Cross-correlation analysis of the maintained discharge of rabbit retinal ganglion cells

On the Functional Role of Gamma Synchronization in the Retinogeniculate System of the Cat

Fast gamma oscillations, generated within the retina, and transmitted to the cortex via the lateral geniculate nucleus (LGN), are thought to carry information about stimulus size and continuity. This hypothesis relies mainly on studies conducted under anesthesia and the extent to which it holds under more naturalistic conditions remains unclear. Using multielectrode recordings of spiking activity in the retina and the LGN of both male and female cats, we show that visually driven gamma oscillations are absent for awake states and are highly dependent on halothane (or isoflurane). Under ketamine, responses were nonoscillatory, as in the awake condition. Response entrainment to the monitor refresh was commonly observed up to 120 Hz and was superseded by the gamma oscillatory responses induced by halothane. Given that retinal gamma oscillations are contingent on halothane anesthesia and absent in the awake cat, such oscillations should be considered artifactual, thus playing no functional role in vision.SIGNIFICANCE STATEMENT Gamma rhythms have been proposed to be a robust encoding mechanism critical for visual processing. In the retinogeniculate system of the cat, many studies have shown gamma oscillations associated with responses to static stimuli. Here, we extend these observations to dynamic stimuli. An unexpected finding was that retinal gamma responses strongly depend on halothane concentration levels and are absent in the awake cat. These results weaken the notion that gamma in the retina is relevant for vision. Notably, retinal gamma shares many of the properties of cortical gamma. In this respect, oscillations induced by halothane in the retina may serve as a valuable preparation, although artificial, for studying oscillatory dynamics.

Effects of tACS-Like Electrical Stimulation on Correlated Firing of Retinal Ganglion Cells: Part III

Purpose

Transcranial alternating current stimulation (tACS) is a stimulation protocol used for learning enhancement and mitigation of cognitive dysfunction. Correlated firing has been postulated to be a meta-code that links neuronal spike responses associated with a single entity and may be an important component of high-level cognitive functions. Thus, changes in the covariance firing structure of CNS neurons such as retinal ganglion cells are one potential mechanism by which tACS can exert its effects.

Materials and Methods

We used microelectrode arrays to record light-evoked spike responses of 24 retinal ganglion cells in 7 rabbit eyecup preparations and analyzed the covariance between 30 pairs of neighboring retinal ganglion cells before, during, and after 10-minute application of alternating currents of 1 microampere at 10 or 20 Hz.

Results

tACS stimulation significantly changed the covariance structure of correlated firing in 60% of simultaneously recorded retinal ganglion cells. Application of tACS in the retinal preparation increased cross-covariance in 26% of cell pairs, an effect usually associated with increased light-evoked ganglion cell firing. tACS associated decreases in cross-covariance occurred in 37% of cell pairs. Increased covariance was more common in response to the first, 10-minute application of tACS in isolated retina preparation. Changes in covariance were rare after repeated stimulation, and more likely to result in decreased covariance.

Conclusion

Retinal ganglion cell correlated firing is modulated by 1 microampere tACS currents showing that electrical stimulation can significantly and persistently change the structure of the correlated firing of simultaneously recorded rabbit retinal ganglion cells.

PTEN Expression Regulates Gap Junction Connectivity in the Retina

Manipulation of the phosphatase and tensin homolog (PTEN) pathway has been suggested as a therapeutic approach to treat or prevent vision loss due to retinal disease. In this study, we investigated the effects of deleting one copy of Pten in a well-characterized class of retinal ganglion cells called α-ganglion cells in the mouse retina. In Pten +/- retinas, α-ganglion cells did not exhibit major changes in their dendritic structure, although most cells developed a few, unusual loop-forming dendrites. By contrast, α-ganglion cells exhibited a significant decrease in heterologous and homologous gap junction mediated cell coupling with other retinal ganglion and amacrine cells. Additionally, the majority of OFF α-ganglion cells (12/18 cells) formed novel coupling to displaced amacrine cells. The number of connexin36 puncta, the predominant connexin that mediates gap junction communication at electrical synapses, was decreased by at least 50% on OFF α-ganglion cells. Reduced and incorrect gap junction connectivity of α-ganglion cells will affect their functional properties and alter visual image processing in the retina. The anomalous connectivity of retinal ganglion cells would potentially limit future therapeutic approaches involving manipulation of the Pten pathway for treating ganglion cell degeneration in diseases like glaucoma, traumatic brain injury, Parkinson's, and Alzheimer's diseases.

Cross-synaptic synchrony and transmission of signal and noise across the mouse retina

Cross-synaptic synchrony—correlations in transmitter release across output synapses of a single neuron—is a key determinant of how signal and noise traverse neural circuits. The anatomical connectivity between rod bipolar and A17 amacrine cells in the mammalian retina, specifically that neighboring A17s often receive input from many of the same rod bipolar cells, provides a rare technical opportunity to measure cross-synaptic synchrony under physiological conditions. This approach reveals that synchronization of rod bipolar cell synapses is near perfect in the dark and decreases with increasing light level. Strong synaptic synchronization in the dark minimizes intrinsic synaptic noise and allows rod bipolar cells to faithfully transmit upstream signal and noise to downstream neurons. Desynchronization in steady light lowers the sensitivity of the rod bipolar output to upstream voltage fluctuations. This work reveals how cross-synaptic synchrony shapes retinal responses to physiological light inputs and, more generally, signaling in complex neural networks.

DOI:http://dx.doi.org/10.7554/eLife.03892.001

The human eye is capable of detecting a single photon of starlight. This level of sensitivity is made possible by the high sensitivity of photoreceptors called rods. There are around 120 million rods in the retina, and they support vision in levels of light that are too low to activate the photoreceptors called cones that allow us to see in color. This is why we cannot see colors in the dark.

Signals are relayed through the retina via a circuit made up of multiple types of neurons. The activation of rods leads to activation of cells known as ‘rod bipolar cells’ which, in turn, activate amacrine cells and ganglion cells, with the latter sending signals via the optic nerve to the brain. All of these neurons communicate with one another at junctions called synapses. Activation of a rod bipolar cell, for example, triggers the release of molecules called neurotransmitters: these molecules bind to and activate receptors on the amacrine cells, enabling the signal to be transmitted.

For the brain to detect that a single photon has struck a rod, the eye must transmit information along this chain of neurons in a way that is highly reliable while adding very little noise to the signal. Grimes et al. have now revealed a key step in how this is achieved.

Electrical recordings from the mouse retina revealed that, in the dark, small fluctuations in the activity of rod bipolar cells lead to the near-deterministic release of neurotransmitters. This reduces the impact of random fluctuations in neurotransmitter release produced at individual synapses and ensures that the signals from rod bipolar cells (and thus from rods) are transmitted faithfully through the circuit with minimal added noise. As light levels increase, this tight synchrony of transmitter release breaks down, reducing the sensitivity to individual photons.

Given that many other brain regions share the features that enable retinal cells to coordinate the release of neurotransmitters, this mechanism might be used throughout the brain to increase the signal-to-noise ratio for the transmission of information through neural circuits.

