolution. It identified multiple receptive fields reflecting the main excitatory and suppressive components of the response of each neuron. Significantly, our model accurately identified ON-OFF cells and distinguished their distinct ON and OFF receptive fields, and it demonstrated a diversity of suppressive receptive fields in the RGC population. In total, our method offers a rich description of RGC computation and sets a foundation for relating it to retinal circuitry.

Temporal Neuromodulation of Retinal Ganglion Cells by Low-Frequency Focused Ultrasound Stimulation

Significant progress has been made recently in treating neurological blindness using implantable visual prostheses. However, implantable medical devices are highly invasive and subject to many safety, efficacy, and cost issues. The discovery that ultrasound (US) may be useful as a noninvasive neuromodulation tool has aroused great interest in the field of acoustic retinal prostheses (ARPs). We have investigated the responsiveness of rat retinal ganglion cells (RGCs) to low-frequency focused US stimulation (LFUS) at 2.25 MHz and characterized the neurophysiological properties of US responses by performing in vitro multielectrode array recordings. The results show that LFUS can reliably activate RGCs. The US-induced responses did not correspond to the standard light responses and varied greatly among cell types. Moreover, dual-peak responses to US stimulation were observed that have not been reported previously. The temporal response properties of RGCs, including their latency, firing rate, and response type, were modulated by the acoustic intensity. These findings suggest the presence of a temporal neuromodulation effect of LFUS and potentially open a new avenue in the development of ARP.

The dynamic receptive fields of retinal ganglion cells

Retinal ganglion cells (RGCs) were one of the first classes of sensory neurons to be described in terms of a receptive field (RF). Over the last six decades, our understanding of the diversity of RGC types and the nuances of their response properties has grown exponentially. We will review the current understanding of RGC RFs mostly from studies in mammals, but including work from other vertebrates as well. We will argue for a new paradigm that embraces the fluidity of RGC RFs with an eye toward the neuroethology of vision. Specifically, we will focus on (1) different methods for measuring RGC RFs, (2) RF models, (3) feature selectivity and the distinction between fluid and stable RF properties, and (4) ideas about the future of understanding RGC RFs.

Bioinspired Approach to Modeling Retinal Ganglion Cells Using System Identification Techniques

The processing capabilities of biological vision systems are still vastly superior to artificial vision, even though this has been an active area of research for over half a century. Current artificial vision techniques integrate many insights from biology yet they remain far-off the capabilities of animals and humans in terms of speed, power, and performance. A key aspect to modeling the human visual system is the ability to accurately model the behavior and computation within the retina. In particular, we focus on modeling the retinal ganglion cells (RGCs) as they convey the accumulated data of real world images as action potentials onto the visual cortex via the optic nerve. Computational models that approximate the processing that occurs within RGCs can be derived by quantitatively fitting the sets of physiological data using an input-output analysis where the input is a known stimulus and the output is neuronal recordings. Currently, these input-output responses are modeled using computational combinations of linear and nonlinear models that are generally complex and lack any relevance to the underlying biophysics. In this paper, we illustrate how system identification techniques, which take inspiration from biological systems, can accurately model retinal ganglion cell behavior, and are a viable alternative to traditional linear-nonlinear approaches.

Refinement of Spatial Receptive Fields in the Developing Mouse Lateral Geniculate Nucleus Is Coordinated with Excitatory and Inhibitory Remodeling

