Efficient coding of natural scenes improves neural system identification.

Neural system identification aims at learning the response function of neurons to arbitrary stimuli using experimentally recorded data, but typically does not leverage normative principles such as efficient coding of natural environments. Visual systems, however, have evolved to efficiently process input from the natural environment. Here, we present a normative network regularization for system identification models by incorporating, as a regularizer, the efficient coding hypothesis, which states that neural response properties of sensory representations are strongly shaped by the need to preserve most of the stimulus information with limited resources. Using this approach, we explored if a system identification model can be improved by sharing its convolutional filters with those of an autoencoder which aims to efficiently encode natural stimuli. To this end, we built a hybrid model to predict the responses of retinal neurons to noise stimuli. This approach did not only yield a higher performance than the "stand-alone" system identification model, it also produced more biologically plausible filters, meaning that they more closely resembled neural representation in early visual systems. We found these results applied to retinal responses to different artificial stimuli and across model architectures. Moreover, our normatively regularized model performed particularly well in predicting responses of direction-of-motion sensitive retinal neurons. The benefit of natural scene statistics became marginal, however, for predicting the responses to natural movies. In summary, our results indicate that efficiently encoding environmental inputs can improve system identification models, at least for noise stimuli, and point to the benefit of probing the visual system with naturalistic stimuli.

Inhibition decorrelates visual feature representations in the inner retina.

The retina extracts visual features for transmission to the brain. Different types of bipolar cell split the photoreceptor input into parallel channels and provide the excitatory drive for downstream visual circuits. Mouse bipolar cell types have been described at great anatomical and genetic detail, but a similarly deep understanding of their functional diversity is lacking. Here, by imaging light-driven glutamate release from more than 13,000 bipolar cell axon terminals in the intact retina, we show that bipolar cell functional diversity is generated by the interplay of dendritic excitatory inputs and axonal inhibitory inputs. The resulting centre and surround components of bipolar cell receptive fields interact to decorrelate bipolar cell output in the spatial and temporal domains. Our findings highlight the importance of inhibitory circuits in generating functionally diverse excitatory pathways and suggest that decorrelation of parallel visual pathways begins as early as the second synapse of the mouse visual system.

Retinal bipolar cells: elementary building blocks of vision.

Retinal bipolar cells are the first 'projection neurons' of the vertebrate visual system­all of the information needed for vision is relayed by this intraretinal connection. Each of the at least 13 distinct types of bipolar cells systematically transforms the photoreceptor input in a different way, thereby generating specific channels that encode stimulus properties, such as polarity, contrast, temporal profile and chromatic composition. As a result, bipolar cell output signals represent elementary 'building blocks' from which the microcircuits of the inner retina derive a feature-oriented description of the visual world.

Calcium dynamics change in degenerating cone photoreceptors.

Cone photoreceptors (cones) are essential for high-resolution daylight vision and colour perception. Loss of cones in hereditary retinal diseases has a dramatic impact on human vision. The mechanisms underlying cone death are poorly understood, and consequently, there are no treatments available. Previous studies suggest a central role for calcium (Ca2+) homeostasis deficits in photoreceptor degeneration; however, direct evidence for this is scarce and physiological measurements of Ca2+ in degenerating mammalian cones are lacking.Here, we took advantage of the transgenic HR2.1:TN-XL mouse line that expresses a genetically encoded Ca2+ biosensor exclusively in cones. We cross-bred this line with mouse models for primary ("cone photoreceptor function loss-1", cpfl1) and secondary ("retinal degeneration-1", rd1) cone degeneration, respectively, and assessed resting Ca2+ levels and light-evoked Ca2+ responses in cones using two-photon imaging. We found that Ca2+ dynamics were altered in cpfl1 cones, showing higher noise and variable Ca2+ levels, with significantly wider distribution than for wild-type and rd1 cones. Unexpectedly, up to 21% of cpfl1 cones still displayed light-evoked Ca2+ responses, which were larger and slower than wild-type responses. In contrast, genetically intact rd1 cones were characterized by lower noise and complete lack of visual function.Our study demonstrates alterations in cone Ca2+ dynamics in both primary and secondary cone degeneration. Our results are consistent with the view that higher (fluctuating) cone Ca2+ levels are involved in photoreceptor cell death in primary (cpfl1) but not in secondary (rd1) cone degeneration. These findings may guide the future development of therapies targeting photoreceptor Ca2+ homeostasis.

