Computing complex visual features with retinal spike times

Anti-Hebbian plasticity drives sequence learning in striatum

Spatio-temporal activity patterns have been observed in a variety of brain areas in spontaneous activity, prior to or during action, or in response to stimuli. Biological mechanisms endowing neurons with the ability to distinguish between different sequences remain largely unknown. Learning sequences of spikes raises multiple challenges, such as maintaining in memory spike history and discriminating partially overlapping sequences. Here, we show that anti-Hebbian spike-timing dependent plasticity (STDP), as observed at cortico-striatal synapses, can naturally lead to learning spike sequences. We design a spiking model of the striatal output neuron receiving spike patterns defined as sequential input from a fixed set of cortical neurons. We use a simple synaptic plasticity rule that combines anti-Hebbian STDP and non-associative potentiation for a subset of the presented patterns called rewarded patterns. We study the ability of striatal output neurons to discriminate rewarded from non-rewarded patterns by firing only after the presentation of a rewarded pattern. In particular, we show that two biological properties of striatal networks, spiking latency and collateral inhibition, contribute to an increase in accuracy, by allowing a better discrimination of partially overlapping sequences. These results suggest that anti-Hebbian STDP may serve as a biological substrate for learning sequences of spikes.

Retinal ganglion cells encode differently in the myopic mouse retina?

The etiology of myopia remains unclear. This study investigated whether retinal ganglion cells (RGCs) in the myopic retina encode visual information differently from the normal retina and to determine the role of Connexin (Cx) 36 in this process. Generalized linear models (GLMs), which can capture stimulus-dependent changes in real neurons with spike timing precision and reliability, were used to predict RGCs responses to focused and defocused images in the retinas of wild-type (normal) and Lens-Induced Myopia (LIM) mice. As the predominant subunit of gap junctions in the mouse retina and a plausible modulator in myopia development, Cx36 knockout (KO) mice were used as a control for an intact retinal circuit. The kinetics of excitatory postsynaptic currents (EPSCs) of a single αRGC could reflect projection of both focused and defocused images in the retinas of normal and LIM, but not in the Cx36 knockout mice. Poisson GLMs revealed that RGC encoding of visual stimuli in the LIM retina was similar to that of the normal retina. In the LIM retinas, the linear-Gaussian GLM model with offset was a better fit for predicting the spike count under a focused image than the defocused image. Akaike information criterion (AIC) indicated that nonparametric GLM (np-GLM) model predicted focused/defocused images better in both LIM and normal retinas. However, the spike counts in 33% of αRGCs in LIM retinas were better fitted by exponential GLM (exp-GLM) under defocus, compared to only 13% αRGCs in normal retinas. The differences in encoding performance between LIM and normal retinas indicated the possible amendment and plasticity of the retinal circuit in myopic retinas. The absence of a similar response between Cx36 KO mice and normal/LIM mice might suggest that Cx36, which is associated with myopia development, plays a role in encoding focused and defocused images.

SmaRT2P: a software for generating and processing smart line recording trajectories for population two-photon calcium imaging

Two-photon fluorescence calcium imaging allows recording the activity of large neural populations with subcellular spatial resolution, but it is typically characterized by low signal-to-noise ratio (SNR) and poor accuracy in detecting single or few action potentials when large number of neurons are imaged. We recently showed that implementing a smart line scanning approach using trajectories that optimally sample the regions of interest increases both the SNR fluorescence signals and the accuracy of single spike detection in population imaging in vivo. However, smart line scanning requires highly specialised software to design recording trajectories, interface with acquisition hardware, and efficiently process acquired data. Furthermore, smart line scanning needs optimized strategies to cope with movement artefacts and neuropil contamination. Here, we develop and validate SmaRT2P, an open-source, user-friendly and easy-to-interface Matlab-based software environment to perform optimized smart line scanning in two-photon calcium imaging experiments. SmaRT2P is designed to interface with popular acquisition software (e.g., ScanImage) and implements novel strategies to detect motion artefacts, estimate neuropil contamination, and minimize their impact on functional signals extracted from neuronal population imaging. SmaRT2P is structured in a modular way to allow flexibility in the processing pipeline, requiring minimal user intervention in parameter setting. The use of SmaRT2P for smart line scanning has the potential to facilitate the functional investigation of large neuronal populations with increased SNR and accuracy in detecting the discharge of single and few action potentials.