Receptive field properties of individual visual neurons are dictated by the precise patterns of synaptic connections they receive, including the arrangement of inputs in visual space and features such as polarity (On vs Off). The inputs from the retina to the lateral geniculate nucleus (LGN) in the mouse undergo significant refinement during development. However, it is unknown how this refinement corresponds to the establishment of functional visual response properties. Here we conducted in vivo and in vitro recordings in the mouse LGN, beginning just after natural eye opening, to determine how receptive fields develop as excitatory and feedforward inhibitory retinal afferents refine. Experiments used both male and female subjects. For in vivo assessment of receptive fields, we performed multisite extracellular recordings in awake mice. Spatial receptive fields at eye-opening were >2 times larger than in adulthood, and decreased in size over the subsequent week. This topographic refinement was accompanied by other spatial changes, such as a decrease in spot size preference and an increase in surround suppression. Notably, the degree of specificity in terms of On/Off and sustained/transient responses appeared to be established already at eye opening and did not change. We performed in vitro recordings of the synaptic responses evoked by optic tract stimulation across the same time period. These recordings revealed a pairing of decreased excitatory and increased feedforward inhibitory convergence, providing a potential mechanism to explain the spatial receptive field refinement.SIGNIFICANCE STATEMENT The development of precise patterns of retinogeniculate connectivity has been a powerful model system for understanding the mechanisms underlying the activity-dependent refinement of sensory systems. Here we link the maturation of spatial receptive field properties in the lateral geniculate nucleus (LGN) to the remodeling of retinal and inhibitory feedforward convergence onto LGN neurons. These findings should thus provide a starting point for testing the cell type-specific plasticity mechanisms that lead to refinement of different excitatory and inhibitory inputs, and for determining the effect of these mechanisms on the establishment of mature receptive fields in the LGN.

Characterization of retinal ganglion cell, horizontal cell, and amacrine cell types expressing the neurotrophic receptor tyrosine kinase Ret

We report the retinal expression pattern of Ret, a receptor tyrosine kinase for the glial derived neurotrophic factor (GDNF) family ligands (GFLs), during development and in the adult mouse. Ret is initially expressed in retinal ganglion cells (RGCs), followed by horizontal cells (HCs) and amacrine cells (ACs), beginning with the early stages of postmitotic development. Ret expression persists in all three classes of neurons in the adult. Using RNA sequencing, immunostaining and random sparse recombination, we show that Ret is expressed in at least three distinct types of ACs, and ten types of RGCs. Using intersectional genetics, we describe the dendritic arbor morphologies of RGC types expressing Ret in combination with each of the three members of the POU4f/Brn3 family of transcription factors. Ret expression overlaps with Brn3a in 4 RGC types, with Brn3b in 5 RGC types, and with Brn3c in one RGC type, respectively. Ret⁺ RGCs project to the lateral geniculate nucleus (LGN), pretectal area (PTA) and superior colliculus (SC), and avoid the suprachiasmatic nucleus and accessory optic system. Brn3a⁺ Ret⁺ and Brn3c⁺ Ret⁺ RGCs project preferentially to contralateral retinorecipient areas, while Brn3b⁺ Ret⁺ RGCs shows minor ipsilateral projections to the olivary pretectal nucleus and the LGN. Our findings establish intersectional genetic approaches for the anatomic and developmental characterization of individual Ret⁺ RGC types. In addition, they provide necessary information for addressing the potential interplay between GDNF neurotrophic signaling and transcriptional regulation in RGC type specification.

Different functional susceptibilities of mouse retinal ganglion cell subtypes to optic nerve crush injury

In optic neuropathies, the progressive deterioration of retinal ganglion cell (RGC) function leads to irreversible vision loss. Increasing experimental evidence suggests differing susceptibility for RGC functional subtypes. Here with multi-electrode array recordings, RGC functional loss was characterized at multiple time points in a mouse model of optic nerve crush. Firing rate, latency of response and receptive field size were analyzed for ON, OFF and ON-OFF RGCs separately. It was observed that responses and receptive fields of OFF cells were impaired earlier than ON cells after the injury. For the ON-OFF cells, the OFF component of response was also more susceptible to optic nerve injury than the ON component. Moreover, more ON transient cells survived than ON sustained cells post the crush, implying a diversified vulnerability for ON cells. Together, these data support the contention that RGCs' functional degeneration in optic nerve injury is subtype dependent, a fact that needs to be considered when developing treatments of glaucomatous retinal ganglion cell degeneration and other optic neuropathies.