How do horizontal cells 'talk' to cone photoreceptors? Different levels of complexity at the cone-horizontal cell synapse.

The first synapse of the retina plays a fundamental role in the visual system. Due to its importance, it is critical that it encodes information from the outside world with the greatest accuracy and precision possible. Cone photoreceptor axon terminals contain many individual synaptic sites, each represented by a presynaptic structure called a 'ribbon'. These synapses are both highly sophisticated and conserved. Each ribbon relays the light signal to one ON cone bipolar cell and several OFF cone bipolar cells, while two dendritic processes from a GABAergic interneuron, the horizontal cell, modulate the cone output via parallel feedback mechanisms. The presence of these three partners within a single synapse has raised numerous questions, and its anatomical and functional complexity is still only partially understood. However, the understanding of this synapse has recently evolved, as a consequence of progress in understanding dendritic signal processing and its role in facilitating global versus local signalling. Indeed, for the downstream retinal network, dendritic processing in horizontal cells may be essential, as they must support important functional operations such as contrast enhancement, which requires spatial averaging of the photoreceptor array, while at the same time preserving accurate spatial information. Here, we review recent progress made towards a better understanding of the cone synapse, with an emphasis on horizontal cell function, and discuss why such complexity might be necessary for early visual processing.

Connectomics of synaptic microcircuits: lessons from the outer retina.

Photoreceptors form a sophisticated synaptic complex with bipolar and horizontal cells, transmitting the signals generated by the phototransduction cascade to downstream retinal circuitry. The cone photoreceptor synapse shows several characteristic anatomical connectivity motifs that shape signal transfer: typically, ON-cone bipolar cells receive photoreceptor input through invaginating synapses; OFF-cone bipolar cells form basal synapses with photoreceptors. Both ON- and OFF-cone bipolar cells are believed to sample from all cone photoreceptors within their dendritic span. Electron microscopy and immunolabelling studies have established the robustness of these motifs, but have been limited by trade-offs in sample size and spatial resolution, respectively, constraining precise quantitative investigation to a few individual cells. 3D-serial electron microscopy overcomes these limitations and has permitted complete sets of neurons to be reconstructed over a comparatively large section of retinal tissue. Although the published mouse dataset lacks labels for synaptic structures, the characteristic anatomical motifs at the photoreceptor synapse can be exploited to identify putative synaptic contacts, which has enabled the development of a quantitative description of outer retinal connectivity. This revealed unexpected exceptions to classical motifs, including substantial interaction between rod and cone pathways at the photoreceptor synapse, sparse photoreceptor sampling and atypical contacts. Here, we summarize what was learned from this study in a more general context: we consider both the implications and limitations of the study and identify promising avenues for future research.

Deletion of myosin VI causes slow retinal optic neuropathy and age-related macular degeneration (AMD)-relevant retinal phenotype.

The unconventional myosin VI, a member of the actin-based motor protein family of myosins, is expressed in the retina. Its deletion was previously shown to reduce amplitudes of the a- and b-waves of the electroretinogram. Analyzing wild-type and myosin VI-deficient Snell's Waltzer mice in more detail, the expression pattern of myosin VI in retinal pigment epithelium, outer limiting membrane, and outer plexiform layer could be linked with differential progressing ocular deficits. These encompassed reduced a-waves and b-waves and disturbed oscillatory potentials in the electroretinogram, photoreceptor cell death, retinal microglia infiltration, and formation of basal laminar deposits. A phenotype comprising features of glaucoma (neurodegeneration) and age-related macular degeneration could thus be uncovered that suggests dysfunction of myosin VI and its variable cargo adaptor proteins for membrane sorting and autophagy, as possible candidate mediators for both disease forms.

