Evolution of neuronal cell classes and types in the vertebrate retina.

The basic plan of the retina is conserved across vertebrates, yet species differ profoundly in their visual needs1. Retinal cell types may have evolved to accommodate these varied needs, but this has not been systematically studied. Here we generated and integrated single-cell transcriptomic atlases of the retina from 17 species: humans, two non-human primates, four rodents, three ungulates, opossum, ferret, tree shrew, a bird, a reptile, a teleost fish and a lamprey. We found high molecular conservation of the six retinal cell classes (photoreceptors, horizontal cells, bipolar cells, amacrine cells, retinal ganglion cells (RGCs) and Müller glia), with transcriptomic variation across species related to evolutionary distance. Major subclasses were also conserved, whereas variation among cell types within classes or subclasses was more pronounced. However, an integrative analysis revealed that numerous cell types are shared across species, based on conserved gene expression programmes that are likely to trace back to an early ancestral vertebrate. The degree of variation among cell types increased from the outer retina (photoreceptors) to the inner retina (RGCs), suggesting that evolution acts preferentially to shape the retinal output. Finally, we identified rodent orthologues of midget RGCs, which comprise more than 80% of RGCs in the human retina, subserve high-acuity vision, and were previously believed to be restricted to primates2. By contrast, the mouse orthologues have large receptive fields and comprise around 2% of mouse RGCs. Projections of both primate and mouse orthologous types are overrepresented in the thalamus, which supplies the primary visual cortex. We suggest that midget RGCs are not primate innovations, but are descendants of evolutionarily ancient types that decreased in size and increased in number as primates evolved, thereby facilitating high visual acuity and increased cortical processing of visual information.

A Fast Algorithm for All-Pairs-Shortest-Paths Suitable for Neural Networks.

Given a directed graph of nodes and edges connecting them, a common problem is to find the shortest path between any two nodes. Here we show that the shortest path distances can be found by a simple matrix inversion: if the edges are given by the adjacency matrix Aij, then with a suitably small value of γ, the shortest path distances are Dij=ceil(logγ[(I-γA)-1]ij).We derive several graph-theoretic bounds on the value of γ and explore its useful range with numerics on different graph types. Even when the distance function is not globally accurate across the entire graph, it still works locally to instruct pursuit of the shortest path. In this mode, it also extends to weighted graphs with positive edge weights. For a wide range of dense graphs, this distance function is computationally faster than the best available alternative. Finally, we show that this method leads naturally to a neural network solution of the all-pairs-shortest-path problem.

Role of the cerebro-placental-uterine ratio in predicting adverse perinatal outcome in low-risk pregnancies at term.

PURPOSE: The cerebroplacental ratio (CPR) is associated with adverse perinatal outcome (APO) in low-risk pregnancies near term. A Doppler parameter, which also includes information from the uterine vessels could potentially improve detection of subclinical placental dysfunction. The aim of this study is to investigate the performance of cerebro-placental-uterine ratio (CPUR) related to APO prediction in low-risk term pregnancies in > 40 + 0 weeks. METHODS: This is a retrospective cohort study. All low-risk pregnancies in which feto-maternal Doppler was examined from 40 + 0 weeks and an appropriate for gestational age fetus was present were included. ROC (receiver operating characteristic curves) analyses were performed to assess the predictive value of CPUR. The presence of at least one of the following outcome parameters was defined as composite APO (CAPO): operative delivery (OD) due to intrapartum fetal compromise (IFC), admission to the neonatal intensive care unit, umbilical cord arterial pH ≤ 7.15, 5 min APGAR ≤ 7. RESULTS: A total of n = 114 cases were included. Mean gestational age at examination and delivery were 40 + 3 weeks and 40 + 6 weeks, respectively. Overall, CAPO occurred in 38 of 114 cases (33.3%). ROC analyses showed a significant association of CPUR (AUC = 0.67, p = 0.004) and CPR (AUC = 0.68, p = 0.002) with CAPO. Additionally, CPUR (AUC = 0.64, p = 0.040) showed a predictive value for OD due to IFC. CONCLUSION: The CPUR in > 40 + 0 weeks showed a predictive value for CAPO and OD due to IFC in low-risk pregnancies. However, the extent to which CPUR can be used to optimize delivery management warrants further investigations in prospective interventional studies.