Pan-retinal characterisation of Light Responses from Ganglion Cells in the Developing Mouse Retina

We have investigated the ontogeny of light-driven responses in mouse retinal ganglion cells (RGCs). Using a large-scale, high-density multielectrode array, we recorded from hundreds to thousands of RGCs simultaneously at pan-retinal level, including dorsal and ventral locations. Responses to different contrasts not only revealed a complex developmental profile for ON, OFF and ON-OFF responses, but also unveiled differences between dorsal and ventral RGC responses. At eye-opening, dorsal RGCs of all types were more responsive to light, perhaps indicating an environmental priority to nest viewing for pre-weaning pups. The developmental profile of ON and OFF responses exhibited antagonistic behaviour, with the strongest ON responses shortly after eye-opening, followed by an increase in the strength of OFF responses later on. Further, we found that with maturation receptive field (RF) center sizes decrease, spike-triggered averaged responses to white noise become stronger, and centers become more circular while maintaining differences between RGC types. We conclude that the maturation of retinal functionality is not spatially homogeneous, likely reflecting ecological requirements that favour earlier maturation of the dorsal retina.

Graphics Processing Unit-Accelerated Code for Computing Second-Order Wiener Kernels and Spike-Triggered Covariance

Sensory neuroscience seeks to understand and predict how sensory neurons respond to stimuli. Nonlinear components of neural responses are frequently characterized by the second-order Wiener kernel and the closely-related spike-triggered covariance (STC). Recent advances in data acquisition have made it increasingly common and computationally intensive to compute second-order Wiener kernels/STC matrices. In order to speed up this sort of analysis, we developed a graphics processing unit (GPU)-accelerated module that computes the second-order Wiener kernel of a system's response to a stimulus. The generated kernel can be easily transformed for use in standard STC analyses. Our code speeds up such analyses by factors of over 100 relative to current methods that utilize central processing units (CPUs). It works on any modern GPU and may be integrated into many data analysis workflows. This module accelerates data analysis so that more time can be spent exploring parameter space and interpreting data.

Differential stimulation of the retina with subretinally injected exogenous neurotransmitter: A biomimetic alternative to electrical stimulation

Subretinal stimulation of the retina with neurotransmitters, the normal means of conveying visual information, is a potentially better alternative to electrical stimulation widely used in current retinal prostheses for treating blindness from photoreceptor degenerative diseases. Yet, no subretinal electrical or chemical stimulation study has stimulated the OFF and ON pathways differentially through inner retinal activation. Here, we demonstrate the feasibility of differentially stimulating retinal ganglion cells (RGCs) through the inner nuclear layer of the retina with glutamate, a primary neurotransmitter chemical, in a biomimetic way. We show that controlled pulsatile delivery of glutamate into the subsurface of explanted wild-type rat retinas elicits highly localized simultaneous inhibitory and excitatory spike rate responses in OFF and ON RGCs. We also present the spatiotemporal characteristics of RGC responses to subretinally injected glutamate and the therapeutic stimulation parameters. Our findings could pave the way for future development of a neurotransmitter-based subretinal prosthesis offering more naturalistic vision and better visual acuity than electrical prostheses.

Transplanted neurons integrate into adult retinas and respond to light

Retinal ganglion cells (RGCs) degenerate in diseases like glaucoma and are not replaced in adult mammals. Here we investigate whether transplanted RGCs can integrate into the mature retina. We have transplanted GFP-labelled RGCs into uninjured rat retinas in vivo by intravitreal injection. Transplanted RGCs acquire the general morphology of endogenous RGCs, with axons orienting towards the optic nerve head of the host retina and dendrites growing into the inner plexiform layer. Preliminary data show in some cases GFP(+) axons extending within the host optic nerves and optic tract, reaching usual synaptic targets in the brain, including the lateral geniculate nucleus and superior colliculus. Electrophysiological recordings from transplanted RGCs demonstrate the cells' electrical excitability and light responses similar to host ON, ON-OFF and OFF RGCs, although less rapid and with greater adaptation. These data present a promising approach to develop cell replacement strategies in diseased retinas with degenerating RGCs.

Increasing Spontaneous Retinal Activity before Eye Opening Accelerates the Development of Geniculate Receptive Fields

Visually evoked activity is necessary for the normal development of the visual system. However, little is known about the capacity for patterned spontaneous activity to drive the maturation of receptive fields before visual experience. Retinal waves provide instructive retinotopic information for the anatomical organization of the visual thalamus. To determine whether retinal waves also drive the maturation of functional responses, we increased the frequency of retinal waves pharmacologically in the ferret (Mustela putorius furo) during a period of retinogeniculate development before eye opening. The development of geniculate receptive fields after receiving these increased neural activities was measured using single-unit electrophysiology. We found that increased retinal waves accelerate the developmental reduction of geniculate receptive field sizes. This reduction is due to a decrease in receptive field center size rather than an increase in inhibitory surround strength. This work reveals an instructive role for patterned spontaneous activity in guiding the functional development of neural circuits.

