Dendritic computation of direction selectivity by retinal ganglion cells

Regular mosaic of synaptic contacts among three retinal neurons

Retinal bipolar, amacrine, and ganglion cells contact each other within precisely defined synaptic laminae, but the spatial distribution of contacts between the cells is generally treated as random. Here we show that not to be the case. Excitatory inputs to inner retinal neurons were visualized by introduction of a plasmid coding for the postsynaptic protein PSD95-GFP. Our initial finding was that synapses on the dendrites of retinal ganglion cells are regularly spaced, at 2-3-μm intervals, along the dendrites. Thus, the presence of a PSD95 punctum creates a nearby zone from which other inputs appear to be excluded. Despite their great variation in size and different morphologies, the spacing is similar for the arbors of different retinal ganglion cell types. Regular spacing was also observed for the starburst amacrine cells. This regularity is mirrored in the spacing of axonal varicosities of the stratified bipolar cells, which have a regular, nonrandom interval consistent with that of the PSD95 puncta on ganglion cells. Thus, for each level of the inner plexiform layer all three cell types participate in a single 2D mosaic of synaptic contacts. These findings raise a new set of questions: How does the self-avoidance of synaptic sites along an individual dendrite arise and how is it physically maintained? Why is a regular spacing of inputs important for the computational function of the cells? Finally, which of the three players, if any, is developmentally responsible for the initial establishment of the pattern?

Separability of stimulus parameter encoding by on-off directionally selective rabbit retinal ganglion cells

The ganglion cell output of the retina constitutes a bottleneck in sensory processing in that ganglion cells must encode multiple stimulus parameters in their responses. Here we investigate encoding strategies of On-Off directionally selective retinal ganglion cells (On-Off DS RGCs) in rabbits, a class of cells dedicated to representing motion. The exquisite axial discrimination of these cells to preferred vs. null direction motion is well documented: it is invariant with respect to speed, contrast, spatial configuration, spatial frequency, and motion extent. However, these cells have broad direction tuning curves and their responses also vary as a function of other parameters such as speed and contrast. In this study, we examined whether the variation in responses across multiple stimulus parameters is systematic, that is the same for all cells, and separable, such that the response to a stimulus is a product of the effects of each stimulus parameter alone. We extracellularly recorded single On-Off DS RGCs in a superfused eyecup preparation while stimulating them with moving bars. We found that spike count responses of these cells scaled as independent functions of direction, speed, and luminance. Moreover, the speed and luminance functions were common across the whole sample of cells. Based on these findings, we developed a model that accurately predicted responses of On-Off DS RGCs as products of separable functions of direction, speed, and luminance (r = 0.98; P < 0.0001). Such a multiplicatively separable encoding strategy may simplify the decoding of these cells' outputs by the higher visual centers.

Asymmetric inhibitory connections enhance directional selectivity in a three-layer simulation model of retinal networks

In this paper, we found that spatial and temporal asymmetricity of excitatory connections are able to generate directional selectivity which can be enhanced by asymmetrical inhibitory connections by reconstructing a hexagonally-arranged three-layered simulation model of retina by NEURON simulator. Asymmetric excitatory inputs to ganglion cells with randomly arborizing dendrites were able to generate weaker directional selectivity to moving stimuli whose speed was less than 10 μm/msec. By just adding asymmetric inhibitory connections via inhibitory amacrine cells, directional selectivity became stronger to respond to moving stimuli at ten times faster speed (< 100 μm/msec). In conclusion, an excitatory mechanism appeared to generate directional selectivity while asymmetric inhibitory connections enhance directional selectivity in retina.

Mechanisms of direction selectivity in cat primary visual cortex as revealed by visual adaptation