Evaluation of right ventricular myocardial deformation properties in fetal hypoplastic left heart by two-dimensional speckle tracking echocardiography.

PURPOSE: Right ventricular (RV) function influences the outcome of hypoplastic left heart (HLH) patients. This study aimed to confirm the assumption of prenatal RV remodeling and possible influencing factors of myocardial restructuring using two-dimensional speckle tracking echocardiography (2D STE). METHODS: This is a retrospective cross-sectional cohort study including HLH fetuses and gestational age-matched controls. Based on a four-chamber view, cine loops were stored with 60 frames per second. Global longitudinal peak systolic strain (GLPSS) of the RV was retrospectively determined and compared to healthy controls. Furthermore, HLH subgroups were built according to the presence of left ventricular endocardial fibroelastosis (LV-EFE) and restrictive foramen ovale (FO) to investigate the effect of these compromising factors on myocardial deformation. RESULTS: A total of 41 HLH fetuses and 101 controls were included. Gestational age at fetal assessment was similarly distributed in both groups (controls: 26.0 ± 5.6 weeks vs. HLH: 29.1 ± 5.6 weeks). Relating to RV-GLPSS values, fetuses with HLH demonstrated lower mean values than healthy control fetuses (- 15.65% vs. - 16.80%, p = 0.065). Cases with LV-EFE (n = 11) showed significantly lower mean values compared to such without LV-EFE (n = 30) (RV-GLPSS: - 12.12% vs. - 16.52%, p = 0.003). No significant differences were observed for cases with FO restriction (n = 10). CONCLUSIONS: In HLH the RV undergoes prenatal remodeling, leading to an adaptation of myocardial function to LV conditions. Further explorations by STE should expand knowledge about RV contraction properties in HLH and its impact on surgical outcome.

A neuronal circuit for colour vision based on rod-cone opponency.

In bright light, cone-photoreceptors are active and colour vision derives from a comparison of signals in cones with different visual pigments. This comparison begins in the retina, where certain retinal ganglion cells have 'colour-opponent' visual responses-excited by light of one colour and suppressed by another colour. In dim light, rod-photoreceptors are active, but colour vision is impossible because they all use the same visual pigment. Instead, the rod signals are thought to splice into retinal circuits at various points, in synergy with the cone signals. Here we report a new circuit for colour vision that challenges these expectations. A genetically identified type of mouse retinal ganglion cell called JAMB (J-RGC), was found to have colour-opponent responses, OFF to ultraviolet (UV) light and ON to green light. Although the mouse retina contains a green-sensitive cone, the ON response instead originates in rods. Rods and cones both contribute to the response over several decades of light intensity. Remarkably, the rod signal in this circuit is antagonistic to that from cones. For rodents, this UV-green channel may play a role in social communication, as suggested by spectral measurements from the environment. In the human retina, all of the components for this circuit exist as well, and its function can explain certain experiences of colour in dim lights, such as a 'blue shift' in twilight. The discovery of this genetically defined pathway will enable new targeted studies of colour processing in the brain.

Orientation columns in the mouse superior colliculus.

More than twenty types of retinal ganglion cells conduct visual information from the eye to the rest of the brain. Each retinal ganglion cell type tessellates the retina in a regular mosaic, so that every point in visual space is processed for visual primitives such as contrast and motion. This information flows to two principal brain centres: the visual cortex and the superior colliculus. The superior colliculus plays an evolutionarily conserved role in visual behaviours, but its functional architecture is poorly understood. Here we report on population recordings of visual responses from neurons in the mouse superior colliculus. Many neurons respond preferentially to lines of a certain orientation or movement axis. We show that cells with similar orientation preferences form large patches that span the vertical thickness of the retinorecipient layers. This organization is strikingly different from the randomly interspersed orientation preferences in the mouse's visual cortex; instead, it resembles the orientation columns observed in the visual cortices of large mammals. Notably, adjacent superior colliculus orientation columns have only limited receptive field overlap. This is in contrast to the organization of visual cortex, where each point in the visual field activates neurons with all preferred orientations. Instead, the superior colliculus favours specific contour orientations within ∼30° regions of the visual field, a finding with implications for behavioural responses mediated by this brain centre.

Functional diversity among sensory neurons from efficient coding principles.