Role of AMPA receptor desensitization in short term depression - lessons from retinogeniculate synapses

Repetitive synapse activity induces various forms of short-term plasticity. The role of presynaptic mechanisms such as residual Ca²⁺ and vesicle depletion in short-term facilitation and short-term depression is well established. On the other hand, the contribution of postsynaptic mechanisms such as receptor desensitization and saturation to short-term plasticity is less well known and often ignored. In this review, I will describe short-term plasticity in retinogeniculate synapses of relay neurons of the dorsal lateral geniculate nucleus (dLGN) to exemplify the synaptic properties that facilitate the contribution of AMPA receptor desensitization to short-term plasticity. These include high vesicle release probability, glutamate spillover and, importantly, slow recovery from desensitization of AMPA receptors. The latter is strongly regulated by the interaction of AMPA receptors with auxiliary proteins such as CKAMP44. Finally, I discuss the relevance of short-term plasticity in retinogeniculate synapses for the processing of visual information by LGN relay neurons.

The impact of synchronous versus asynchronous electrical stimulation in artificial vision

Visual prosthesis devices designed to restore sight to the blind have been under development in the laboratory for several decades. Clinical translation continues to be challenging, due in part to gaps in our understanding of critical parameters such as how phosphenes, the electrically-generated pixels of artificial vision, can be combined to form images. In this review we explore the effects that synchronous and asynchronous electrical stimulation across multiple electrodes have in evoking phosphenes. Understanding how electrical patterns influence phosphene generation to control object binding and perception of visual form is fundamental to creation of a clinically successful prosthesis.

Reconstruction of natural images from responses of primate retinal ganglion cells

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

The covariance perceptron: A new paradigm for classification and processing of time series in recurrent neuronal networks

Learning in neuronal networks has developed in many directions, in particular to reproduce cognitive tasks like image recognition and speech processing. Implementations have been inspired by stereotypical neuronal responses like tuning curves in the visual system, where, for example, ON/OFF cells fire or not depending on the contrast in their receptive fields. Classical models of neuronal networks therefore map a set of input signals to a set of activity levels in the output of the network. Each category of inputs is thereby predominantly characterized by its mean. In the case of time series, fluctuations around this mean constitute noise in this view. For this paradigm, the high variability exhibited by the cortical activity may thus imply limitations or constraints, which have been discussed for many years. For example, the need for averaging neuronal activity over long periods or large groups of cells to assess a robust mean and to diminish the effect of noise correlations. To reconcile robust computations with variable neuronal activity, we here propose a conceptual change of perspective by employing variability of activity as the basis for stimulus-related information to be learned by neurons, rather than merely being the noise that corrupts the mean signal. In this new paradigm both afferent and recurrent weights in a network are tuned to shape the input-output mapping for covariances, the second-order statistics of the fluctuating activity. When including time lags, covariance patterns define a natural metric for time series that capture their propagating nature. We develop the theory for classification of time series based on their spatio-temporal covariances, which reflect dynamical properties. We demonstrate that recurrent connectivity is able to transform information contained in the temporal structure of the signal into spatial covariances. Finally, we use the MNIST database to show how the covariance perceptron can capture specific second-order statistical patterns generated by moving digits.

The dynamics in cortex is characterized by highly fluctuating activity: Even under the very same experimental conditions the activity typically does not reproduce on the level of individual spikes. Given this variability, how then does the brain realize its quasi-deterministic function? One obvious solution is to compute averages over many cells, assuming that the mean activity, or rate, is actually the decisive signal. Variability across trials of an experiment is thus considered noise. We here explore the opposite view: Can fluctuations be used to actually represent information? And if yes, is there a benefit over a representation using the mean rate? We find that a fluctuation-based scheme is not only powerful in distinguishing signals into several classes, but also that networks can efficiently be trained in the new paradigm. Moreover, we argue why such a scheme of representation is more consistent with known forms of synaptic plasticity than rate-based network dynamics.