In contrast to neurons of the lateral geniculate nucleus (LGN), neurons in the primary visual cortex (V1) are selective for the direction of visual motion. Cortical direction selectivity could emerge from the spatiotemporal configuration of inputs from thalamic cells, from intracortical inhibitory interactions, or from a combination of thalamic and intracortical interactions. To distinguish between these possibilities, we studied the effect of adaptation (prolonged visual stimulation) on the direction selectivity of intracellularly recorded cortical neurons. It is known that adaptation selectively reduces the responses of cortical neurons, while largely sparing the afferent LGN input. Adaptation can therefore be used as a tool to dissect the relative contribution of afferent and intracortical interactions to the generation of direction selectivity. In both simple and complex cells, adaptation caused a hyperpolarization of the resting membrane potential (-2.5 mV, simple cells, -0.95 mV complex cells). In simple cells, adaptation in either direction only slightly reduced the visually evoked depolarization; this reduction was similar for preferred and null directions. In complex cells, adaptation strongly reduced visual responses in a direction-dependent manner: the reduction was largest when the stimulus direction matched that of the adapting motion. As a result, adaptation caused changes in the direction selectivity of complex cells: direction selectivity was reduced after preferred direction adaptation and increased after null direction adaptation. Because adaptation in the null direction enhanced direction selectivity rather than reduced it, it seems unlikely that inhibition from the null direction is the primary mechanism for creating direction selectivity.

Tetrodotoxin-resistant sodium channels contribute to directional responses in starburst amacrine cells

The biophysical mechanisms that give rise to direction selectivity in the retina remain uncertain. Current evidence suggests that the directional signal first arises within the dendrites of starburst amacrine cells (SBACs). Two models have been proposed to explain this phenomenon, one based on mutual inhibitory interactions between SBACs, and the other positing an intrinsic dendritic mechanism requiring a voltage-gradient depolarizing towards the dendritic tips. We tested these models by recording current and voltage responses to visual stimuli in SBACs. In agreement with previous work, we found that the excitatory currents in the SBACs were directional, and remained directional when GABA receptors were blocked. Contrary to the mutual-inhibitory model, stimuli that produce strong directional signals in ganglion cells failed to reveal a significant inhibitory input to SBACs. Suppression of the tonic excitatory conductance, proposed to generate the dendritic voltage-gradient required for the dendrite autonomous model, failed to eliminate the directional signal in SBACs. However, selective block of tetrodotoxin-resistant sodium channels did reduce the strength of the directional excitatory signal in the SBACs. These results indicate that current models of direction-selectivity in the SBACs are inadequate, and suggest that voltage-gated excitatory channels, specifically tetrodotoxin-resistant sodium channels, are important elements in directional signaling. This is the first physiological evidence that tetrodotoxin-resistant sodium channels play a role in retinal information processing.

Dendritic spikes amplify the synaptic signal to enhance detection of motion in a simulation of the direction-selective ganglion cell

The On-Off direction-selective ganglion cell (DSGC) in mammalian retinas responds most strongly to a stimulus moving in a specific direction. The DSGC initiates spikes in its dendritic tree, which are thought to propagate to the soma with high probability. Both dendritic and somatic spikes in the DSGC display strong directional tuning, whereas somatic PSPs (postsynaptic potentials) are only weakly directional, indicating that spike generation includes marked enhancement of the directional signal. We used a realistic computational model based on anatomical and physiological measurements to determine the source of the enhancement. Our results indicate that the DSGC dendritic tree is partitioned into separate electrotonic regions, each summing its local excitatory and inhibitory synaptic inputs to initiate spikes. Within each local region the local spike threshold nonlinearly amplifies the preferred response over the null response on the basis of PSP amplitude. Using inhibitory conductances previously measured in DSGCs, the simulation results showed that inhibition is only sufficient to prevent spike initiation and cannot affect spike propagation. Therefore, inhibition will only act locally within the dendritic arbor. We identified the role of three mechanisms that generate directional selectivity (DS) in the local dendritic regions. First, a mechanism for DS intrinsic to the dendritic structure of the DSGC enhances DS on the null side of the cell's dendritic tree and weakens it on the preferred side. Second, spatially offset postsynaptic inhibition generates robust DS in the isolated dendritic tips but weak DS near the soma. Third, presynaptic DS is apparently necessary because it is more robust across the dendritic tree. The pre- and postsynaptic mechanisms together can overcome the local intrinsic DS. These local dendritic mechanisms can perform independent nonlinear computations to make a decision, and there could be analogous mechanisms within cortical circuitry.