In many sensory systems the neural signal is coded by the coordinated response of heterogeneous populations of neurons. What computational benefit does this diversity confer on information processing? We derive an efficient coding framework assuming that neurons have evolved to communicate signals optimally given natural stimulus statistics and metabolic constraints. Incorporating nonlinearities and realistic noise, we study optimal population coding of the same sensory variable using two measures: maximizing the mutual information between stimuli and responses, and minimizing the error incurred by the optimal linear decoder of responses. Our theory is applied to a commonly observed splitting of sensory neurons into ON and OFF that signal stimulus increases or decreases, and to populations of monotonically increasing responses of the same type, ON. Depending on the optimality measure, we make different predictions about how to optimally split a population into ON and OFF, and how to allocate the firing thresholds of individual neurons given realistic stimulus distributions and noise, which accord with certain biases observed experimentally.

Atrial and Ventricular Deformation Analysis in Normal Fetal Hearts Using Two-Dimensional Speckle Tracking Echocardiography.

OBJECTIVE: Two-dimensional speckle tracking echocardiography (2D-STE)-based strain values of the left and the right ventricle have been established; however, less is known about atrial deformation. The aim of our study was to assess both atrial strain and ventricular strain using 2D-STE in a cardiac 4-chamber view and to investigate the effect of possible influencing factors such as gestational age. METHODS: Fetal echocardiography was performed on a Toshiba Aplio 500 ultrasound system. Based on an apical or basal 4-chamber view of the fetal heart, left and right ventricular longitudinal peak systolic strain (LVLPSS and RVLPSS) as well as left and right atrial longitudinal peak systolic strain (LALPSS and RALPSS) were assessed by 2D-STE. RESULTS: A total of 101 healthy fetuses were included. The mean gestational age (GA) was 26.0 ± 5.6 weeks. GA was significantly positively correlated (p < 0.05) with LVLPSS and RVLPSS and significantly negatively correlated (p < 0.05) with LALPSS and RALPSS. The mean values for LVLPSS and RVLPSS were -17.44 ± 2.29% and -16.89 ± 1.72%. The mean values for LALPSS and RALPSS were 34.09 ± 4.17% and 35.36 ± 2.90%. CONCLUSION: Ventricular and atrial deformation analysis in 2D-STE was technically feasible and showed comparable values to current data. For future research on myocardial function (MF) of the fetus, considering GA as an influencing factor for deformation analysis seems to be adequate. Especially, atrial deformation analysis allows the assessment of diastolic myocardial function. Further research needs to clarify the clinical meaning of these myocardial deformation indices in fetuses at risk.

Relevance of peripheral cholinesterase activity on postoperative delirium in adult surgical patients (CESARO): A prospective observational cohort study.

BACKGROUND: The cholinergic system is considered to play a key role in the development of postoperative delirium (POD), which is a common complication after surgery. OBJECTIVES: To determine whether peri-operative acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) activities are associated with the development of POD in in-hospital surgical patients, and raise hypotheses on cholinergic regulatory mechanisms in POD. DESIGN: A prospective multicentre observational study by the Peripheral Cholinesterase-activity on Neurocognitive Dysfunctions in Surgical Patients (CESARO) study group. SETTING: Nine German hospitals. PATIENTS: Patients of at least 18 years of age scheduled for inpatient elective surgery for a variety of surgical procedures. A total of 650 patients (mean age 61.5 years, 52.8% male) were included. METHODS: Clinical variables, and peripheral AChE and BuChE activities, were assessed throughout the peri-operative period using bedside point-of-care measurements (one pre-operative and two postoperative measurements). POD screening was conducted postoperatively for at least 24 h and up to the third postoperative day using a validated screening tool (nursing delirium screening scale). RESULTS: In all, 179 patients (27.5%) developed POD within the early postoperative phase. There was a lower BuChE activity in patients with delirium compared with patients without delirium pre-operatively (Cohen's r = 0.07, P = 0.091), on postoperative day 1 (Cohen's r = 0.12, P = 0.003) and on postoperative day 2 (Cohen's r = 0.12, P = 0.002). In contrast, there was a significantly higher AChE activity in patients with delirium compared with patients without delirium pre-operatively (Cohen's r = 0.10, P = 0.012), on postoperative day 1 (Cohen's r = 0.11, P = 0.004) and on postoperative day 2 (Cohen's r = 0.13, P = 0.002). After adjusting for covariates in multiple logistic regression, a significant association between both BuChE and AChE activities and POD was not found. However, in the multivariable analysis using the Generalized Estimating Equation, cholinesterase activities showed that a decrease of BuChE activity by 100 U L increased the risk of a delirium by approximately 2.1% (95% CI 1.6 to 2.8%) and for each 1 U g of haemoglobin increase in AChE activity, there was a 1.4% (95% CI 0.6 to 2.2%) increased risk of POD. CONCLUSION: Peri-operative peripheral cholinesterase activities may be related to the development of POD, but the clinical implications remain unclear. Further studies, in homogeneous patient groups with a strict protocol for measurement time points, are needed to investigate the relationship between cholinesterase activities and POD. TRIAL REGISTRATION: www.clinicaltrials.gov. Identifier NCT01964274.