Differential regulation of cone calcium signals by different horizontal cell feedback mechanisms in the mouse retina.

Controlling neurotransmitter release by modulating the presynaptic calcium level is a key mechanism to ensure reliable signal transmission from one neuron to the next. In this study, we investigated how the glutamatergic output of cone photoreceptors (cones) in the mouse retina is shaped by different feedback mechanisms from postsynaptic GABAergic horizontal cells (HCs) using a combination of two-photon calcium imaging and pharmacology at the level of individual cone axon terminals. We provide evidence that hemichannel-mediated (putative ephaptic) feedback sets the cone output gain by defining the basal calcium level, a mechanism that may be crucial for adapting cones to the ambient light level. In contrast, pH-mediated feedback did not modulate the cone basal calcium level but affected the size and shape of light-evoked cone calcium signals in a contrast-dependent way: low-contrast light responses were amplified, whereas high-contrast light responses were reduced. Finally, we provide functional evidence that GABA shapes light-evoked calcium signals in cones. Because we could not localize ionotropic GABA receptors on cone axon terminals using electron microscopy, we suggest that GABA may act through GABA autoreceptors on HCs, thereby possibly modulating hemichannel- and/or pH-mediated feedback. Together, our results suggest that at the cone synapse, hemichannel-mediated (ephaptic) and pH-mediated feedback fulfill distinct functions to adjust the output of cones to changing ambient light levels and stimulus contrasts and that the efficacy of these feedback mechanisms is likely modulated by GABA release in the outer retina.

BK channels mediate pathway-specific modulation of visual signals in the in vivo mouse retina.

The modulatory role of large-conductance Ca(2+)-activated K(+) (BK) channels in the nervous system has been extensively studied. In the retina, it has been shown that BK channels play a pivotal role in modulating feedback from A17 amacrine cells to rod bipolar cells (RBCs). Here, we used electroretinography to examine the functional role of BK channels for rod and cone vision in the retina in vivo using a genetically engineered mouse lacking functional BK channels (Bk(-/-)). Under dark-adapted and light-adapted conditions, the lack of BK channels had no effect on photoreceptor activity, suggesting that these ion channels do not modulate photoreceptor responses. At the bipolar cell level, the ERG signals attributed to RBCs in Bk(-/-) mice were not different from those in wild-type mice at low scotopic stimulus intensities. However, at high scotopic and low mesopic stimulus intensities, close to RBC saturation, a significant reduction of ERG signals reflecting RBC activity was present in the Bk(-/-) retina. At higher mesopic stimulus intensities activating both RBCs and cone bipolar cells (CBCs), no difference in ERG signals between Bk(-/-) and wild-type mice was found. In photopic stimulus paradigms, activity of ON- and OFF-CBCs in Bk(-/-) and wild-type retinae was indistinguishable. These findings demonstrate that BK channels modulate visual responses in vivo at the bipolar cell level at intermediate stimulus conditions.

Light-driven calcium signals in mouse cone photoreceptors.

Calcium mediates various neuronal functions. The complexity of neuronal Ca²âº signaling is well exemplified by retinal cone photoreceptors, which, with their distinct compartmentalization, offer unique possibilities for studying the diversity of Ca²âº functions in a single cell. Measuring subcellular Ca²âº signals in cones under physiological conditions is not only fundamental for understanding cone function, it also bears important insights into pathophysiological processes governing retinal neurodegeneration. However, due to the proximity of light-sensitive outer segments to other cellular compartments, optical measurements of light-evoked Ca²âº responses in cones are challenging. We addressed this problem by generating a transgenic mouse (HR2.1:TN-XL) in which both short- and middle-wavelength-sensitive cones selectively express the genetically encoded ratiometric Ca²âº biosensor TN-XL. We show that HR2.1:TN-XL allows recording of light-evoked Ca²âº responses using two-photon imaging in individual cone photoreceptor terminals and to probe phototransduction and its diverse regulatory mechanisms with pharmacology at subcellular resolution. To further test this system, we asked whether the classical, nitric oxide (NO)-soluble guanylyl-cyclase (sGC)-cGMP pathway could modulate Ca²âº in cone terminals. Surprisingly, NO reduced Ca²âº resting levels in mouse cones, without evidence for direct sGC involvement. In conclusion, HR2.1:TN-XL mice offer unprecedented opportunities to elucidate light-driven Ca²âº dynamics and their (dys)regulation in cone photoreceptors.