DOI:http://dx.doi.org/10.7554/eLife.03892.002

Light increases the gap junctional coupling of retinal ganglion cells

We examined the effect of light adaptation on the gap junctional coupling of α-ganglion cells (α-GCs) in rabbit and mouse retinas. We assayed changes in coupling by measuring parameters of tracer coupling following injection of α-GCs with Neurobiotin and the concerted spike activity of α-GC neighbours under dark- and light-adapted conditions. We found that light adaptation using mesopic or photopic background lights resulted in a dramatic increase in the labelling intensity, number, and spatial extent of ganglion and amacrine cells coupled to OFF α-GCs when compared to levels seen under dark adaptation. While this augmentation of coupling by light did not produce an increase in the concerted spontaneous activity of OFF α-GC neighbours, it did significantly increase correlated light-evoked spiking. This was seen as an increase in the number of correlated spikes for α-GC neighbours and an extension of correlations to second-tier neighbours that was not seen under dark-adapted conditions. Pharmacological studies in the rabbit retina indicated that dopamine mediates the observed changes in coupling by differentially activating D1 and D2 receptors under different adaptation states. In this scheme, activation of dopamine D1 receptors following light exposure triggers cAMP-mediated intracellular pathways resulting in an increase in gap junctional conductance. Overall, our results indicate that as we move from night to day there is an enhanced electrical coupling between α-GCs, thereby increasing the concerted activity believed to strengthen the capacity and efficiency of information flow across the optic nerve.

Connexin36 is required for gap junctional coupling of most ganglion cell subtypes in the mouse retina

Converging evidence indicates that electrical synaptic transmission via gap junctions plays a crucial role in signal processing in the retina. In particular, amacrine and ganglion cells express numerous gap junctions, resulting in extensive electrical networks in the proximal retina. Both connexin36 (Cx36) and connexin45 (Cx45) subunits are widely distributed in the inner plexiform layer (IPL) and therefore are likely contribute to gap junctions formed by a number of ganglion cell subtypes. In the present study, we used the gap junction-permeant tracer Neurobiotin to compare the coupling pattern of different ganglion cell subtypes in wild-type (WT) and Cx36 knockout (KO) mouse retinas. We found that homologous ganglion-to-ganglion cell coupling was lost for two subtypes after deletion of Cx36, whereas two other ganglion cell subtypes retained homologous coupling in the KO mouse. In contrast, deletion of Cx36 resulted in a partial or complete loss of ganglion-to-amacrine cell heterologous coupling in 9 of 10 ganglion cell populations studied. Overall, our results indicate that Cx36 is the predominant subunit of gap junctions in the proximal mouse retina, expressed by most ganglion cell subtypes, and thereby likely plays a major role in the concerted activity generated by electrical synapses.

Functional connectivity map of retinal ganglion cells for retinal prosthesis

Retinal prostheses are being developed to restore vision for the blind with retinal diseases such as retinitis pigmentosa (RP) or age-related macular degeneration (AMD). Among the many issues for prosthesis development, stimulation encoding strategy is one of the most essential electrophysiological issues. The more we understand the retinal circuitry how it encodes and processes visual information, the greater it could help decide stimulation encoding strategy for retinal prosthesis. Therefore, we examined how retinal ganglion cells (RGCs) in in-vitro retinal preparation act together to encode a visual scene with multielectrode array (MEA). Simultaneous recording of many RGCs with MEA showed that nearby neurons often fired synchronously, with spike delays mostly within 1 ms range. This synchronized firing - narrow correlation - was blocked by gap junction blocker, heptanol, but not by glutamatergic synapse blocker, kynurenic acid. By tracking down all the RGC pairs which showed narrow correlation, we could harvest 40 functional connectivity maps of RGCs which showed the cell cluster firing together. We suggest that finding functional connectivity map would be useful in stimulation encoding strategy for the retinal prosthesis since stimulating the cluster of RGCs would be more efficient than separately stimulating each individual RGC.

The diverse functional roles and regulation of neuronal gap junctions in the retina

Electrical synaptic transmission through gap junctions underlies direct and rapid neuronal communication in the CNS. The diversity of functional roles that electrical synapses have is perhaps best exemplified in the vertebrate retina, in which gap junctions are formed by each of the five major neuron types. These junctions are dynamically regulated by ambient illumination and by circadian rhythms acting through light-activated neuromodulators such as dopamine and nitric oxide, which in turn activate intracellular signalling pathways in the retina.The networks formed by electrically coupled neurons are plastic and reconfigurable, and those in the retina are positioned to play key and diverse parts in the transmission and processing of visual information at every retinal level.

The structure of large-scale synchronized firing in primate retina

Synchronized firing among neurons has been proposed to constitute an elementary aspect of the neural code in sensory and motor systems. However, it remains unclear how synchronized firing affects the large-scale patterns of activity and redundancy of visual signals in a complete population of neurons. We recorded simultaneously from hundreds of retinal ganglion cells in primate retina, and examined synchronized firing in completely sampled populations of approximately 50-100 ON-parasol cells, which form a major projection to the magnocellular layers of the lateral geniculate nucleus. Synchronized firing in pairs of cells was a subset of a much larger pattern of activity that exhibited local, isotropic spatial properties. However, a simple model based solely on interactions between adjacent cells reproduced 99% of the spatial structure and scale of synchronized firing. No more than 20% of the variability in firing of an individual cell was predictable from the activity of its neighbors. These results held both for spontaneous firing and in the presence of independent visual modulation of the firing of each cell. In sum, large-scale synchronized firing in the entire population of ON-parasol cells appears to reflect simple neighbor interactions, rather than a unique visual signal or a highly redundant coding scheme.

Tracer coupling patterns of the ganglion cell subtypes in the mouse retina

It is now clear that electrical coupling via gap junctions is prevalent across the retina, expressed by each of the five main neuronal types. With the introduction of mutants in which selective gap junction connexins are deleted, the mouse has recently become an important model for studying the function of coupling between retinal neurons. In this study we examined the tracer-coupling pattern of ganglion cells by injecting them with the gap junction-permanent tracer Neurobiotin to provide, for the first time, a comprehensive survey of ganglion cell coupling in the wildtype mouse retina. Murine ganglion cells were differentiated into 22 morphologically distinct subtypes based on soma-dendritic parameters. Most (16/22) ganglion cell subtypes were tracer-coupled to neighboring ganglion and/or amacrine cells. The amacrine cells coupled to ganglion cells displayed either polyaxonal or wide-field morphologies with extensive arbors. We found that different subtypes of ganglion cells were never coupled to one another, indicating that they subserved independent electrical networks. Finally, we found that the tracer-coupling patterns of the 22 ganglion cell populations were largely stereotypic across the 71 retinas studied. Our results indicate that electrical coupling is extensive in the inner retina of the mouse, suggesting 0

Synchronized firing in the retina

Synchronized firing in neural populations has been proposed to constitute an elementary aspect of the neural code, but a complete understanding of its origins and significance has been elusive. Synchronized firing has been extensively documented in retinal ganglion cells, the output neurons of the retina. However, differences in synchronized firing across species and cell types have led to varied conclusions about its mechanisms and role in visual signaling. Recent work on two identified cell populations in the primate retina, the ON-parasol and OFF-parasol cells, permits a more unified understanding. Intracellular recordings reveal that synchronized firing in these cell types arises primarily from common synaptic input to adjacent pairs of cells. Statistical analysis indicates that local pairwise interactions can explain the pattern of synchronized firing in the entire parasol cell population. Computational analysis reveals that the aggregate impact of synchronized firing on the visual signal is substantial. Thus, in the parasol cells, the origin and impact of synchronized firing on the neural code may be understood as locally shared input which influences the visual signals transmitted from eye to brain.