Understanding spike-triggered covariance using Wiener theory for receptive field identification

Receptive field identification is a vital problem in sensory neurophysiology and vision. Much research has been done in identifying the receptive fields of nonlinear neurons whose firing rate is determined by the nonlinear interactions of a small number of linear filters. Despite more advanced methods that have been proposed, spike-triggered covariance (STC) continues to be the most widely used method in such situations due to its simplicity and intuitiveness. Although the connection between STC and Wiener/Volterra kernels has often been mentioned in the literature, this relationship has never been explicitly derived. Here we derive this relationship and show that the STC matrix is actually a modified version of the second-order Wiener kernel, which incorporates the input autocorrelation and mixes first- and second-order dynamics. It is then shown how, with little modification of the STC method, the Wiener kernels may be obtained and, from them, the principal dynamic modes, a set of compact and efficient linear filters that essentially combine the spike-triggered average and STC matrix and generalize to systems with both continuous and point-process outputs. Finally, using Wiener theory, we show how these obtained filters may be corrected when they were estimated using correlated inputs. Our correction technique is shown to be superior to those commonly used in the literature for both correlated Gaussian images and natural images.

Spike-Triggered Covariance Analysis Reveals Phenomenological Diversity of Contrast Adaptation in the Retina

When visual contrast changes, retinal ganglion cells adapt by adjusting their sensitivity as well as their temporal filtering characteristics. The latter has classically been described by contrast-induced gain changes that depend on temporal frequency. Here, we explored a new perspective on contrast-induced changes in temporal filtering by using spike-triggered covariance analysis to extract multiple parallel temporal filters for individual ganglion cells. Based on multielectrode-array recordings from ganglion cells in the isolated salamander retina, we found that contrast adaptation of temporal filtering can largely be captured by contrast-invariant sets of filters with contrast-dependent weights. Moreover, differences among the ganglion cells in the filter sets and their contrast-dependent contributions allowed us to phenomenologically distinguish three types of filter changes. The first type is characterized by newly emerging features at higher contrast, which can be reproduced by computational models that contain response-triggered gain-control mechanisms. The second type follows from stronger adaptation in the Off pathway as compared to the On pathway in On-Off-type ganglion cells. Finally, we found that, in a subset of neurons, contrast-induced filter changes are governed by particularly strong spike-timing dynamics, in particular by pronounced stimulus-dependent latency shifts that can be observed in these cells. Together, our results show that the contrast dependence of temporal filtering in retinal ganglion cells has a multifaceted phenomenology and that a multi-filter analysis can provide a useful basis for capturing the underlying signal-processing dynamics.

Our sensory systems have to process stimuli under a wide range of environmental conditions. To cope with this challenge, the involved neurons adapt by adjusting their signal processing to the recently encountered intensity range. In the visual system, one finds, for example, that higher visual contrast leads to changes in how visual signals are temporally filtered, making signal processing faster and more band-pass-like at higher contrast. By analyzing signals from neurons in the retina of salamanders, we here found that these adaptation effects can be described by a fixed set of filters, independent of contrast, whose relative contributions change with contrast. Also, we found that different phenomena contribute to this adaptation. In particular, some cells change their relative sensitivity to light increments and light decrements, whereas other cells are influenced by a strong contrast-dependence of the exact timing of their responses. Our results show that contrast adaptation in the retina is not an entirely homogeneous phenomenon, and that models with multiple filters can help in characterizing sensory adaptation.

Progressive degeneration of retinal and superior collicular functions in mice with sustained ocular hypertension

PURPOSE

We investigated the progressive degeneration of retinal and superior collicular functions in a mouse model of sustained ocular hypertension.