Spike desensitisation as a mechanism for high-contrast selectivity in retinal ganglion cells.

In the vertebrate retina, several dozens of parallel channels relay information about the visual world to the brain. These channels are represented by the different types of retinal ganglion cells (RGCs), whose responses are rendered selective for distinct sets of visual features by various mechanisms. These mechanisms can be roughly grouped into synaptic interactions and cell-intrinsic mechanisms, with the latter including dendritic morphology as well as ion channel complement and distribution. Here, we investigate how strongly ion channel complement can shape RGC output by comparing two mouse RGC types, the well-described ON alpha cell and a little-studied ON cell that is EGFP-labelled in the Igfbp5 mouse line and displays an unusual selectivity for stimuli with high contrast. Using patch-clamp recordings and computational modelling, we show that a higher activation threshold and a pronounced slow inactivation of the voltage-gated Na+ channels contribute to the distinct contrast tuning and transient responses in ON Igfbp5 RGCs, respectively. In contrast, such a mechanism could not be observed in ON alpha cells. This study provides an example for the powerful role that the last stage of retinal processing can play in shaping RGC responses.

Super-resolution STED imaging in the inner and outer whole-mount mouse retina.

Since its invention, super-resolution microscopy has become a popular tool for advanced imaging of biological structures, allowing visualisation of subcellular structures at a spatial scale below the diffraction limit. Thus, it is not surprising that recently, different super-resolution techniques are being applied in neuroscience, e.g. to resolve the clustering of neurotransmitter receptors and protein complex composition in presynaptic terminals. Still, the vast majority of these experiments were carried out either in cell cultures or very thin tissue sections, while there are only a few examples of super-resolution imaging in deeper layers (30 - 50 µm) of biological samples. In that context, the mammalian whole-mount retina has rarely been studied with super-resolution microscopy. Here, we aimed at establishing a stimulated-emission-depletion (STED) microscopy protocol for imaging whole-mount retina. To this end, we developed sample preparation including horizontal slicing of retinal tissue, an immunolabeling protocol with STED-compatible fluorophores and optimised the image acquisition settings. We labelled subcellular structures in somata, dendrites, and axons of retinal ganglion cells in the inner mouse retina. By measuring the full width at half maximum of the thinnest filamentous structures in our preparation, we achieved a resolution enhancement of two or higher compared to conventional confocal images. When combined with horizontal slicing of the retina, these settings allowed visualisation of putative GABAergic horizontal cell synapses in the outer retina. Taken together, we successfully established a STED protocol for reliable super-resolution imaging in the whole-mount mouse retina at depths between 30 and 50 µm, which enables investigating, for instance, protein complex composition and cytoskeletal ultrastructure at retinal synapses in health and disease.

Minimal genetically encoded tags for fluorescent protein labeling in living neurons.

Modern light microscopy, including super-resolution techniques, has brought about a demand for small labeling tags that bring the fluorophore closer to the target. This challenge can be addressed by labeling unnatural amino acids (UAAs) with bioorthogonal click chemistry. The minimal size of the UAA and the possibility to couple the fluorophores directly to the protein of interest with single-residue precision in living cells make click labeling unique. Here, we establish click labeling in living primary neurons and use it for fixed-cell, live-cell, dual-color pulse-chase, and super-resolution microscopy of neurofilament light chain (NFL). We also show that click labeling can be combined with CRISPR/Cas9 genome engineering for tagging endogenous NFL. Due to its versatile nature and compatibility with advanced multicolor microscopy techniques, we anticipate that click labeling will contribute to novel discoveries in the neurobiology field.

Suppression without inhibition: how retinal computation contributes to saccadic suppression.