Properties of stimulus-dependent synchrony in retinal ganglion cells

Neighboring retinal ganglion cells often spike synchronously, but the possible function and mechanism of this synchrony is unclear. Recently, the strength of the fast correlation between ON-OFF directionally selective cells of the rabbit retina was shown to be stimulus dependent. Here, we extend that study, investigating stimulus-dependent correlation among multiple ganglion-cell classes, using multi-electrode recordings. Our results generalized those for directionally selective cells. All cell pairs exhibiting significant spike synchrony did it for an extended edge but rarely for full-field stimuli. The strength of this synchrony did not depend on the amplitude of the response and correlations could be present even when the cells' receptive fields did not overlap. In addition, correlations tended to be orientation selective in a manner predictable by the relative positions of the receptive fields. Finally, extended edges and full-field stimuli produced significantly greater and smaller correlations than predicted by chance respectively. We propose an amacrine-network model for the enhancement and depression of correlation. Such an apparently purposeful control of correlation adds evidence for retinal synchrony playing a functional role in vision.

Signals and noise in an inhibitory interneuron diverge to control activity in nearby retinal ganglion cells

Information about sensory stimuli is represented by spatiotemporal patterns of neural activity. The complexity of the central nervous system, however, frequently obscures the origin and properties of signals and noise that underlie these activity patterns. We minimized this constraint by examining mechanisms governing correlated activity in mouse retinal ganglion cells (RGCs) under conditions in which light-evoked responses traverse a specific circuit, the rod bipolar pathway. Signals and noise in this circuit produced correlated synaptic input to neighboring On and Off RGCs. Temporal modulation of light intensity did not alter the degree to which noise in the input to nearby RGCs was correlated, and action potential generation in individual RGCs was largely insensitive to differences in network noise generated by dynamic and static light stimuli. Together, these features enable noise in shared circuitry to diminish simultaneous action potential generation in neighboring On and Off RGCs under a variety of conditions.

Simultaneity in the millisecond range as a requirement for effective shape recognition

Neurons of the visual system are capable of firing with millisecond precision, and synchrony of firing may provide a mechanism for "binding" stimulus elements in the image for purposes of recognition. While the neurophysiology is suggestive, there has been relatively little behavioral work to support the proposition that synchrony contributes to object recognition. The present experiments examined this issue by briefly flashing dots that were positioned at the outer boundary of namable objects, similar to silhouettes. Display of a given dot lasted only 0.1 ms, and temporal proximity of dot pairs, and among dot pairs, was varied as subjects were asked to name each object. In Exp 1, where the display of dots pairs was essentially simultaneous (0.2 ms to show both), there was a linear decline in recognition of the shapes as the interval between pairs increased from 0 ms to 6 ms. Compared with performance at 0 ms of delay, even the 2 ms interval between pairs produced a significant decrease in recognition. In Exp 2 the interval between pairs was constant at 3 ms, and the interval between pair members was varied. Here also a linear decline was observed as the interval between pair members increased from 0 ms to 1.5 ms, with the difference between 0 ms and 0.5 ms being significant. Thus minimal transient discrete cues can be integrated for purposes of shape recognition to the extent that they are synchronously displayed, and coincidence in the millisecond and even submillisecond range is needed for effective encoding of image data.

Can the theory of "whitening" explain the center-surround properties of retinal ganglion cell receptive fields?

To account for the spatial and temporal response properties of the retina, a number of studies have proposed that these properties serve to "whiten" the visual input. In particular, it has been argued that the sensitivity of retinal ganglion cells is matched to the spatial frequency spectrum of natural scenes, resulting in a flattened or "whitened" response spectrum across a range of frequencies. However, we argue that there are two distinct hypotheses regarding the flattening of the spectrum. The decorrelation hypothesis proposes that the magnitude of each ganglion cell tuning curve rises with spatial frequency, resulting in a flattened response spectrum for natural scene stimuli. With appropriate sampling, this scheme allows neighboring neurons to be uncorrelated with each other. The response equalization hypothesis proposes that the overall response magnitude of neurons increases with spatial frequency. The proposed goal of this model is to allow neurons with different receptive field sizes to produce the same average response to natural scenes. The response equalization hypothesis proposes an explanation for the relative gain of different ganglion cells and we show that this proposal fits well with published data. We suggest that both hypotheses are important in understanding the tuning and sensitivity of ganglion cells. However, using a simulation, both models are shown to be insufficient to explain the center-surround receptive field organization of ganglion cells. We discuss other factors, including representational sparseness, which could be related to the goals of ganglion cell spatial processing. We suggest three constraints needed to describe the basic linear properties of P-type ganglion cells: decorrelation, response equalization, and a minimal wiring or minimal size constraint.

Light-induced changes in spike synchronization between coupled ON direction selective ganglion cells in the mammalian retina

Although electrical coupling via gap junctions is prevalent among ganglion cells in the vertebrate retina, there have been few direct studies of their influence on the light-evoked signaling of these cells. Here, we describe the pattern and function of coupling between the ON direction selective (DS) ganglion cells, a unique subtype whose signals are transmitted to the accessory optic system (AOS) where they initiate the optokinetic response. ON DS cells are coupled indirectly via gap junctions made with a subtype of polyaxonal amacrine cell. This coupling underlies synchronization of the spontaneous and light-evoked spike activity of neighboring ON DS cells. However, we find that ON DS cell pairs show robust synchrony for all directions of stimulus movement, except for the null direction. Null stimulus movement evokes a GABAergic inhibition that temporally shifts firing of ON DS cell neighbors, resulting in a desynchronization of spike activity. Thus, detection of null stimulus movement appears key to the direction selectivity of ON DS cells, evoking both an attenuation of spike frequency and a desynchronization of neighbors. We posit that active desynchronization reduces summation of synaptic potentials at target AOS cells and thus provides a secondary mechanism by which ON DS cell ensembles can signal direction of stimulus motion to the brain.

Stimulus-dependent correlated firing in directionally selective retinal ganglion cells

Synchronous spiking has been postulated to be a meta-signal in visual cortex and other CNS loci that tags neuronal spike responses to a single entity. In retina, however, synchronized spikes have been postulated to arise via mechanisms that would largely preclude their carrying such a code. One such mechanism is gap junction coupling, in which synchronous spikes would be a by-product of lateral signal sharing. Synchronous spikes have also been postulated to arise from common-source inputs to retinal ganglion cells having overlapping receptive fields, and thus code for stimulus location in the overlap area. On-Off directionally selective ganglion cells of the rabbit retina exhibit a highly precise tiling pattern in which gap junction coupling occurs between some neighboring, same-preferred-direction cells. Depending on how correlated spikes arise, and for what purpose, one could postulate that synchronized spikes in this system (1) always arise in some subset of same-direction cells because of gap junctions, but never in non-same-preferred-directional cells; (2) never arise in same-directional cells because their receptive fields do not overlap, but arise only in different-directional cells whose receptive fields overlap, as a code for location in the overlap region; or (3) arise in a stimulus-dependent manner for both same- and different-preferred-direction cells for a function similar to that postulated for neurons in visual cortex. Simultaneous, extracellular recordings were obtained from neighboring On-Off directionally selective (DS) ganglion cells having the same and different preferred directions in an isolated rabbit retinal preparation. Stimulation by large flashing spots elicited responses from DS ganglion-cell pairs that typically showed little synchronous firing. Movement of extended bars, however, often produced synchronous spikes in cells having similar or orthogonal preferred directions. Surprisingly, correlated firing could occur for the opposite contrast polarity edges of moving stimuli when the leading edge of a sweeping bar excited the receptive field of one cell as its trailing edge stimulated another. Pharmacological manipulations showed that the spike synchronization is enhanced by excitatory cholinergic amacrine-cell inputs, and reduced by inhibitory GABAergic inputs, in a motion-specific manner. One possible interpretation is that this synchronous firing could be a signal to higher centers that the outputs of the two DS ganglion cells should be "bound" together as responding to a contour of a common object.