Benefits of pathway splitting in sensory coding.

In many sensory systems, the neural signal splits into multiple parallel pathways. For example, in the mammalian retina, ~20 types of retinal ganglion cells transmit information about the visual scene to the brain. The purpose of this profuse and early pathway splitting remains unknown. We examine a common instance of splitting into ON and OFF neurons excited by increments and decrements of light intensity in the visual scene, respectively. We test the hypothesis that pathway splitting enables more efficient encoding of sensory stimuli. Specifically, we compare a model system with an ON and an OFF neuron to one with two ON neurons. Surprisingly, the optimal ON-OFF system transmits the same information as the optimal ON-ON system, if one constrains the maximal firing rate of the neurons. However, the ON-OFF system uses fewer spikes on average to transmit this information. This superiority of the ON-OFF system is also observed when the two systems are optimized while constraining their mean firing rate. The efficiency gain for the ON-OFF split is comparable with that derived from decorrelation, a well known processing strategy of early sensory systems. The gain can be orders of magnitude larger when the ecologically important stimuli are rare but large events of either polarity. The ON-OFF system also provides a better code for extracting information by a linear downstream decoder. The results suggest that the evolution of ON-OFF diversification in sensory systems may be driven by the benefits of lowering average metabolic cost, especially in a world in which the relevant stimuli are sparse.

The most numerous ganglion cell type of the mouse retina is a selective feature detector.

The retina reports the visual scene to the brain through many parallel channels, each carried by a distinct population of retinal ganglion cells. Among these, the population with the smallest and densest receptive fields encodes the neural image with highest resolution. In human retina, and those of cat and macaque, these high-resolution ganglion cells act as generic pixel encoders: They serve to represent many different visual inputs and convey a neural image of the scene downstream for further processing. Here we identify and analyze high-resolution ganglion cells in the mouse retina, using a transgenic line in which these cells, called "W3", are labeled fluorescently. Counter to the expectation, these ganglion cells do not participate in encoding generic visual scenes, but remain silent during most common visual stimuli. A detailed study of their response properties showed that W3 cells pool rectified excitation from both On and Off bipolar cells, which makes them sensitive to local motion. However, they also receive unusually strong lateral inhibition, both pre- and postsynaptically, triggered by distant motion. As a result, the W3 cell can detect small moving objects down to the receptive field size of bipolar cells, but only if the background is featureless or stationary--an unusual condition. A survey of naturalistic stimuli shows that W3 cells may serve as alarm neurons for overhead predators.

Nonlinear dynamics support a linear population code in a retinal target-tracking circuit.

A basic task faced by the visual system of many organisms is to accurately track the position of moving prey. The retina is the first stage in the processing of such stimuli; the nature of the transformation here, from photons to spike trains, constrains not only the ultimate fidelity of the tracking signal but also the ease with which it can be extracted by other brain regions. Here we demonstrate that a population of fast-OFF ganglion cells in the salamander retina, whose dynamics are governed by a nonlinear circuit, serve to compute the future position of the target over hundreds of milliseconds. The extrapolated position of the target is not found by stimulus reconstruction but is instead computed by a weighted sum of ganglion cell outputs, the population vector average (PVA). The magnitude of PVA extrapolation varies systematically with target size, speed, and acceleration, such that large targets are tracked most accurately at high speeds, and small targets at low speeds, just as is seen in the motion of real prey. Tracking precision reaches the resolution of single photoreceptors, and the PVA algorithm performs more robustly than several alternative algorithms. If the salamander brain uses the fast-OFF cell circuit for target extrapolation as we suggest, the circuit dynamics should leave a microstructure on the behavior that may be measured in future experiments. Our analysis highlights the utility of simple computations that, while not globally optimal, are efficiently implemented and have close to optimal performance over a limited but ethologically relevant range of stimuli.