METHODS

Focal laser illumination and injection of polystyrene microbeads were used to induce chronic ocular hypertension. Retinal ganglion cell (RGC) loss was characterized by in vivo optical coherence tomography (OCT) and immunohistochemistry. Retinal dysfunction was also monitored by the full-field ERG. Retinal ganglion cell light responses were recorded using a 256-channel multielectrode array (MEA), and RGC subtypes were characterized by noncentered spike-triggered covariance (STC-NC) analysis. Single-unit extracellular recordings from superficial layers of the superior colliculus (SC) were performed to examine the receptive field (RF) properties of SC neurons.

RESULTS

The elevation of intraocular pressure (IOP) lasted 4 months in mice treated with a combination of laser photocoagulation and microbead injection. Progressive RGC loss and functional degeneration were confirmed in ocular hypertensive (OHT) mice. These mice had fewer visually responsive RGCs than controls. Using the STC-NC analysis, we classified RGCs into ON, OFF, and ON-OFF functional subtypes. We showed that ON and OFF RGCs were more susceptible to the IOP elevation than ON-OFF RGCs. Furthermore, SC neurons of OHT mice had weakened responses to visual stimulation and exhibited mismatched ON and OFF subfields and irregular RF structure.

CONCLUSIONS

We demonstrated that the functional degeneration of RGCs is subtype-dependent and that the ON and OFF pathways from the retina to the SC were disrupted. Our study provides a foundation to investigate the mechanisms underlying the progressive vision loss in experimental glaucoma.

Rapid mapping of visual receptive fields by filtered back projection: application to multi-neuronal electrophysiology and imaging

Neurons in the visual system vary widely in the spatiotemporal properties of their receptive fields (RFs), and understanding these variations is key to elucidating how visual information is processed. We present a new approach for mapping RFs based on the filtered back projection (FBP), an algorithm used for tomographic reconstructions. To estimate RFs, a series of bars were flashed across the retina at pseudo-random positions and at a minimum of five orientations. We apply this method to retinal neurons and show that it can accurately recover the spatial RF and impulse response of ganglion cells recorded on a multi-electrode array. We also demonstrate its utility for in vivo imaging by mapping the RFs of an array of bipolar cell synapses expressing a genetically encoded Ca²⁺ indicator. We find that FBP offers several advantages over the commonly used spike-triggered average (STA): (i) ON and OFF components of a RF can be separated; (ii) the impulse response can be reconstructed at sample rates of 125 Hz, rather than the refresh rate of a monitor; (iii) FBP reveals the response properties of neurons that are not evident using STA, including those that display orientation selectivity, or fire at low mean spike rates; and (iv) the FBP method is fast, allowing the RFs of all the bipolar cell synaptic terminals in a field of view to be reconstructed in under 4 min. Use of the FBP will benefit investigations of the visual system that employ electrophysiology or optical reporters to measure activity across populations of neurons.

Subtype-dependent postnatal development of direction- and orientation-selective retinal ganglion cells in mice

The direction-selective ganglion cells (DSGCs) and orientation-selective ganglion cells (OSGCs) encode the directional and the orientational information of a moving object, respectively. It is unclear how DSGCs and OSGCs mature in the mouse retina during postnatal development. Here we investigated the development of DSGCs and OSGCs after eye-opening. We show that 1) DSGCs and OSGCs are present at postnatal day 12 (P12), just before eye-opening; 2) the fractions of both DSGCs and OSGCs increase from P12 to P30; 3) the development of DSGCs and OSGCs is subtype dependent; and 4) direction and orientation selectivity are two separate features of retinal ganglion cells (RGCs) in the mouse retina. We classified RGCs into different functional subtypes based on their light response properties. Compared with P12, the direction and orientation selectivity of ON-OFF RGCs but not ON RGCs became stronger at P30. The tuning width of DSGCs for both ON and ON-OFF subtypes decreased with age. For OSGCs, we divided them into non-direction-selective (non-DS) OSGCs and direction-selective OSGCs (DS&OSGCs). For DS&OSGCs, we found that there was no correlation between the direction and orientation selectivity, and that the tuning width of both ON and ON-OFF subtypes remained unchanged with age. For non-DS OSGCs, the tuning width of ON but not ON-OFF subtype decreased with development. These findings provide a foundation to reveal the molecular and synaptic mechanisms underlying the development of the direction and orientation selectivity in the retina.