Spontaneous oscillatory activity of starburst amacrine cells in the mouse retina

Using patch-clamp techniques, we investigated the characteristics of the spontaneous oscillatory activity displayed by starburst amacrine cells in the mouse retina. At a holding potential of -70 mV, oscillations appeared as spontaneous, rhythmic inward currents with a frequency of approximately 3.5 Hz and an average maximal amplitude of approximately 120 pA. Application of TEA, a potassium channel blocker, increased the amplitude of oscillatory currents by >70% but reduced their frequency by approximately 17%. The TEA effects did not appear to result from direct actions on starburst cells, but rather a modulation of their synaptic inputs. Oscillatory currents were inhibited by 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX), an antagonist of AMPA/kainate receptors, indicating that they were dependent on a periodic glutamatergic input likely from presynaptic bipolar cells. The oscillations were also inhibited by the calcium channel blockers cadmium and nifedipine, suggesting that the glutamate release was calcium dependent. Application of AP4, an agonist of mGluR6 receptors on on-center bipolar cells, blocked the oscillatory currents in starburst cells. However, application of TEA overcame the AP4 blockade, suggesting that the periodic glutamate release from bipolar cells is intrinsic to the inner plexiform layer in that, under experimental conditions, it can occur independent of photoreceptor input. The GABA receptor antagonists picrotoxin and bicuculline enhanced the amplitude of oscillations in starburst cells prestimulated with TEA. Our results suggest that this enhancement was due to a reduction of a GABAergic feedback inhibition from amacrine cells to bipolar cells and the resultant increased glutamate release. Finally, we found that some ganglion cells and other types of amacrine cell also displayed rhythmic activity, suggesting that oscillatory behavior is expressed by a number of inner retinal neurons.

The potential coding utility of intercell cross-correlations in the retina

The action potentials (impulses) produced by pairs of neighboring retinal ganglion cells often show a tendency either to fire in close temporal synchrony or to avoid temporal synchrony. This cross-correlation (a rate of "coincidences" that differs from that expected by chance) has been exploited as a window into retinal processing, but its possible functional significance has proven elusive. Previous work has failed to show that the coincidences serve as a direct code for visual stimuli. In this analysis it is shown that the coincidences serve neither as a key for reducing variability nor as a key for improving the coding by the individual cells. The residual impulse trains (trains with coincidences deleted) are more variable than the raw impulse trains and provide an inferior coding to that of the raw impulse trains. There is negative correlation between the firing rate of the residual impulse trains and that of the coincidence impulse trains, which is consistent with the lower variance of the raw impulse trains. There is no consistent cross-correlation between the rates of residual impulse trains of cells in pairs showing cross-correlation; however, it is found that this observation does not discriminate among models for generating coincidences.

Functional connectivity of disparity-tuned neurons in the visual cortex

Different mechanisms have been proposed concerning how disparity-tuned neurons might be connected to produce the signals for depth perception. Here we present neurophysiological evidence providing insight on this issue. We have recorded simultaneously from pairs of disparity-tuned neurons in the cat's striate cortex. The purpose was to determine the relationships between disparity tuning and functional connectivity revealed through neural cross-correlograms. Monosynaptic connections tend to be stronger between pairs of cells with similar disparity tuning. Pairs of complex cells tend to have either similar tuning or nearly opposite tuning with an absence of quadrature relations. Pairs with at least one simple cell do have some nearly quadrature relationships when they are recorded from the same electrode. Coarse-to-fine connections (i.e., the presynaptic cell has lower disparity frequency and larger disparity range) tend to be stronger but less frequent than those of a fine-to-coarse nature. Our results are consistent with a system that produces weighted averaging across cells that are tuned to similar disparities but different disparity scales to reduce false matches.

Gap junctional coupling underlies the short-latency spike synchrony of retinal alpha ganglion cells

We examined whether coupling between neighboringalpha-type ganglion cells (alpha-GCs) in the rabbit retina underlies their synchronous spike activity. Simultaneous recordings were made from arrays of alpha-GCs to determine the synchrony of both spontaneous and light-evoked spike activity. One cell within each array was then injected with the biotinylated tracer Neurobiotin to determine which of the cells were coupled via gap junctions. Cross-correlation analyses indicated that neighboring off-center alpha-GCs maintain short-latency (approximately 2.5 msec) synchronous spiking, whereas the spontaneous spike activities of on-centeralpha-GC neighbors are not correlated. Without exception, those off-centeralpha-GCs showing synchronous spiking were found to be tracer coupled to both amacrine cells and neighboring off-centeralpha-GCs. In contrast, on-center alpha-GCs were never tracer coupled. Furthermore, whereas spikes initiated in an off-center alpha-GC with extrinsic current injection resulted in short-latency synchronized spiking in neighboring off-center alpha-GCs, this was never seen between on-center alpha-GCs. These results indicate that electrical coupling via gap junctions underlies the short-latency concerted spike activity of neighboring alpha-GCs.

Gap junctional coupling underlies the short-latency spike synchrony of retinal alpha ganglion cells

We examined whether coupling between neighboringalpha-type ganglion cells (alpha-GCs) in the rabbit retina underlies their synchronous spike activity. Simultaneous recordings were made from arrays of alpha-GCs to determine the synchrony of both spontaneous and light-evoked spike activity. One cell within each array was then injected with the biotinylated tracer Neurobiotin to determine which of the cells were coupled via gap junctions. Cross-correlation analyses indicated that neighboring off-center alpha-GCs maintain short-latency (approximately 2.5 msec) synchronous spiking, whereas the spontaneous spike activities of on-centeralpha-GC neighbors are not correlated. Without exception, those off-centeralpha-GCs showing synchronous spiking were found to be tracer coupled to both amacrine cells and neighboring off-centeralpha-GCs. In contrast, on-center alpha-GCs were never tracer coupled. Furthermore, whereas spikes initiated in an off-center alpha-GC with extrinsic current injection resulted in short-latency synchronized spiking in neighboring off-center alpha-GCs, this was never seen between on-center alpha-GCs. These results indicate that electrical coupling via gap junctions underlies the short-latency concerted spike activity of neighboring alpha-GCs.