Dynamical adaptation in photoreceptors.

Adaptation is at the heart of sensation and nowhere is it more salient than in early visual processing. Light adaptation in photoreceptors is doubly dynamical: it depends upon the temporal structure of the input and it affects the temporal structure of the response. We introduce a non-linear dynamical adaptation model of photoreceptors. It is simple enough that it can be solved exactly and simulated with ease; analytical and numerical approaches combined provide both intuition on the behavior of dynamical adaptation and quantitative results to be compared with data. Yet the model is rich enough to capture intricate phenomenology. First, we show that it reproduces the known phenomenology of light response and short-term adaptation. Second, we present new recordings and demonstrate that the model reproduces cone response with great precision. Third, we derive a number of predictions on the response of photoreceptors to sophisticated stimuli such as periodic inputs, various forms of flickering inputs, and natural inputs. In particular, we demonstrate that photoreceptors undergo rapid adaptation of response gain and time scale, over ∼ 300[Formula: see text] ms-i. e., over the time scale of the response itself-and we confirm this prediction with data. For natural inputs, this fast adaptation can modulate the response gain more than tenfold and is hence physiologically relevant.

Molecular identification of a retinal cell type that responds to upward motion.

The retina contains complex circuits of neurons that extract salient information from visual inputs. Signals from photoreceptors are processed by retinal interneurons, integrated by retinal ganglion cells (RGCs) and sent to the brain by RGC axons. Distinct types of RGC respond to different visual features, such as increases or decreases in light intensity (ON and OFF cells, respectively), colour or moving objects. Thus, RGCs comprise a set of parallel pathways from the eye to the brain. The identification of molecular markers for RGC subsets will facilitate attempts to correlate their structure with their function, assess their synaptic inputs and targets, and study their diversification. Here we show, by means of a transgenic marking method, that junctional adhesion molecule B (JAM-B) marks a previously unrecognized class of OFF RGCs in mice. These cells have asymmetric dendritic arbors aligned in a dorsal-to-ventral direction across the retina. Their receptive fields are also asymmetric and respond selectively to stimuli moving in a soma-to-dendrite direction; because the lens reverses the image of the world on the retina, these cells detect upward motion in the visual field. Thus, JAM-B identifies a unique population of RGCs in which structure corresponds remarkably to function.

A fast algorithm for All-Pairs-Shortest-Paths suitable for neural networks.

Given a directed graph of nodes and edges connecting them, a common problem is to find the shortest path between any two nodes. Here we show that the shortest path distances can be found by a simple matrix inversion: If the edges are given by the adjacency matrix A i j then with a suitably small value of γ the shortest path distances are D i j = ceil ( log γ [ ( I - γ A ) - 1 ] i j ) We derive several graph-theoretic bounds on the value of γ , and explore its useful range with numerics on different graph types. Even when the distance function is not globally accurate across the entire graph, it still works locally to instruct pursuit of the shortest path. In this mode, it also extends to weighted graphs with positive edge weights. For a wide range of dense graphs this distance function is computationally faster than the best available alternative. Finally we show that this method leads naturally to a neural network solution of the all-pairs-shortest-path problem.

Endotaxis: A neuromorphic algorithm for mapping, goal-learning, navigation, and patrolling.

An animal entering a new environment typically faces three challenges: explore the space for resources, memorize their locations, and navigate towards those targets as needed. Here we propose a neural algorithm that can solve all these problems and operates reliably in diverse and complex environments. At its core, the mechanism makes use of a behavioral module common to all motile animals, namely the ability to follow an odor to its source. We show how the brain can learn to generate internal "virtual odors" that guide the animal to any location of interest. This endotaxis algorithm can be implemented with a simple 3-layer neural circuit using only biologically realistic structures and learning rules. Several neural components of this scheme are found in brains from insects to humans. Nature may have evolved a general mechanism for search and navigation on the ancient backbone of chemotaxis.

Bayesian model of dynamic image stabilization in the visual system.