Dark rearing alters the normal development of spatiotemporal response properties but not of contrast detection threshold in mouse retinal ganglion cells

The mouse visual system is immature when the eyes open two weeks after birth. As in other mammals, some of the maturation that occurs in the subsequent weeks is known to depend on visual experience. Development of the retina, which as the first stage of vision provides the visual information to the brain, also depends on light-driven activity for proper development but has been less well studied than visual cortical development. The critical properties for retinal encoding of images include detection of contrast and responsiveness to the broad range of temporal stimulus frequencies present in natural stimuli. Here we show that contrast detection threshold and temporal frequency response characteristics of ON and OFF retinal ganglion cells (RGCs), which are poor at eye opening, subsequently undergo maturation, improving RGC performance. Further, we find that depriving mice of visual experience from before birth by rearing them in the dark causes ON and OFF RGCs to have smaller receptive field centers but does not affect their contrast detection threshold development. The modest developmental increase in temporal frequency responsiveness of RGCs in mice reared on a normal light cycle was inhibited by dark rearing only in ON but not OFF RGCs. Thus, these RGC response characteristics are in many ways unaffected by the experience-dependent changes to synaptic and spontaneous activity known to occur in the mouse retina in the two weeks after eye opening, but specific differences are apparent in the ON vs. OFF RGC populations.

Differential progression of structural and functional alterations in distinct retinal ganglion cell types in a mouse model of glaucoma

Intraocular pressure (IOP) elevation is a principal risk factor for glaucoma. Using a microbead injection technique to chronically raise IOP for 15 or 30 d in mice, we identified the early changes in visual response properties of different types of retinal ganglion cells (RGCs) and correlated these changes with neuronal morphology before cell death. Microbead-injected eyes showed reduced optokinetic tracking as well as cell death. In such eyes, multielectrode array recordings revealed that four RGC types show diverse alterations in their light responses upon IOP elevation. OFF-transient RGCs exhibited a more rapid decline in both structural and functional organizations compared with other RGCs. In contrast, although the light-evoked responses of OFF-sustained RGCs were perturbed, the dendritic arbor of this cell type remained intact. ON-transient and ON-sustained RGCs had normal functional receptive field sizes but their spontaneous and light-evoked firing rates were reduced. ON- and OFF-sustained RGCs lost excitatory synapses across an otherwise structurally normal dendritic arbor. Together, our observations indicate that there are changes in spontaneous activity and light-evoked responses in RGCs before detectable dendritic loss. However, when dendrites retract, we found corresponding changes in receptive field center size. Importantly, the effects of IOP elevation are not uniformly manifested in the structure and function of diverse RGC populations, nor are distinct RGC types perturbed within the same time-frame by such a challenge.

Orientation-selective responses in the mouse lateral geniculate nucleus

The dorsal lateral geniculate nucleus (dLGN) receives visual information from the retina and transmits it to the cortex. In this study, we made extracellular recordings in the dLGN of both anesthetized and awake mice, and found that a surprisingly high proportion of cells were selective for stimulus orientation. The orientation selectivity of dLGN cells was unchanged after silencing the visual cortex pharmacologically, indicating that it is not due to cortical feedback. The orientation tuning of some dLGN cells correlated with their elongated receptive fields, while in others orientation selectivity was observed despite the fact that their receptive fields were circular, suggesting that their retinal input might already be orientation selective. Consistently, we revealed orientation/axis-selective ganglion cells in the mouse retina using multielectrode arrays in an in vitro preparation. Furthermore, the orientation tuning of dLGN cells was largely maintained at different stimulus contrasts, which could be sufficiently explained by a simple linear feedforward model. We also compared the degree of orientation selectivity in different visual structures under the same recording condition. Compared with the dLGN, orientation selectivity is greatly improved in the visual cortex, but is similar in the superior colliculus, another major retinal target. Together, our results demonstrate prominent orientation selectivity in the mouse dLGN, which may potentially contribute to visual processing in the cortex.