The On-Off direction-selective ganglion cell (DSGC) found in mammalian retinas generates a directional signal, responding most strongly to a stimulus moving in a specific direction. The DSGC initiates spikes in its dendritic tree which are thought to propagate to the soma and brain with high probability. Both dendritic and somatic spikes in the DSGC display strong directional tuning, whereas postsynaptic potentials (PSPs) recorded in the soma are only weakly directional, indicating that postsynaptic spike generation markedly enhances the directional signal. We constructed a realistic computational model to determine the source of the enhancement. Our results indicate that the DSGC dendritic tree is partitioned into separate computational regions. Within each region, the local spike threshold produces nonlinear amplification of the preferred response over the null response on the basis of PSP amplitude. The simulation results showed that inhibition acts locally within the dendritic arbor and will not stop dendritic spikes from propagating. We identified the role of three mechanisms that generate direction selectivity in the local dendritic regions, which suggests the origin of the previously described “non-direction-selective region,” and also suggests that the known DS in the synaptic inputs is apparently necessary for robust DS across the dendritic tree.

Selective activation of mGluR8 receptors modulates retinal ganglion cell light responses

Extracellular and whole-cell light-evoked responses of mouse retinal ganglion cells were recorded in the presence of the mGluR8 selective agonist, (S)-3,4-dicarboxy-phenylglycine (DCPG). Off-light responses were reversibly reduced in the presence of DCPG in wild-type but not in mGluR8-deficient retinas. On-responses were only marginally modulated by DCPG. During Off-responses, DCPG suppressed both excitatory and inhibitory synaptic conductances suggesting that mGluR8 receptor activity reduces glutamate release from bipolar cell terminals and possibly also the release of an inhibitory neurotransmitter from amacrine cell processes.

Spinogenesis and pruning from early visual onset to adulthood: an intracellular injection study of layer III pyramidal cells in the ventral visual cortical pathway of the macaque monkey

Neocortical pyramidal cells are characterized by markedly different structure among cortical areas in the mature brain. In the ventral visual pathway of adult primates, pyramidal cells become increasingly more branched and more spinous with anterior progression through the primary (V1), second (V2), and fourth (V4) visual areas and cytoarchitectonic areas TEO and TE. It is not known how these regional specializations in neuron structure develop. Here, we report that the basal dendritic trees of layer III pyramidal cells in V1, V2, V4, TEO, and TE were characterized by unique growth profiles. Different numbers of spines were grown in the dendritic trees of cells among these cortical areas and then subsequently pruned. In V1, V2, and V4, more spines were pruned than grew resulting in a net decrease in the number of spines in the dendritic trees following the onset of visual experience. In TEO and TE, neurons grew more spines than they pruned from visual onset to adulthood. These data suggest that visual experience may influence neuronal maturation in different ways in different cortical areas.

Differences in the time course of short-term depression across receptive fields are correlated with directional selectivity in electrosensory neurons

Directional selectivity, in which neurons respond preferentially to one direction of movement ("preferred") over the opposite direction ("null"), is a critical computation that is found in the nervous systems of many animals. Here we show the first experimental evidence for a correlation between differences in short-term depression and direction-selective responses to moving objects. As predicted by quantitative models, the observed differences in the time courses of short-term depression at different locations within receptive fields were correlated with measures of direction selectivity in awake, behaving weakly electric fish (Apteronotus leptorhynchus). Because short-term depression is ubiquitous in the central nervous systems of vertebrate animals, it may be a common mechanism used for the generation of directional selectivity and other spatiotemporal computations.

Synaptic physiology of direction selectivity in the retina

Detection of the direction of image movement is accomplished first in the retina by an elegant neuronal circuit, which integrates multiple levels of spatially asymmetric synaptic interactions among subsets of bipolar, amacrine and ganglion cells. Central to these interactions is the asymmetric GABAergic inhibition exerted by the starburst amacrine cell (SAC), a cholinergic and GABAergic interneuron with a radially symmetric dendritic tree. SACs make reciprocal GABAergic synapses on each other to create a direct inhibitory receptive field surround, which suppresses the response of each SAC to centripetal image movement. Each radially projecting branch of a SAC responds to image movement with a centrifugal bias and, through directionally asymmetric synaptic connections with the dendrites of direction-selective ganglion cells (DSGCs), exerts a spatially offset inhibition that vetoes the response of DSGCs to image movement in a specific (null) direction. Recent physiological studies have greatly advanced our understanding of the mechanism of direction selectivity and also revealed a new level of complexity that remains to be understood.