Humans can resolve the fine details of visual stimuli although the image projected on the retina is constantly drifting relative to the photoreceptor array. Here we demonstrate that the brain must take this drift into account when performing high acuity visual tasks. Further, we propose a decoding strategy for interpreting the spikes emitted by the retina, which takes into account the ambiguity caused by retinal noise and the unknown trajectory of the projected image on the retina. A main difficulty, addressed in our proposal, is the exponentially large number of possible stimuli, which renders the ideal Bayesian solution to the problem computationally intractable. In contrast, the strategy that we propose suggests a realistic implementation in the visual cortex. The implementation involves two populations of cells, one that tracks the position of the image and another that represents a stabilized estimate of the image itself. Spikes from the retina are dynamically routed to the two populations and are interpreted in a probabilistic manner. We consider the architecture of neural circuitry that could implement this strategy and its performance under measured statistics of human fixational eye motion. A salient prediction is that in high acuity tasks, fixed features within the visual scene are beneficial because they provide information about the drifting position of the image. Therefore, complete elimination of peripheral features in the visual scene should degrade performance on high acuity tasks involving very small stimuli.

Functional cell types in the mouse superior colliculus.

The superior colliculus (SC) represents a major visual processing station in the mammalian brain that receives input from many types of retinal ganglion cells (RGCs). How many parallel channels exist in the SC, and what information does each encode? Here, we recorded from mouse superficial SC neurons under a battery of visual stimuli including those used for classification of RGCs. An unsupervised clustering algorithm identified 24 functional types based on their visual responses. They fall into two groups: one that responds similarly to RGCs and another with more diverse and specialized stimulus selectivity. The second group is dominant at greater depths, consistent with a vertical progression of signal processing in the SC. Cells of the same functional type tend to cluster near each other in anatomical space. Compared to the retina, the visual representation in the SC has lower dimensionality, consistent with a sifting process along the visual pathway.

Evolution of neuronal cell classes and types in the vertebrate retina.

The basic plan of the retina is conserved across vertebrates, yet species differ profoundly in their visual needs (Baden et al., 2020). One might expect that retinal cell types evolved to accommodate these varied needs, but this has not been systematically studied. Here, we generated and integrated single-cell transcriptomic atlases of the retina from 17 species: humans, two non-human primates, four rodents, three ungulates, opossum, ferret, tree shrew, a teleost fish, a bird, a reptile and a lamprey. Molecular conservation of the six retinal cell classes (photoreceptors, horizontal cells, bipolar cells, amacrine cells, retinal ganglion cells [RGCs] and Muller glia) is striking, with transcriptomic differences across species correlated with evolutionary distance. Major subclasses are also conserved, whereas variation among types within classes or subclasses is more pronounced. However, an integrative analysis revealed that numerous types are shared across species based on conserved gene expression programs that likely trace back to the common ancestor of jawed vertebrates. The degree of variation among types increases from the outer retina (photoreceptors) to the inner retina (RGCs), suggesting that evolution acts preferentially to shape the retinal output. Finally, we identified mammalian orthologs of midget RGCs, which comprise >80% of RGCs in the human retina, subserve high-acuity vision, and were believed to be primate-specific (Berson, 2008); in contrast, the mouse orthologs comprise <2% of mouse RGCs. Projections both primate and mouse orthologous types are overrepresented in the thalamus, which supplies the primary visual cortex. We suggest that midget RGCs are not primate innovations, but descendants of evolutionarily ancient types that decreased in size and increased in number as primates evolved, thereby facilitating high visual acuity and increased cortical processing of visual information.

The projective field of a retinal amacrine cell.

In sensory systems, neurons are generally characterized by their receptive field, namely the sensitivity to activity patterns at the input of the circuit. To assess the role of the neuron in the system, one must also know its projective field, namely the spatiotemporal effects the neuron exerts on all of the outputs of the circuit. We studied both the receptive and projective fields of an amacrine interneuron in the salamander retina. This amacrine type has a sustained OFF response with a small receptive field, but its output projects over a much larger region. Unlike other amacrine cells, this type is remarkably promiscuous and affects nearly every ganglion cell within reach of its dendrites. Its activity modulates the sensitivity of visual responses in ganglion cells but leaves their kinetics unchanged. The projective field displays a center-surround structure: depolarizing a single amacrine suppresses the visual sensitivity of ganglion cells nearby and enhances it at greater distances. This change in sign is seen even within the receptive field of one ganglion cell; thus, the modulation occurs presynaptically on bipolar cell terminals, most likely via GABA(B) receptors. Such an antagonistic projective field could contribute to the mechanisms of the retina for predictive coding.