A cooperation and competition based simple cell receptive field model and study of feed-forward linear and nonlinear contributions to orientation selectivity

We present a model for development of orientation selectivity in layer IV simple cells. Receptive field (RF) development in the model, is determined by diffusive cooperation and resource limited competition guided axonal growth and retraction in geniculocortical pathway. The simulated cortical RFs resemble experimental RFs. The receptive field model is incorporated in a three-layer visual pathway model consisting of retina, LGN and cortex. We have studied the effect of activity dependent synaptic scaling on orientation tuning of cortical cells. The mean value of hwhh (half width at half the height of maximum response) in simulated cortical cells is 58 degrees when we consider only the linear excitatory contribution from LGN. We observe a mean improvement of 22.8 degrees in tuning response due to the non-linear spiking mechanisms that include effects of threshold voltage and synaptic scaling factor.

Multineuronal firing patterns in the signal from eye to brain

Population codes in the brain have generally been characterized by recording responses from one neuron at a time. This approach will miss codes that rely on concerted patterns of action potentials from many cells. Here we analyze visual signaling in populations of ganglion cells recorded from the isolated salamander retina. These neurons tend to fire synchronously far more frequently than expected by chance. We present an efficient algorithm to identify what groups of cells cooperate in this way. Such groups can include up to seven or more neurons and may account for more than 50% of all the spikes recorded from the retina. These firing patterns represent specific messages about the visual stimulus that differ significantly from what one would derive by single-cell analysis.

Feed-forward synchronization: propagation of temporal patterns along the retinothalamocortical pathway

Visual responses in the cortex and lateral geniculate nucleus (LGN) are often associated with synchronous oscillatory patterning. In this short review, we examine the possible relationships between subcortical and cortical synchronization mechanisms. Our results obtained from simultaneous multi-unit recordings show strong synchronization of oscillatory responses between retina, LGN and cortex, indicating that cortical neurons can be synchronized by oscillatory activity relayed through the LGN. This feed-forward synchronization mechanism operating in the 60 to 120 Hz frequency range was observed mostly for static stimuli. In response to moving stimuli, by contrast, cortical synchronization was independent of oscillatory inputs from the LGN, with oscillation frequency in the range of 30 to 60 Hz. The functional implications of synchronization of activity from parallel channels are discussed, in particular its significance for signal transmission and cortical integration processes.

Firing coincidences between neighboring retinal ganglion cells: inside information or epiphenomenon?

Retinal ganglion cells often fire impulses in synchrony; is this synchronization an irrelevant by-product of processing shared inputs, or does it encode information? We examined the rate of occurrence of coincident impulses from pairs of ganglion cells responding to stimuli that varied along several dimensions. We find that coincidences convey little if any additional information about simple static stimuli beyond what could be determined from the firing rates of the two cells considered separately. In fact, at least one of the separate cells generally provided a better information channel than the coincidence rate, implying that under these conditions ganglion cells do not employ a strategy of encoding by coincidences.

Spontaneous retinal activity is tonic and does not drive tectal activity during activity-dependent refinement in regeneration

During development, waves of activity periodically spread across retina to produce correlated activity that is thought to drive activity-dependent ordering in optic fibers. We asked whether similar waves of activity are produced in the retina of adult goldfish during activity-dependent refinement by regenerating optic fibers. Dual-electrode recordings of spontaneous activity were made at different distances across retina but revealed no evidence of retinal waves in normal retina or during regeneration. Retinal activity was tonic and lacked the episodic bursting associated with waves. Cross-correlation analysis showed that the correlated activity that was normally restricted to near neighbors (typically seen across 100-200 microm and absent at >500 microm) was not altered during regeneration. The only change associated with regeneration was a twofold reduction in ganglion cell firing rates. Because spontaneous retinal activity is known to be sufficient to generate refinement during regeneration in goldfish, we examined its effect on tectal activity. In normal fish, acutely eliminating retinal activity with TTX rapidly reduced tectal unit activity by >90%. Surprisingly, during refinement at 4-6 weeks, eliminating retinal activity had no detectable effect on tectal activity. Similar results were obtained in recordings from torus longitudinalis. After refinement at 3 months, tectal activity was again highly dependent on ongoing retinal activity. We conclude that spontaneous retinal activity drives tectal cells in normal fish and after regeneration but not during activity-dependent refinement. The implications of these results for the role of presynaptic activity in refinement are considered.

Rate coding versus temporal order coding: what the retinal ganglion cells tell the visual cortex

It is often supposed that the messages sent to the visual cortex by the retinal ganglion cells are encoded by the mean firing rates observed on spike trains generated with a Poisson process. Using an information transmission approach, we evaluate the performances of two such codes, one based on the spike count and the other on the mean interspike interval, and compare the results with a rank order code, where the first ganglion cells to emit a spike are given a maximal weight. Our results show that the rate codes are far from optimal for fast information transmission and that the temporal structure of the spike train can be efficiently used to maximize the information transfer rate under conditions where each cell needs to fire only one spike.

High-frequency oscillations (20 to 120 Hz) and their role in visual processing

Oscillatory firing of neurons in response to visual stimuli has been observed to occur with different frequencies at multiple levels of the visual system. In the cat retina, oscillatory firing patterns occur with frequencies in the range of 60 to 120 Hz (omega-oscillations). These millisecond-precise temporal patterns are transmitted reliably to the cortex and may provide a feed-forward mechanism of response synchronization. In the cortex, visual responses often show oscillatory patterning with frequencies between 20 and 60 Hz (gamma-oscillations), which are not phase locked to the stimulus onset and therefore do not show up in regularly averaged evoked potentials. Gamma-oscillatory responses synchronize with millisecond precision over long distances and are mediated by the reciprocal corticocortical connectivity. Modulatory systems like the ascending reticular activating system facilitate synchronization and increase the strength of gamma-oscillations. During states of such functional cortical activation, the dominant frequency of the EEG is shifted from lower frequencies in the delta-/theta-range to higher frequencies in the gamma-range. Therefore, functional states indicate different degrees of temporal precision with which large neuronal populations interact. Response synchronization also depends on relations of global stimulus features. This suggests that millisecond-precise neuronal interactions serve as a fundamental mechanism for visual information processing.

Increased spontaneous unit activity and appearance of spontaneous negative potentials in the goldfish tectum during refinement of the optic projection

Spontaneous (not retinally driven) postsynaptic activity was examined during activity-dependent refinement of optic fibers in the goldfish tectum. Unit recordings in vivo and in vitro demonstrated that spontaneous tectal activity increased to 150% of normal during refinement at 1-2 months after optic nerve crush and subsequently returned to baseline over the next month. This increase was not mimicked by long-term denervation indicating an effect specifically influenced by regenerating fibers. Loss of optic input was also found to induce spontaneous negative potentials (SNPs) rapidly in the tectum. SNPs were negative, monophasic potentials of 70-120 msec duration and -0.15 to -1.5 mV amplitude. SNPs occurred with no apparent periodicity at a frequency of approximately 0.3-0.6 Hz. Multiple electrode recordings and depth analysis showed that SNPs were localized events occurring in columnar domains of tectum a few hundred micrometers wide. Cross-correlation analysis revealed that SNPs were strongly correlated with local unit bursting, suggesting SNPs are generated by the summed synaptic and spike currents of coactive cells in small regions of the tectum. SNPs were suppressed by a low concentration of APV indicating they were regulated by NMDA receptors. During regeneration, the number and size of SNPs reached a peak during refinement and subsequently decreased, eventually disappearing. This temporal association with refinement suggests that these patterns of postsynaptic activity may have functional relevance. It is hypothesized that SNPs or the underlying activity that produces them increases the excitability of target cells, allowing the weak, less-convergent input from regenerating axons to drive target groups of cells in the tectum during refinement.