Functional stability of retinal ganglion cells after degeneration-induced changes in synaptic input

Glutamate released from photoreceptors controls the activity and output of parallel pathways in the retina. When photoreceptors die because of degenerative diseases, surviving retinal networks are left without their major source of input, but little is known about how photoreceptor loss affects ongoing synaptic activity and retinal output. Here, we use patch-clamp recording and two-photon microscopy to investigate morphological and physiological properties of identified types of ON and OFF retinal ganglion cells (RGCs) in the adult (36-210 d old) retinal degeneration rd-1/rd-1 mouse. We find that strong rhythmic synaptic input drives ongoing oscillatory spike activity in both ON and OFF RGCs at a fundamental "beating" frequency of approximately 10 Hz. Despite this aberrant activity, ON and OFF cells maintain their characteristic dendritic stratification, intrinsic firing properties, including rebound firing in OFF cells, balance of synaptic excitation and inhibition, and dendritic calcium signaling. Thus, RGCs are inherently stable during degeneration-induced retinal activity.

Elevated intracellular chloride level in albino visual cortex neurons is mediated by Na-K-Cl co-transporter

Background

During development the switch from a depolarizing to a hyperpolarizing action of GABA is a consequence of a decrease of the Na⁺-K⁺-2Cl^- co-transporter (NKCC1, Cl^--uptake) and increase of the K⁺-Cl^- co-transporter (KCC2, Cl^--extrusion) expression. However albino visual cortex neurons don't show a corresponding decrease in intracellular chloride concentration during development of the visual system as compared to pigmented animals.

Results

Our study revealed that more cells express NKCC1 in albinos compared to pigmented rat visual cortex neurons whereas KCC2 is expressed in all cells in both strains. We determined a positive relationship between the presence of NKCC1 and an inhibitory deficit in single neurons of the albino visual cortex. After pharmacological blockade of NKCC1 function with its specific inhibitor, bumetanide, the reversal potential of electrically evoked GABAA receptor-mediated postsynaptic currents and, as a consequence, [Cl-]_i in albino visual cortex neurons shifted to the pigmented rat brain value. In conclusion, our pharmacological experiments and subsequent single cell real time PCR analysis of the co-transporter mRNA demonstrated that the inhibitory deficit present in the albino visual cortical network is almost exclusively mediated by NKCC1.

Conclusion

Our findings suggest that blocking of NKCC1 in albino visual cortex neurons could improve processing in visual cortex and therefore might be beneficial for vision in albinos.

Symmetric synaptic patterns between starburst amacrine cells and direction selective ganglion cells in the rabbit retina

Inputs from starburst amacrine cells (SACs) to ON-OFF direction selective ganglion cells (DSGCs) in the rabbit retina are themselves directional. However, the synaptic asymmetry between SACs and DSGCs required for generating direction selectivity has been controversial. We investigated dendritic contacts and distribution of inhibitory synapses between SACs and their overlapped DSGCs. Double injection of SAC/DSGC pairs and quantitative analysis revealed no obvious asymmetry of dendritic contacts between SACs and DSGCs. Furthermore, examination of the inhibitory input pattern on the dendrites of DSGCs using antibodies against GABA(A) receptors also suggested an isotropic arrangement with the overlapping SACs in both the preferred and the null directions. Therefore, the presynaptic mechanism of direction selectivity upon DSGCs may not result from a simple asymmetric arrangement with overlapping SACs. Multiple layer interactions and sophisticated synaptic connections between SACs and DSGCs are necessary.

Robust directional computation in on-off directionally selective ganglion cells of rabbit retina

The spatial and temporal interactions in the receptive fields of On-Off directionally selective (DS) ganglion cells endow them with directional selectivity. Using a variety of stimuli, such as sinusoidal gratings, we show that these interactions make directional selectivity of the DS ganglion cell robust with respect to stimulus parameters such as contrast, speed, spatial frequency, and extent of motion. Moreover, unlike the directional selectivity of striate-cortex cells, On-Off DS ganglion cells display directional selectivity to motions not oriented perpendicularly to the contour of the objects. We argue that these cells may achieve such high robustness by combining multiple mechanisms of directional selectivity.

Cellular mechanisms for direction selectivity in the retina

Direction selectivity represents a fundamental computation found across multiple sensory systems. In the mammalian visual system, direction selectivity appears first in the retina, where excitatory and inhibitory interneurons release neurotransmitter most rapidly during movement in a preferred direction. Two parallel sets of interneuron signals are integrated by a direction-selective ganglion cell, which creates a direction preference for both bright and dark moving objects. Direction selectivity of synaptic input becomes amplified by action potentials in the ganglion cell dendrites. Recent work has elucidated direction-selective mechanisms in inhibitory circuitry, but mechanisms in excitatory circuitry remain unexplained.