Visual perception remains stable across saccadic eye movements, despite the concurrent strongly disruptive visual flow. This stability is partially associated with a reduction in visual sensitivity, known as saccadic suppression, which already starts in the retina with reduced ganglion cell sensitivity. However, the retinal circuit mechanisms giving rise to such suppression remain unknown. Here, we describe these mechanisms using electrophysiology in mouse, pig, and macaque retina, 2-photon calcium imaging, computational modeling, and human psychophysics. We find that sequential stimuli, like those that naturally occur during saccades, trigger three independent suppressive mechanisms in the retina. The main mechanism is triggered by contrast-reversing sequential stimuli and originates within the receptive field center of ganglion cells. It does not involve inhibition or other known suppressive mechanisms like saturation or adaptation. Instead, it relies on temporal filtering of the inherently slow response of cone photoreceptors coupled with downstream nonlinearities. Two further mechanisms of suppression are present predominantly in ON ganglion cells and originate in the receptive field surround, highlighting another disparity between ON and OFF ganglion cells. The mechanisms uncovered here likely play a role in shaping the retinal output following eye movements and other natural viewing conditions where sequential stimulation is ubiquitous.

TTC30A and TTC30B Redundancy Protects IFT Complex B Integrity and Its Pivotal Role in Ciliogenesis.

Intraflagellar transport (IFT) is a microtubule-based system that supports the assembly and maintenance of cilia. The dysfunction of IFT leads to ciliopathies of variable severity. Two of the IFT-B components are the paralogue proteins TTC30A and TTC30B. To investigate whether these proteins constitute redundant functions, CRISPR/Cas9 was used to generate single TTC30A or B and double-knockout hTERT-RPE1 cells. Ciliogenesis assays showed the redundancy of both proteins while the polyglutamylation of cilia was affected in single knockouts. The localization of other IFT components was not affected by the depletion of a single paralogue. A loss of both proteins led to a severe ciliogenesis defect, resulting in no cilia formation, which was rescued by TTC30A or B. The redundancy can be explained by the highly similar interaction patterns of the paralogues; both equally interact with the IFT-B machinery. Our study demonstrates that a loss of one TTC30 paralogue can mostly be compensated by the other, thus preventing severe ciliary defects. However, cells assemble shorter cilia, which are potentially limited in their function, especially because of impaired polyglutamylation. A complete loss of both proteins leads to a deficit in IFT complex B integrity followed by disrupted IFT and subsequently no cilia formation.

Center-surround interactions underlie bipolar cell motion sensitivity in the mouse retina.

Motion sensing is a critical aspect of vision. We studied the representation of motion in mouse retinal bipolar cells and found that some bipolar cells are radially direction selective, preferring the origin of small object motion trajectories. Using a glutamate sensor, we directly observed bipolar cells synaptic output and found that there are radial direction selective and non-selective bipolar cell types, the majority being selective, and that radial direction selectivity relies on properties of the center-surround receptive field. We used these bipolar cell receptive fields along with connectomics to design biophysical models of downstream cells. The models and additional experiments demonstrated that bipolar cells pass radial direction selective excitation to starburst amacrine cells, which contributes to their directional tuning. As bipolar cells provide excitation to most amacrine and ganglion cells, their radial direction selectivity may contribute to motion processing throughout the visual system.

Retinal horizontal cells use different synaptic sites for global feedforward and local feedback signaling.

In the outer plexiform layer (OPL) of the mammalian retina, cone photoreceptors (cones) provide input to more than a dozen types of cone bipolar cells (CBCs). In the mouse, this transmission is modulated by a single horizontal cell (HC) type. HCs perform global signaling within their laterally coupled network but also provide local, cone-specific feedback. However, it is unknown how HCs provide local feedback to cones at the same time as global forward signaling to CBCs and where the underlying synapses are located. To assess how HCs simultaneously perform different modes of signaling, we reconstructed the dendritic trees of five HCs as well as cone axon terminals and CBC dendrites in a serial block-face electron microscopy volume and analyzed their connectivity. In addition to the fine HC dendritic tips invaginating cone axon terminals, we also identified "bulbs," short segments of increased dendritic diameter on the primary dendrites of HCs. These bulbs are in an OPL stratum well below the cone axon terminal base and make contacts with other HCs and CBCs. Our results from immunolabeling, electron microscopy, and glutamate imaging suggest that HC bulbs represent GABAergic synapses that do not receive any direct photoreceptor input. Together, our data suggest the existence of two synaptic strata in the mouse OPL, spatially separating cone-specific feedback and feedforward signaling to CBCs. A biophysical model of a HC dendritic branch and voltage imaging support the hypothesis that this spatial arrangement of synaptic contacts allows for simultaneous local feedback and global feedforward signaling by HCs.