Increased spontaneous unit activity and appearance of spontaneous negative potentials in the goldfish tectum during refinement of the optic projection

Spontaneous (not retinally driven) postsynaptic activity was examined during activity-dependent refinement of optic fibers in the goldfish tectum. Unit recordings in vivo and in vitro demonstrated that spontaneous tectal activity increased to 150% of normal during refinement at 1-2 months after optic nerve crush and subsequently returned to baseline over the next month. This increase was not mimicked by long-term denervation indicating an effect specifically influenced by regenerating fibers. Loss of optic input was also found to induce spontaneous negative potentials (SNPs) rapidly in the tectum. SNPs were negative, monophasic potentials of 70-120 msec duration and -0.15 to -1.5 mV amplitude. SNPs occurred with no apparent periodicity at a frequency of approximately 0.3-0.6 Hz. Multiple electrode recordings and depth analysis showed that SNPs were localized events occurring in columnar domains of tectum a few hundred micrometers wide. Cross-correlation analysis revealed that SNPs were strongly correlated with local unit bursting, suggesting SNPs are generated by the summed synaptic and spike currents of coactive cells in small regions of the tectum. SNPs were suppressed by a low concentration of APV indicating they were regulated by NMDA receptors. During regeneration, the number and size of SNPs reached a peak during refinement and subsequently decreased, eventually disappearing. This temporal association with refinement suggests that these patterns of postsynaptic activity may have functional relevance. It is hypothesized that SNPs or the underlying activity that produces them increases the excitability of target cells, allowing the weak, less-convergent input from regenerating axons to drive target groups of cells in the tectum during refinement.

Pattern formation by retinal afferents in the ferret lateral geniculate nucleus: developmental segregation and the role of N-methyl-D-aspartate receptors

The projection from the retina to the lateral geniculate nucleus (LGN) in ferrets segregates during development into eye-specific layers and ON/OFF sublayers. The projection pattern and the morphology of single axons was examined at several postnatal ages. The axons progress from a simple, sparsely branched morphology at birth to crude arbors at postnatal day 7 (P7). At P14-P15, axons have terminal arbors that span one eye-specific layer. By P19-P21, retinal afferents in the A layers have segregated into inner and outer sublaminae that correspond to ON- and OFF-center cells. Sublaminae form mainly by directed growth of terminal arbors in appropriately positioned regions of the LGN, along with elimination of extraneous branches in inappropriate regions. From P28 to P35, the LGN assumes an adult-like shape, and retinogeniculate axons form terminal boutons on branch endings. During the period between P14 and P21, when retinogeniculate axons segregate into ON/OFF sublaminae, N-methyl-D-aspartate (NMDA) receptors were blocked with chronic infusion of specific antagonists into the LGN. NMDA receptor blockade prevents the retinal afferent segregation into ON/OFF sublaminae. Some individual retinogeniculate axons have arbors that are not restricted appropriately, and most are restricted in size but are located inappropriately within the eye-specific laminae. Thus, NMDA receptor blockade prevents the positioning of retinogeniculate arbors that lead to the formation of ON/OFF sublaminae in the LGN. These results indicate that the activity of postsynaptic cells, and the activation of NMDA receptors in particular, can influence significantly the patterning of inputs and the structure of presynaptic afferents during development.

Synchronous oscillations in the cat retina

Retinal ganglion cells exhibit oscillatory responses which are precisely synchronized over large distances. Here we examined, with multi-electrode recordings, the time course of synchronization during spontaneous and stimulus-driven oscillatory activity. Spontaneous discharges showed synchronized oscillations at approximately 30 Hz, which were occasionally associated with slower superimposed oscillations in the range of 1-5 Hz. Stationary stimuli or moving gratings induced synchronous oscillations at higher frequencies (mean of 79.0 +/- 20.0 Hz for OFF- and 91.7 +/- 11.7 Hz for ON-responses), with time lags of a few milliseconds. At response onset, the first few oscillatory cycles were occasionally time locked to the stimulus. Thereafter, synchronization became independent of stimulus coordination and was exclusively due to neuronal interactions. Oscillatory modulation emerged rapidly and was sustained throughout the responses while oscillation frequency decreased gradually. This periodic patterning of responses persisted despite brief and local occlusion of stimuli, suggesting that synchronous oscillations emerge from population dynamics and entrain cells even if they are intermittently silenced.

Spontaneous activity in developing turtle retinal ganglion cells: pharmacological studies

Extracellular recordings were obtained from the ganglion cell (GC) layer during correlated spontaneous bursting activity (SBA) in the immature turtle retina. Pharmacological agents were bath-applied, and their effects on burst and correlation parameters were determined. SBA requires synaptic transmission. It was blocked in the presence of curare and mecamylamine, two cholinergic nicotinic antagonists, and enhanced with neostigmine, a cholinesterase inhibitor. SBA was profoundly inhibited during blockade of glutamatergic receptors with the broad spectrum antagonist kynurenate and it vanished with 6,7-dinitroquinoxaline-2-3-dione (DNQX) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), two AMPA/kainate receptor antagonists. Blockade of NMDA receptors with D(-)-2-amino-5-phosphonopentanoic acid (D-AP-5) led only to a modest reduction in SBA. Blockade of GABAA receptors with bicuculline prolonged the duration of the bursts. Inhibition of GABA uptake with nipecotic acid led to a decrease in burst rate. Blockade of K+ channels with cesium (Cs+) and tetraethylammonium (TEA) led to a dramatic decrease in excitability. Burst propagation between neighboring GCs was reduced by K+ channel blockade. Gap junction blockade had no consistent effect on bursts or correlation parameters. None of these drugs had a strong effect on the refractory period between bursts. We conclude that correlated SBA in immature turtle GCs requires both cholinergic nicotinic and glutamatergic (mainly through AMPA/kainate receptors) synaptic transmission. GABAergic activity modulates the intensity and the duration of the bursts. Extracellular K+ is involved in lateral activity propagation and increases retinal excitability, which may be required for burst generation.

Spontaneous activity in developing turtle retinal ganglion cells: pharmacological studies

Extracellular recordings were obtained from the ganglion cell (GC) layer during correlated spontaneous bursting activity (SBA) in the immature turtle retina. Pharmacological agents were bath-applied, and their effects on burst and correlation parameters were determined. SBA requires synaptic transmission. It was blocked in the presence of curare and mecamylamine, two cholinergic nicotinic antagonists, and enhanced with neostigmine, a cholinesterase inhibitor. SBA was profoundly inhibited during blockade of glutamatergic receptors with the broad spectrum antagonist kynurenate and it vanished with 6,7-dinitroquinoxaline-2-3-dione (DNQX) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), two AMPA/kainate receptor antagonists. Blockade of NMDA receptors with D(-)-2-amino-5-phosphonopentanoic acid (D-AP-5) led only to a modest reduction in SBA. Blockade of GABAA receptors with bicuculline prolonged the duration of the bursts. Inhibition of GABA uptake with nipecotic acid led to a decrease in burst rate. Blockade of K+ channels with cesium (Cs+) and tetraethylammonium (TEA) led to a dramatic decrease in excitability. Burst propagation between neighboring GCs was reduced by K+ channel blockade. Gap junction blockade had no consistent effect on bursts or correlation parameters. None of these drugs had a strong effect on the refractory period between bursts. We conclude that correlated SBA in immature turtle GCs requires both cholinergic nicotinic and glutamatergic (mainly through AMPA/kainate receptors) synaptic transmission. GABAergic activity modulates the intensity and the duration of the bursts. Extracellular K+ is involved in lateral activity propagation and increases retinal excitability, which may be required for burst generation.