Characterization of Voltage-Gated Ionic Channels in Cholinergic Amacrine Cells in the Mouse Retina

Recent studies have shown that cholinergic amacrine cells possess unique membrane properties. However, voltage-gated ionic channels in cholinergic amacrine cells have not been characterized systematically. In this study, using electrophysiological and immunohistochemical techniques, we examined voltage-gated ionic channels in a transgenic mouse line the cholinergic amacrine cells of which were selectively labeled with green fluorescent protein (GFP). Voltage-gated K+ currents contained a 4-aminopyridine-sensitive current (A current) and a tetraethylammonium-sensitive current (delayed rectifier K+ current). Voltage-gated Ca2+ currents contained a ω-conotoxin GVIA-sensitive component (N-type) and a ω-Aga IVA-sensitive component (P/Q-type). Tetrodotoxin-sensitive Na+ currents and dihydropyridine-sensitive Ca2+ currents (L-type) were not observed. Immunoreactivity for the Na channel subunit (Pan Nav), the K channel subunits (the A-current subunits [Kv. 3.3 and Kv 3.4]) and the Ca channel subunits (α1A [P/Q-type], α1B [N-type] and α1C [L-type]) was detected in the membrane fraction of the mouse retina by Western blot analysis. Immunoreactivity for the Kv. 3.3, Kv 3.4, α1A [P/Q-type], and α1B [N-type] was colocalized with the GFP signals. Immunoreactivity for α1C [L-type] was not colocalized with the GFP signals. Immunoreactivity for Pan Nav did not exist on the membrane surface of the GFP-positive cells. Our findings indicate that signal propagation in cholinergic amacrine cells is mediated by a combination of two types of voltage-gated K+ currents (the A current and the delayed rectifier K+ current) and two types of voltage-gated Ca2+ currents (the P/Q-type and the N-type) in the mouse retina.

The potential importance of saturating and supersaturating contrast response functions in visual cortex

Most cortical visual neurons do not respond linearly with contrast. Generally, they show saturated responses to stimuli of high contrast, a feature often characterized by a divisive normalization function. This nonlinearity is generally thought to be useful in focusing the dynamic response range of the neuron on a particular region of contrast space, optimizing contrast gain. Some neurons not only saturate but also supersaturate; at high contrast, the response of the neuron decreases rather than plateaus. Under the contrast gain control theory, these cells would seem to reflect a nonoptimal normalization pool that provides excessive inhibition to the neurons. Since very few data on supersaturation are available, this article examines the frequency with which such neurons occur in macaque visual cortex by considering an extension of the Naka-Rushton equation with the capacity to represent nonmonotonic functions. The prevalence of gain-control theories for saturation has occluded an additional computational function for saturation, namely, in detecting the conjunction of certain features. A saturating nonlinearity is a critical part of the selective detection of compound stimuli over their components. In this role, the existence of saturating contrast response functions might be considered necessary rather than simply optimal.

Identification of synaptic pattern of kainate glutamate receptor subtypes on direction-selective retinal ganglion cells

In this article we investigate the distributions of kainate glutamate receptor subtypes GluR5-7 and KA1, 2 on the dendritic arbors of direction-selective (DS) retinal ganglion cells (RGCs) of the rabbit retina to search for anisotropies, which might contribute to a directional preference of DS RGCs. The distribution of the kainate receptor subunits on the DS RGCs was determined using antibody immunocytochemistry. DS RGCs were injected with Lucifer yellow and the cells were identified by their characteristic morphology. The double-labeled images of dendrites and receptors were visualized using confocal microscopy and were reconstructed from high-resolution confocal images. We found no evidence of asymmetry in any of the kainate receptor subunits examined on the dendritic arbors of both On and Off layers of DS RGCs. Our results indicate that direction selectivity appears to lie in the neuronal circuitry afferent to the ganglion cell.