Type-specific dendritic integration in mouse retinal ganglion cells.

Neural computation relies on the integration of synaptic inputs across a neuron's dendritic arbour. However, it is far from understood how different cell types tune this process to establish cell-type specific computations. Here, using two-photon imaging of dendritic Ca2+ signals, electrical recordings of somatic voltage and biophysical modelling, we demonstrate that four morphologically distinct types of mouse retinal ganglion cells with overlapping excitatory synaptic input (transient Off alpha, transient Off mini, sustained Off, and F-mini Off) exhibit type-specific dendritic integration profiles: in contrast to the other types, dendrites of transient Off alpha cells were spatially independent, with little receptive field overlap. The temporal correlation of dendritic signals varied also extensively, with the highest and lowest correlation in transient Off mini and transient Off alpha cells, respectively. We show that differences between cell types can likely be explained by differences in backpropagation efficiency, arising from the specific combinations of dendritic morphology and ion channel densities.

Systematic spatiotemporal mapping reveals divergent cell death pathways in three mouse models of hereditary retinal degeneration.

Calcium (Ca2+ ) dysregulation has been linked to neuronal cell death, including in hereditary retinal degeneration. Ca2+ dysregulation is thought to cause rod and cone photoreceptor cell death. Spatial and temporal heterogeneities in retinal disease models have hampered validation of this hypothesis. We examined the role of Ca2+ in photoreceptor degeneration, assessing the activation pattern of Ca2+ -dependent calpain proteases, generating spatiotemporal maps of the entire retina in the cpfl1 mouse model for primary cone degeneration, and in the rd1 and rd10 models for primary rod degeneration. We used Gaussian process models to distinguish the temporal sequences of degenerative molecular processes from other variability sources.In the rd1 and rd10 models, spatiotemporal pattern of increased calpain activity matched the progression of primary rod degeneration. High calpain activity coincided with activation of the calpain-2 isoform but not with calpain-1, suggesting differential roles for both calpain isoforms. Primary rod loss was linked to upregulation of apoptosis-inducing factor, although only a minute fraction of cells showed activity of the apoptotic marker caspase-3. After primary rod degeneration concluded, caspase-3 activation appeared in cones, suggesting apoptosis as the dominant mechanism for secondary cone loss. Gaussian process models highlighted calpain activity as a key event during primary rod photoreceptor cell death. Our data suggest a causal link between Ca2+ dysregulation and primary, nonapoptotic degeneration of photoreceptors and a role for apoptosis in secondary degeneration of cones, highlighting the importance of the spatial and temporal location of key molecular events, which may guide the evaluation of new therapies.

Neural circuits in the mouse retina support color vision in the upper visual field.

Color vision is essential for an animal's survival. It starts in the retina, where signals from different photoreceptor types are locally compared by neural circuits. Mice, like most mammals, are dichromatic with two cone types. They can discriminate colors only in their upper visual field. In the corresponding ventral retina, however, most cones display the same spectral preference, thereby presumably impairing spectral comparisons. In this study, we systematically investigated the retinal circuits underlying mouse color vision by recording light responses from cones, bipolar and ganglion cells. Surprisingly, most color-opponent cells are located in the ventral retina, with rod photoreceptors likely being involved. Here, the complexity of chromatic processing increases from cones towards the retinal output, where non-linear center-surround interactions create specific color-opponent output channels to the brain. This suggests that neural circuits in the mouse retina are tuned to extract color from the upper visual field, aiding robust detection of predators and ensuring the animal's survival.