Inferring nonlinear neuronal computation based on physiologically plausible inputs

The computation represented by a sensory neuron's response to stimuli is constructed from an array of physiological processes both belonging to that neuron and inherited from its inputs. Although many of these physiological processes are known to be nonlinear, linear approximations are commonly used to describe the stimulus selectivity of sensory neurons (i.e., linear receptive fields). Here we present an approach for modeling sensory processing, termed the Nonlinear Input Model (NIM), which is based on the hypothesis that the dominant nonlinearities imposed by physiological mechanisms arise from rectification of a neuron's inputs. Incorporating such 'upstream nonlinearities' within the standard linear-nonlinear (LN) cascade modeling structure implicitly allows for the identification of multiple stimulus features driving a neuron's response, which become directly interpretable as either excitatory or inhibitory. Because its form is analogous to an integrate-and-fire neuron receiving excitatory and inhibitory inputs, model fitting can be guided by prior knowledge about the inputs to a given neuron, and elements of the resulting model can often result in specific physiological predictions. Furthermore, by providing an explicit probabilistic model with a relatively simple nonlinear structure, its parameters can be efficiently optimized and appropriately regularized. Parameter estimation is robust and efficient even with large numbers of model components and in the context of high-dimensional stimuli with complex statistical structure (e.g. natural stimuli). We describe detailed methods for estimating the model parameters, and illustrate the advantages of the NIM using a range of example sensory neurons in the visual and auditory systems. We thus present a modeling framework that can capture a broad range of nonlinear response functions while providing physiologically interpretable descriptions of neural computation.

Identifying functional bases for multidimensional neural computations

Current dimensionality-reduction methods can identify relevant subspaces for neural computations but do not favor one basis over the other within the relevant subspace. Finding the appropriate basis can simplify the description of the nonlinear computation with respect to the relevant variables, making it easier to elucidate the underlying neural computation and make hypotheses about the neural circuitry, giving rise to the observed responses. Part of the problem is that although some of the dimensionality reduction methods can identify many of the relevant dimensions, it is usually difficult to map out or interpret the nonlinear transformation with respect to more than a few relevant dimensions simultaneously without some simplifying assumptions. While recent approaches make it possible to create predictive models based on many relevant dimensions simultaneously, there still remains the need to relate such predictive models to the mechanistic descriptions of the operation of underlying neural circuitry. Here we demonstrate that transforming to a basis within the relevant subspace where the neural computation is best described by a given nonlinear function often makes it easier to interpret the computation and describe it with a small number of parameters. We refer to the corresponding basis as the functional basis, and illustrate the utility of such transformation in the context of logical OR and logical AND functions. We show that although dimensionality-reduction methods such as spike-triggered covariance are able to find a relevant subspace, they often produce dimensions that are difficult to interpret and do not correspond to a functional basis. The functional features can be found using a maximum likelihood approach. The results are illustrated using simulated neurons and recordings from retinal ganglion cells. The resulting features are uniquely defined and nonorthogonal, and they make it easier to relate computational and mechanistic models to each other.

Sustained ocular hypertension induces dendritic degeneration of mouse retinal ganglion cells that depends on cell type and location

PURPOSE

Glaucoma is characterized by retinal ganglion cell (RGC) death and frequently associated with elevated IOP. How RGCs degenerate before death is little understood, so we sought to investigate RGC degeneration in a mouse model of ocular hypertension.

METHODS

A laser-induced mouse model of chronic ocular hypertension mimicked human high-tension glaucoma. Immunohistochemistry was used to characterize overall RGC loss and an optomotor behavioral test to measure corresponding changes in visual capacity. Changes in RGC functional properties were characterized by a large-scale multielectrode array (MEA). The transgenic Thy-1-YFP mouse line, in which a small number of RGCs are labeled with yellow fluorescent protein (YFP), permitted investigation of whether subtypes of RGCs or RGCs from particular retinal areas were differentially vulnerable to elevated IOP.

RESULTS

Sustained IOP elevation in mice was achieved by laser photocoagulation. We confirmed RGC loss and decreased visual acuity in ocular hypertensive mice. Furthermore, these mice had fewer visually responsive cells with smaller receptive field sizes compared to controls. We demonstrated that RGC dendritic shrinkage started from the vertical axis of hypertensive eyes and that mono-laminated ON cells were more susceptible to IOP elevation than bi-laminated ON-OFF cells. Moreover, a subgroup of ON RGCs labeled by the SMI-32 antibody exhibited significant dendritic atrophy in the superior quadrant of the hypertensive eyes.