Response Latency Tuning by Retinal Circuits Modulates Signal Efficiency

In the visual system, retinal ganglion cells (RGCs) of various subtypes encode preprocessed photoreceptor signals into a spike output which is then transmitted towards the brain through parallel feature pathways. Spike timing determines how each feature signal contributes to the output of downstream neurons in visual brain centers, thereby influencing efficiency in visual perception. In this study, we demonstrate a marked population-wide variability in RGC response latency that is independent of trial-to-trial variability and recording approach. RGC response latencies to simple visual stimuli vary considerably in a heterogenous cell population but remain reliable when RGCs of a single subtype are compared. This subtype specificity, however, vanishes when the retinal circuitry is bypassed via direct RGC electrical stimulation. This suggests that latency is primarily determined by the signaling speed through retinal pathways that provide subtype specific inputs to RGCs. In addition, response latency is significantly altered when GABA inhibition or gap junction signaling is disturbed, which further supports the key role of retinal microcircuits in latency tuning. Finally, modulation of stimulus parameters affects individual RGC response delays considerably. Based on these findings, we hypothesize that retinal microcircuits fine-tune RGC response latency, which in turn determines the context-dependent weighing of each signal and its contribution to visual perception.

First spike latency of ON/OFF neurons in the optic tectum of pigeons

Some shallow and middle optic tectum (OT) neurons have stable, asymmetric full-screen ON and OFF stimulus response properties, which makes them candidates for delay encoding. In this paper, we investigated the delay encoding mechanism for the neuronal clusters in the OT region of pigeons and determined the mechanism of delay coding in the OT region. By analyzing the responses of the neuron cluster under full-screen switch-on and switch-off stimulation, we found that the delay coding was widespread in the OT region where the ON/OFF stimulation time difference was 4-6 ms. Information theory analysis under grating stimulation and experiments based on single-neuron character reconstruction of neurons showed that OT neuron clusters use the first spike latency (FSL) for the rapid transfer of spatial structure information. Furthermore, 4 models were used to predict the first spike latency of these OT neurons. The best simulation results were obtained using an architecture where the ON and OFF paths of multiple retinal ganglion cells (RGCs) were integrated and summed, respectively.

A single retinal circuit model for multiple computations

Vision is dependent on extracting intricate features of the visual information from the outside world, and complex visual computations begin to take place as soon as at the retinal level. In multiple studies on salamander retinas, the responses of a subtype of retinal ganglion cells, i.e., fast/biphasic-OFF ganglion cells, have been shown to be able to realize multiple functions, such as the segregation of a moving object from its background, motion anticipation, and rapid encoding of the spatial features of a new visual scene. For each of these visual functions, modeling approaches using extended linear-nonlinear cascade models suggest specific preceding retinal circuitries merging onto fast/biphasic-OFF ganglion cells. However, whether multiple visual functions can be accommodated together in a certain retinal circuitry and how specific mechanisms for each visual function interact with each other have not been investigated. Here, we propose a physiologically consistent, detailed computational model of the retinal circuit based on the spatiotemporal dynamics and connections of each class of retinal neurons to implement object motion sensitivity, motion anticipation, and rapid coding in the same circuit. Simulations suggest that multiple computations can be accommodated together, thereby implying that the fast/biphasic-OFF ganglion cell has potential to output a train of spikes carrying multiple pieces of information on distinct features of the visual stimuli.

A transformation from temporal to ensemble coding in a model of piriform cortex

Different coding strategies are used to represent odor information at various stages of the mammalian olfactory system. A temporal latency code represents odor identity in olfactory bulb (OB), but this temporal information is discarded in piriform cortex (PCx) where odor identity is instead encoded through ensemble membership. We developed a spiking PCx network model to understand how this transformation is implemented. In the model, the impact of OB inputs activated earliest after inhalation is amplified within PCx by diffuse recurrent collateral excitation, which then recruits strong, sustained feedback inhibition that suppresses the impact of later-responding glomeruli. We model increasing odor concentrations by decreasing glomerulus onset latencies while preserving their activation sequences. This produces a multiplexed cortical odor code in which activated ensembles are robust to concentration changes while concentration information is encoded through population synchrony. Our model demonstrates how PCx circuitry can implement multiplexed ensemble-identity/temporal-concentration odor coding.

CKAMP44 modulates integration of visual inputs in the lateral geniculate nucleus

Relay neurons in the dorsal lateral geniculate nucleus (dLGN) receive excitatory inputs from retinal ganglion cells (RGCs). Retinogeniculate synapses are characterized by a prominent short-term depression of AMPA receptor (AMPAR)-mediated currents, but the underlying mechanisms and its function for visual integration are not known. Here we identify CKAMP44 as a crucial auxiliary subunit of AMPARs in dLGN relay neurons, where it increases AMPAR-mediated current amplitudes and modulates gating of AMPARs. Importantly, CKAMP44 is responsible for the distinctive short-term depression in retinogeniculate synapses by reducing the rate of recovery from desensitization of AMPARs. Genetic deletion of CKAMP44 strongly reduces synaptic short-term depression, which leads to increased spike probability of relay neurons when activated with high-frequency inputs from retinogeniculate synapses. Finally, in vivo recordings reveal augmented ON- and OFF-responses of dLGN neurons in CKAMP44 knockout (CKAMP44^−/−) mice, demonstrating the importance of CKAMP44 for modulating synaptic short-term depression and visual input integration.