Synchronous activity in the visual system

Synchronous activity among ensembles of neurons is a robust phenomenon observed in many regions of the brain. With the increased use of multielectrode recording techniques, synchronous firing of ensembles of neurons has been found at all levels in the mammalian visual pathway, from the retina to the extrastriate cortex. Here we distinguish three categories of synchrony in the visual system, (a) synchrony from anatomical divergence, (b) stimulus-dependent synchrony, and (c) emergent synchrony (oscillations). Although all three categories have been well documented, their functional significance remains uncertain. We discuss several lines of evidence both for and against a role for synchrony in visual processing: the perceptual consequences of synchronous activity, its ability to carry information, and the transmission of synchronous neural events to subsequent stages of processing.

Correlated firing in rabbit retinal ganglion cells

A ganglion cell's receptive field is defined as that region on the retinal surface in which a light stimulus will produce a response. While neighboring ganglion cells may respond to the same stimulus in a region where their receptive fields overlap, it generally has been assumed that each cell makes an independent decision about whether to fire. Recent recordings from cat and salamander retina using multiple electrodes have challenged this view of independent firing by showing that neighboring ganglion cells have an increased tendency to fire together within +/-5 ms. However, there is still uncertainty about which types of ganglion cells fire together, the mechanisms that produce coordinated spikes, and the overall function of coordinated firing. To address these issues, the responses of up to 80 rabbit retinal ganglion cells were recorded simultaneously using a multielectrode array. Of the 11 classes of rabbit ganglion cells previously identified, coordinated firing was observed in five. Plots of the spike train cross-correlation function suggested that coordinated firing occurred through two mechanisms. In the first mechanism, a spike in an interneuron diverged to produce simultaneous spikes in two ganglion cells. This mechanism predominated in four of the five classes including the ON brisk transient cells. In the second mechanism, ganglion cells appeared to activate each other reciprocally. This was the predominant pattern of correlated firing in OFF brisk transient cells. By comparing the receptive field profiles of ON and OFF brisk transient cells, a peripheral extension of the OFF brisk transient cell receptive field was identified that might be produced by lateral spike spread. Thus an individual OFF brisk transient cell can respond both to a light stimulus directed at the center of its receptive field and to stimuli that activate neighboring OFF brisk transient cells through their receptive field centers.

Roger Sperry and his chemoaffinity hypothesis

In the early 1940s, Roger Sperry performed a series of insightful experiments on the visual system of lower vertebrates that led him to draw two important conclusions: When optic fibers were severed, the regenerating fibers grew back to their original loci in the midbrain tectum to re-establish a topographical set of connections; and the re-establishment of these orderly connections underlay the orderly behavior of the animal. From these conclusions, he inferred that each optic fiber and each tectal neuron possessed cytochemical labels that uniquely denoted their neuronal type and position and that optic fibers could utilize these labels to selectively navigate to their matching target cell. This inference was subsequently formulated into a general explanation of how neurons form ordered interconnections during development and became known as the chemoaffinity hypothesis. The origins of this hypothesis, the controversies that surrounded it for several decades and its eventual acceptance, are discussed in this article.

Cross-correlation between neurons: a source of information about the nervous system

Impulse trains of spiking neurons are a stochastic process; however, the impulse trains of related neurons are generally not statistically independent. Cross-correlation may be due to either a common input that affects the firing of both cells, or to direct synaptic influence of the one neuron upon the other. By analyzing the distributions of intervals preceding and those following the coincident firings relative to that of all the intervals between impulses in each train, it is possible to infer restrictions upon the ways in which the correlating influence interacts with other sources of variability for each cell. This is applied to pairs of cells in two systems, one in which the coincident firing is due to an input common to the two cells, and one in which one cell is presynaptic to the other.

Population coding in the retina

Recent advances in multi-electrode recording have brought us closer to understanding how visual information is encoded by populations of retinal ganglion cells. By monitoring the visual responses of many ganglion cells at once, it is now possible to examine how ganglion cells act together to encode a visual scene.

Mechanisms of concerted firing among retinal ganglion cells

Nearby retinal ganglion cells often fire action potentials in near synchrony. We have investigated the circuit mechanisms that underlie these correlations by recording simultaneously from many ganglion cells in the salamander retina. During spontaneous activity in darkness, three types of correlations were distinguished: broad (firing synchrony within 40-100 ms), medium (10-50 ms), and narrow (<1 ms). When chemical synaptic transmission was blocked, the broad correlations disappeared, but the medium and narrow correlations persisted. Further analysis of the strength and time course of synchronous firing suggests that nearby ganglion cells share inputs from photoreceptors conveyed through interneurons via chemical synapses (broad correlations), share excitation from amacrine cells via electrical junctions (medium), and excite each other via electrical junctions (narrow). It appears that the firing patterns in the optic nerve are strongly shaped by electrical coupling in the inner retina.

An analysis of the cross-correlation between ganglion cells in the retina of goldfish

Neighboring ganglion cells in the retinae of vertebrates show cross-correlation between their impulse trains. Cross-correlation is found both in maintained discharges and during responses to visual stimulation. There has been speculation about the function of this statistical dependence, but little is known about its genesis. This study examines the statistics of the interimpulse intervals preceding and those following impulses that coincide with an impulse in the other train. Short intervals are rarer than expected preceding a coincidence, regardless of the form of the cross-correlation. Short intervals are more common than expected following a coincidence when the cross-correlation is positive, but rarer than expected following coincidences during negative cross-correlation. These results contradict the extant models for cross-correlation, but may be explained by the multiplicative combination of a variable common input and the variability within each cell. In addition, the lag (relative timing of coincident impulses in the two cells) is found to be related to the maintained discharge rates of the cells, implying that the lags may be explained without invoking specific delay circuits.

Long-range synchronization of oscillatory light responses in the cat retina and lateral geniculate nucleus

Visual responses in the retina and the lateral geniculate nucleus (LGN) exhibit oscillatory patterning within a broad range of frequencies. Oscillatory activity is often associated with the synchronization of spatially distributed responses. Here we demonstrate, with simultaneous multi-electrode recordings from the retina and the LGN, that stationary and moving light stimuli evoke in retinal ganglion cells oscillatory responses in the frequency range of 61 to 114 Hz that become synchronized over distances larger than 20 degrees of visual angle across the nasal and temporal halves of the retina. This temporal patterning of retinal responses is transmitted reliably by LGN neurons, such that stimuli crossing the vertical meridian evoke synchronous responses in the LGNs of both hemispheres. The oscillatory responses are not phase-locked to the stimulus onset, indicating that synchronization results from horizontal interactions in the retina. The occurrence of synchronization depends on global stimulus properties such as size and continuity, suggesting that temporal correlation among responses of spatially segregated ganglion cells can be exploited to convey information relevant for perceptual grouping.