CONCLUSIONS

RGC degeneration depends on subtype and location in hypertensive eyes. This study introduces a valuable model to investigate how the structural and functional degeneration of RGCs leads to visual impairments.

Spike-triggered covariance: geometric proof, symmetry properties, and extension beyond Gaussian stimuli

The space of sensory stimuli is complex and high-dimensional. Yet, single neurons in sensory systems are typically affected by only a small subset of the vast space of all possible stimuli. A proper understanding of the input–output transformation represented by a given cell therefore requires the identification of the subset of stimuli that are relevant in shaping the neuronal response. As an extension to the commonly-used spike-triggered average, the analysis of the spike-triggered covariance matrix provides a systematic methodology to detect relevant stimuli. As originally designed, the consistency of this method is guaranteed only if stimuli are drawn from a Gaussian distribution. Here we present a geometric proof of consistency, which provides insight into the foundations of the method, in particular, into the crucial role played by the geometry of stimulus space and symmetries in the stimulus–response relation. This approach leads to a natural extension of the applicability of the spike-triggered covariance technique to arbitrary spherical or elliptic stimulus distributions. The extension only requires a subtle modification of the original prescription. Furthermore, we present a new resampling method for assessing statistical significance of identified relevant stimuli, applicable to spherical and elliptic stimulus distributions. Finally, we exemplify the modified method and compare it to other prescriptions given in the literature.

Second order dimensionality reduction using minimum and maximum mutual information models

Conventional methods used to characterize multidimensional neural feature selectivity, such as spike-triggered covariance (STC) or maximally informative dimensions (MID), are limited to Gaussian stimuli or are only able to identify a small number of features due to the curse of dimensionality. To overcome these issues, we propose two new dimensionality reduction methods that use minimum and maximum information models. These methods are information theoretic extensions of STC that can be used with non-Gaussian stimulus distributions to find relevant linear subspaces of arbitrary dimensionality. We compare these new methods to the conventional methods in two ways: with biologically-inspired simulated neurons responding to natural images and with recordings from macaque retinal and thalamic cells responding to naturalistic time-varying stimuli. With non-Gaussian stimuli, the minimum and maximum information methods significantly outperform STC in all cases, whereas MID performs best in the regime of low dimensional feature spaces.

Neurons are capable of simultaneously encoding information about multiple features of sensory stimuli in their spikes. The dimensionality reduction methods that currently exist to extract those relevant features are either biased for non-Gaussian stimuli or fall victim to the curse of dimensionality. In this paper we introduce two information theoretic extensions of the spike-triggered covariance method. These new methods use the concepts of minimum and maximum mutual information to identify the stimulus features encoded in the spikes of a neuron. Using simulated and experimental neural data, these methods are shown to perform well both in situations where conventional approaches are appropriate and where they fail. These new techniques should improve the characterization of neural feature selectivity in areas of the brain where the application of currently available approaches is restricted.

Receptive field center size decreases and firing properties mature in ON and OFF retinal ganglion cells after eye opening in the mouse

Development of the mammalian visual system is not complete at birth but continues postnatally well after eye opening. Although numerous studies have revealed changes in the development of the thalamus and visual cortex during this time, less is known about the development of response properties of retinal ganglion cells (RGCs). Here, we mapped functional receptive fields of mouse RGCs using a Gaussian white noise checkerboard stimulus and a multielectrode array to record from retinas at eye opening, 3 days later, and 4 wk after birth, when visual responses are essentially mature. Over this time, the receptive field center size of ON and OFF RGC populations decreased. The average receptive field center size of ON RGCs was larger than that of OFF RGCs at eye opening, but they decreased to the same size in the adult. Firing properties were also immature at eye opening. RGCs had longer latencies, lower frequencies of firing, and lower sensitivity than in the adult. Hence, the dramatic maturation of the visual system during the first weeks of visual experience includes the retina.