The function of receptor desensitization in vivo is not well understood. Here, the authors show that deletion of CKAMP44, an AMPAR auxiliary protein that modulates desensitization of AMPAR currents, affects synaptic facilitation at retinogeniculate synapses and visually-evoked firing in awake mice.

CONE: Convex-Optimized-Synaptic Efficacies for Temporally Precise Spike Mapping

Spiking neural networks are well suited to perform time-dependent pattern recognition problems by encoding the temporal dimension in precise spike times. With an appropriate set of weights, a spiking neuron can emit precisely timed action potentials in response to spatiotemporal input spikes. However, deriving supervised learning rules for spike mapping is nontrivial due to the increased complexity. Existing methods rely on heuristic approaches that do not guarantee a convex objective function and, therefore, may not converge to a global minimum. In this paper, we present a novel technique to obtain the weights of spiking neurons by formulating the problem in a convex optimization framework, rendering it be compatible with the established methods. We introduce techniques to influence the weight distribution and membrane trajectory, and then study how these factors affect robustness in the presence of noise. In addition, we show how the existence of a solution can be determined and assess memory capacity limits of a neuron model using synthetic examples. The practical utility of our technique is further assessed by its application to gait-event detection using the experimental data.

Circuit Mechanisms of a Retinal Ganglion Cell with Stimulus-Dependent Response Latency and Activation Beyond Its Dendrites

Center-surround antagonism has been used as the canonical model to describe receptive fields of retinal ganglion cells (RGCs) for decades. We describe a newly identified RGC type in the mouse, called the ON delayed (OND) RGC, with receptive field properties that deviate from center-surround organization. Responding with an unusually long latency to light stimulation, OND RGCs respond earlier as the visual stimulus increases in size. Furthermore, OND RGCs are excited by light falling far beyond their dendrites. We unravel details of the circuit mechanisms behind these phenomena, revealing new roles for inhibition in controlling both temporal and spatial receptive field properties. The non-canonical receptive field properties of the OND RGC-integration of long temporal and large spatial scales-suggest that unlike typical RGCs, it may encode a slowly varying, global property of the visual scene.

Using Matrix and Tensor Factorizations for the Single-Trial Analysis of Population Spike Trains

Advances in neuronal recording techniques are leading to ever larger numbers of simultaneously monitored neurons. This poses the important analytical challenge of how to capture compactly all sensory information that neural population codes carry in their spatial dimension (differences in stimulus tuning across neurons at different locations), in their temporal dimension (temporal neural response variations), or in their combination (temporally coordinated neural population firing). Here we investigate the utility of tensor factorizations of population spike trains along space and time. These factorizations decompose a dataset of single-trial population spike trains into spatial firing patterns (combinations of neurons firing together), temporal firing patterns (temporal activation of these groups of neurons) and trial-dependent activation coefficients (strength of recruitment of such neural patterns on each trial). We validated various factorization methods on simulated data and on populations of ganglion cells simultaneously recorded in the salamander retina. We found that single-trial tensor space-by-time decompositions provided low-dimensional data-robust representations of spike trains that capture efficiently both their spatial and temporal information about sensory stimuli. Tensor decompositions with orthogonality constraints were the most efficient in extracting sensory information, whereas non-negative tensor decompositions worked well even on non-independent and overlapping spike patterns, and retrieved informative firing patterns expressed by the same population in response to novel stimuli. Our method showed that populations of retinal ganglion cells carried information in their spike timing on the ten-milliseconds-scale about spatial details of natural images. This information could not be recovered from the spike counts of these cells. First-spike latencies carried the majority of information provided by the whole spike train about fine-scale image features, and supplied almost as much information about coarse natural image features as firing rates. Together, these results highlight the importance of spike timing, and particularly of first-spike latencies, in retinal coding.