Multineuronal codes in retinal signaling

The visual world is presented to the brain through patterns of action potentials in the population of optic nerve fibers. Single-neuron recordings show that each retinal ganglion cell has a spatially restricted receptive field, a limited integration time, and a characteristic spectral sensitivity. Collectively, these response properties define the visual message conveyed by that neuron's action potentials. Since the size of the optic nerve is strictly constrained, one expects the retina to generate a highly efficient representation of the visual scene. By contrast, the receptive fields of nearby ganglion cells often overlap, suggesting great redundancy among the retinal output signals. Recent multineuron recordings may help resolve this paradox. They reveal concerted firing patterns among ganglion cells, in which small groups of nearby neurons fire synchronously with delays of only a few milliseconds. As there are many more such firing patterns than ganglion cells, such a distributed code might allow the retina to compress a large number of distinct visual messages into a small number of optic nerve fibers. This paper will review the evidence for a distributed coding scheme in the retinal output. The performance limits of such codes are analyzed with simple examples, illustrating that they allow a powerful trade-off between spatial and temporal resolution.

Concerted signaling by retinal ganglion cells

To analyze the rules that govern communication between eye and brain, visual responses were recorded from an intact salamander retina. Parallel observation of many retinal ganglion cells with a microelectrode array showed that nearby neurons often fired synchronously, with spike delays of less than 10 milliseconds. The frequency of such synchronous spikes exceeded the correlation expected from a shared visual stimulus up to 20-fold. Synchronous firing persisted under a variety of visual stimuli and accounted for the majority of action potentials recorded. Analysis of receptive fields showed that concerted spikes encoded information not carried by individual cells; they may represent symbols in a multineuronal code for vision.

Effects of neonatal enucleation on the organization of callosal linkages in striate cortex of the rat

Lewis and Olavarria ([1995] J. Comp. Neurol. 361:119-137) showed that the mediolateral organization of callosal linkages differs markedly between medial and lateral regions of striate cortex in the rat. Thus, callosal fibers originating from medial regions of striate cortex interconnect loci that are mirror-symmetric with respect to the midsagittal plane. In contrast, fibers from lateral regions of striate cortex show a reversed pattern of connections: tracer injections into the 17/18a border produce retrograde cell labeling in regions medial to the contralateral 17/18a border, whereas injections placed somewhat medial to the 17/18a border label cells located at the contralateral 17/18a border. Based on the interpretation that callosal fibers from lateral striate cortex connect retinotopically corresponding loci (Lewis and Olavarria [1995] J. Comp. Neurol. 361:119-137) we propose here that the development of the reversed pattern of connections in lateral portions of striate cortex is guided by activity-dependent cues originating from spontaneously active ganglion cells in temporal retina. In the present study we have attempted to falsify this hypothesis by investigating the effects of neonatal bilateral enucleation on the organization of callosal linkages in striate cortex of the rat. Once enucleated rats reached adulthood, we studied the mediolateral organization of callosal connections by placing small injections of different fluorescent tracers into different loci within medial and lateral striate cortex. The analysis of the distribution of retrogradely labeled callosal cells indicated that connections from lateral portions of striate cortex were no longer organized in a reversed fashion, rather, they resembled the mirror image pattern normally found in the medial callosal region, i.e., injections at the 17/18a border produced labeled cells at the opposite 17/18a border, whereas injections into slightly more medial regions produced labeled cells in the opposite, mirror-symmetric location. In addition, we found that enucleation does not alter the organization of callosal linkages in medial portions of striate cortex. Thus, by showing that enucleation significantly changes the pattern of connections from lateral portions of striate cortex, the present study does not falsify, but rather strengthens the hypothesis that interhemispheric correlated activity driven from the temporal retinal crescent guides the normal development of reversed callosal linkages in lateral portions of rat striate cortex. Furthermore, the present study shows that, in the absence of the eyes, the pattern of callosal linkages in lateral portions of striate cortex resembles the mirror image pattern normally found only in medial striate cortex.

A map of visual space induced in primary auditory cortex

Maps of sensory surfaces are a fundamental feature of sensory cortical areas of the brain. The relative roles of afferents and targets in forming neocortical maps in higher mammals can be examined in ferrets in which retinal inputs are directed into the auditory pathway. In these animals, the primary auditory cortex contains a systematic representation of the retina (and of visual space) rather than a representation of the cochlea (and of sound frequency). A representation of a two-dimensional sensory epithelium, the retina, in cortex that normally represents a one-dimensional epithelium, the cochlea, suggests that the same cortical area can support different types of maps. Topography in the visual map arises both from thalamocortical projections that are characteristic of the auditory pathway and from patterns of retinal activity that provide the input to the map.

Spontaneous Activity as a Determinant of Axonal Connections

To investigate the role of spontaneous retinal activity in map refinement, we studied goldfish kept in darkness during regeneration of a cut optic nerve. In one experiment, such fish (with lenses ablated to blur vision) were maintained for 70 days in stroboscopic light, diurnal light, or total darkness interrupted daily by 15 minutes of stroboscopic light. The retinotectal projection was then assessed for retinotopy by standard methods, using retrograde transport of wheat germ agglutinin-horseradish peroxidase. As in previous work, significantly more refinement was found in diurnal than in stroboscopic light. In darkness, refinement was as complete as in diurnal light. In a second experiment, similar fish were kept in stroboscopic light for 63 days. Some were then assessed to confirm that refinement had been delayed, while others were transferred to darkness or diurnal light for assessment later. After 7 days in either environment, no further refinement was seen; but after 21 days, substantial and significant refinement has occurred in both. Thus the effects of darkness and diurnal light were indistinguishable, and very different from those of stroboscopic light and (in previous studies) tetrodotoxin. Map refinement is evidently activity-dependent but not experience-dependent, and can effectively use the correlated spontaneous firing of neighbouring ganglion cells as its basis. Locally correlated spontaneous activity, which appears also to drive eye- and class-specific axon segregation in mammals, occurs widely in the nervous system. It could potentially generate systematic interconnection patterns even between neuronal populations without an overtly topographic organization.

Activity-dependent tuning and the NMDA receptor

The refinement of the topographic map of visual space within the optic tectum of the frog is activity-dependent. The use of the three-eyed frog preparation to assay the operation of this fine-tuning mechanism indicates that this process is mediated by the NMDA receptor: Chronic in vivo treatment with APV, an NMDA antagonist, disrupts the segregation of retinal afferents into eye-specific zones while NMDA treatment sharpens this pattern. This latter effect is accompanied by a decreased sensitivity of the system to applied NMDA. Activation of the NMDA receptor may mediate the fine-tuning mechanism by initiating the stabilization of appropriate synapses. The requirements for NMDA receptor activation necessitate the convergence of terminals carrying correlated activity patterns. Such patterns of activity are provided by ganglion cells whose cell bodies lie near one another in the retina, and who should therefore, in an accurate visual map, terminate near one another in the tectum. Synapses from ganglion cells who do not neighbor one another in the retina have uncorrelated firing patterns and therefore do not activate the NMDA receptor. These synapses then would not be stabilized relative to one another. In addition to organizing the retinal projection, NMDA receptor activation may also modulate retinal ganglion cell arbor morphology, since chronic in vivo APV or NMDA treatments decrease arbor density. These results are discussed in terms of the effect of NMDA receptor activation on branch initiation and the rate of branch retraction.