Coding Properties of Three Intrinsically Distinct Retinal Ganglion Cells under Periodic Stimuli: A Computational Study

As the sole output neurons in the retina, ganglion cells play significant roles in transforming visual information into spike trains, and then transmitting them to the higher visual centers. However, coding strategies that retinal ganglion cells (RGCs) adopt to accomplish these processes are not completely clear yet. To clarify these issues, we investigate the coding properties of three types of RGCs (repetitive spiking, tonic firing, and phasic firing) by two different measures (spike-rate and spike-latency). Model results show that for periodic stimuli, repetitive spiking RGC and tonic RGC exhibit similar spike-rate patterns. Their spike- rates decrease gradually with increased stimulus frequency, moreover, variation of stimulus amplitude would change the two RGCs' spike-rate patterns. For phasic RGC, it activates strongly at medium levels of frequency when the stimulus amplitude is low. While if high stimulus amplitude is applied, phasic RGC switches to respond strongly at low frequencies. These results suggest that stimulus amplitude is a prominent factor in regulating RGCs in encoding periodic signals. Similar conclusions can be drawn when analyzes spike-latency patterns of the three RGCs. More importantly, the above phenomena can be accurately reproduced by Hodgkin's three classes of neurons, indicating that RGCs can perform the typical three classes of firing dynamics, depending on the distinctions of ion channel densities. Consequently, model results from the three RGCs may be not specific, but can also applicable to neurons in other brain regions which exhibit part(s) or all of the Hodgkin's three excitabilities.

In Vivo Imaging Reveals Composite Coding for Diagonal Motion in the Drosophila Visual System

Understanding information coding is important for resolving the functions of visual neural circuits. The motion vision system is a classic model for studying information coding as it contains a concise and complete information-processing circuit. In Drosophila, the axon terminals of motion-detection neurons (T4 and T5) project to the lobula plate, which comprises four regions that respond to the four cardinal directions of motion. The lobula plate thus represents a topographic map on a transverse plane. This enables us to study the coding of diagonal motion by investigating its response pattern. By using in vivo two-photon calcium imaging, we found that the axon terminals of T4 and T5 cells in the lobula plate were activated during diagonal motion. Further experiments showed that the response to diagonal motion is distributed over the following two regions compared to the cardinal directions of motion—a diagonal motion selective response region and a non-selective response region—which overlap with the response regions of the two vector-correlated cardinal directions of motion. Interestingly, the sizes of the non-selective response regions are linearly correlated with the angle of the diagonal motion. These results revealed that the Drosophila visual system employs a composite coding for diagonal motion that includes both independent coding and vector decomposition coding.

Modeling latency code processing in the electric sense: from the biological template to its VLSI implementation

Understanding the coding of sensory information under the temporal constraints of natural behavior is not yet well resolved. There is a growing consensus that spike timing or latency coding can maximally exploit the timing of neural events to make fast computing elements and that such mechanisms are essential to information processing functions in the brain. The electric sense of mormyrid fish provides a convenient biological model where this coding scheme can be studied. The sensory input is a physically ordered spatial pattern of current densities, which is coded in the precise timing of primary afferent spikes. The neural circuits of the processing pathway are well known and the system exhibits the best known illustration of corollary discharge, which provides the reference to decoding the sensory afferent latency pattern. A theoretical model has been constructed from available electrophysiological and neuroanatomical data to integrate the principal traits of the neural processing structure and to study sensory interaction with motor-command-driven corollary discharge signals. This has been used to explore neural coding strategies at successive stages in the network and to examine the simulated network capacity to reproduce output neuron responses. The model shows that the network has the ability to resolve primary afferent spike timing differences in the sub-millisecond range, and that this depends on the coincidence of sensory and corollary discharge-driven gating signals. In the integrative and output stages of the network, corollary discharge sets up a proactive background filter, providing temporally structured excitation and inhibition within the network whose balance is then modulated locally by sensory input. This complements the initial gating mechanism and contributes to amplification of the input pattern of latencies, conferring network hyperacuity. These mechanisms give the system a robust capacity to extract behaviorally meaningful features of the electric image with high sensitivity over a broad working range. Since the network largely depends on spike timing, we finally discuss its suitability for implementation in robotic applications based on neuromorphic hardware.

Feature-Linking Model for Image Enhancement

Inspired by gamma-band oscillations and other neurobiological discoveries, neural networks research shifts the emphasis toward temporal coding, which uses explicit times at which spikes occur as an essential dimension in neural representations. We present a feature-linking model (FLM) that uses the timing of spikes to encode information. The first spiking time of FLM is applied to image enhancement, and the processing mechanisms are consistent with the human visual system. The enhancement algorithm achieves boosting the details while preserving the information of the input image. Experiments are conducted to demonstrate the effectiveness of the proposed method. Results show that the proposed method is effective.

Rank Order Coding: a Retinal Information Decoding Strategy Revealed by Large-Scale Multielectrode Array Retinal Recordings

How a population of retinal ganglion cells (RGCs) encodes the visual scene remains an open question. Going beyond individual RGC coding strategies, results in salamander suggest that the relative latencies of a RGC pair encode spatial information. Thus, a population code based on this concerted spiking could be a powerful mechanism to transmit visual information rapidly and efficiently. Here, we tested this hypothesis in mouse by recording simultaneous light-evoked responses from hundreds of RGCs, at pan-retinal level, using a new generation of large-scale, high-density multielectrode array consisting of 4096 electrodes. Interestingly, we did not find any RGCs exhibiting a clear latency tuning to the stimuli, suggesting that in mouse, individual RGC pairs may not provide sufficient information. We show that a significant amount of information is encoded synergistically in the concerted spiking of large RGC populations. Thus, the RGC population response described with relative activities, or ranks, provides more relevant information than classical independent spike count- or latency- based codes. In particular, we report for the first time that when considering the relative activities across the whole population, the wave of first stimulus-evoked spikes is an accurate indicator of stimulus content. We show that this coding strategy coexists with classical neural codes, and that it is more efficient and faster. Overall, these novel observations suggest that already at the level of the retina, concerted spiking provides a reliable and fast strategy to rapidly transmit new visual scenes.

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.

Neural population coding: combining insights from microscopic and mass signals

Behavior relies on the distributed and coordinated activity of neural populations. Population activity can be measured using multi-neuron recordings and neuroimaging. Neural recordings reveal how the heterogeneity, sparseness, timing, and correlation of population activity shape information processing in local networks, whereas neuroimaging shows how long-range coupling and brain states impact on local activity and perception. To obtain an integrated perspective on neural information processing we need to combine knowledge from both levels of investigation. We review recent progress of how neural recordings, neuroimaging, and computational approaches begin to elucidate how interactions between local neural population activity and large-scale dynamics shape the structure and coding capacity of local information representations, make them state-dependent, and control distributed populations that collectively shape behavior.

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.

Comparison of the dynamics of neural interactions between current-based and conductance-based integrate-and-fire recurrent networks

Models of networks of Leaky Integrate-and-Fire (LIF) neurons are a widely used tool for theoretical investigations of brain function. These models have been used both with current- and conductance-based synapses. However, the differences in the dynamics expressed by these two approaches have been so far mainly studied at the single neuron level. To investigate how these synaptic models affect network activity, we compared the single neuron and neural population dynamics of conductance-based networks (COBNs) and current-based networks (CUBNs) of LIF neurons. These networks were endowed with sparse excitatory and inhibitory recurrent connections, and were tested in conditions including both low- and high-conductance states. We developed a novel procedure to obtain comparable networks by properly tuning the synaptic parameters not shared by the models. The so defined comparable networks displayed an excellent and robust match of first order statistics (average single neuron firing rates and average frequency spectrum of network activity). However, these comparable networks showed profound differences in the second order statistics of neural population interactions and in the modulation of these properties by external inputs. The correlation between inhibitory and excitatory synaptic currents and the cross-neuron correlation between synaptic inputs, membrane potentials and spike trains were stronger and more stimulus-modulated in the COBN. Because of these properties, the spike train correlation carried more information about the strength of the input in the COBN, although the firing rates were equally informative in both network models. Moreover, the network activity of COBN showed stronger synchronization in the gamma band, and spectral information about the input higher and spread over a broader range of frequencies. These results suggest that the second order statistics of network dynamics depend strongly on the choice of synaptic model.

To spike, or when to spike?

Recent experimental reports have suggested that cortical networks can operate in regimes were sensory information is encoded by relatively small populations of spikes and their precise relative timing. Combined with the discovery of spike timing dependent plasticity, these findings have sparked growing interest in the capabilities of neurons to encode and decode spike timing based neural representations. To address these questions, a novel family of methodologically diverse supervised learning algorithms for spiking neuron models has been developed. These models have demonstrated the high capacity of simple neural architectures to operate also beyond the regime of the well established independent rate codes and to utilize theoretical advantages of spike timing as an additional coding dimension.

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.