Receptive field structure-function correlates in developing turtle retinal ganglion cells

Mature retinal ganglion cells (RGCs) have distinct morphologies that often reflect specialized functional properties such as On and Off responses. But the structural correlates of many complex receptive field (RF) properties (e.g. responses to motion) remain to be deciphered. In this study, we have investigated whether motion anisotropies (non-homogeneities) characteristic of embryonic turtle RGCs arise from immature dendritic arborization in these cells. To test this hypothesis, we have looked at structure-function correlates of developing turtle RGCs from Stage 23 (S23) when light responses emerge, until 15 weeks post-hatching (PH). Using whole cell patch clamp recordings, RGCs were labelled with Lucifer Yellow (LY) while recording their responses to moving edges of light. Comparison of RF and dendritic arbor layouts revealed a weak correlation. To obtain a larger structural sample of developing RGCs, we have looked at dendritic morphology in RGCs retrogradely filled with the tracer horseradish peroxidase (HRP) from S22 (when RGCs become spontaneously active, shortly before they become sensitive to light) until two weeks PH. We found that there was intense dendritic growth from S22 onwards, reaching peak proliferation at S25 (a week before hatching), while RGCs are still exhibiting significant motion anisotropies. Based on these observations, we suggest that immature anisotropic RGC RFs must originate from sparse synaptic inputs onto RGCs rather than from the immaturity of their growing dendritic trees.

Synaptic pattern of AMPA receptor subtypes upon direction-selective retinal ganglion cells

In the search for anisotropies that might contribute to a directional preference of direction-selective (DS) retinal ganglion cells (RGCs), we studied the distributions of AMPA receptor subtypes GluR1, GluR2/3, and GluR4 upon the dendritic arbors of DS RGCs of the rabbit with antibody immunocytochemistry. DS RGCs were injected with Lucifer yellow and the cells were identified by their characteristic morphology. The double-labeled images of dendrites and receptors were visualized by confocal microscopy and were reconstructed from high-resolution confocal images. We found no evidence of asymmetry in any of the AMPA receptor subunits examined upon the dendritic arbors of both On and Off layers of DS RGCs. The present results indicate that direction selectivity appears to lie in presynaptic pattern.

Direction-selective dendritic action potentials in rabbit retina

Dendritic spikes that propagate toward the soma are well documented, but their physiological role remains uncertain. Our in vitro patch-clamp recordings and two-photon calcium imaging show that direction-selective retinal ganglion cells (DSGCs) utilize orthograde dendritic spikes during physiological activity. DSGCs signal the direction of image motion. Excitatory subthreshold postsynaptic potentials are observed in DSGCs for motion in all directions and provide a weakly tuned directional signal. However, spikes are generated over only a narrow range of motion angles, indicating that spike generation greatly enhances directional tuning. Our results indicate that spikes are initiated at multiple sites within the dendritic arbors of DSGCs and that each dendritic spike initiates a somatic spike. We propose that dendritic spike failure, produced by local inhibitory inputs, might be a critical factor that enhances directional tuning of somatic spikes.

BIOPHYSICAL MECHANISMS AND ESSENTIAL TOPOGRAPHY OF DIRECTIONALLY SELECTIVE SUBUNITS IN RABBIT'S RETINA

We commemorate the 40th anniversary of the classical study undertaken by Barlow-Levick with a new challenge: to show how direction selectivity in the dendritic plexus of starburst amacrine cells is being computed. In the rabbit retina, although the cellular locus of direction selectivity is known to occur predominantly in the dendrites of starburst amacrine cells, the biophysical mechanism by which this takes place and its essential topography are yet to be specified with precision. A cotransmission model, involving a conjoint release of excitation/inhibition (i.e., a bisynaptic relay of endogenous ACh and GABA) from the distal varicosities of individual starburst amacrines, will be non-diphasic when the vesicular release of Ach and the non-vesicular, carrier-mediated release of GABA by transporters in the anterograde direction are preferentially suppressed by a negative feedback mechanism involving autoreceptors. Such biophysical mechanisms, including the asymmetric distribution of chloride cotransporters, explain somatofugal motion bias in starburst amacrine cells leading to autonomous functioning "subunits" that underlie the formation of directional selectivity. However, the functional independence of starburst amacrine cell "subunits" is partly a question of their network organization. The topography of directionally selective "subunits" resides in the plexus of crisscrossing dendrites of juxtaposed starburst amacrines, consisting of (i) serial synapses of three or more starburst amacrines and a ON-OFF directionally selective ganglion cell; (ii) a synaptic couplet between two starburst amacrines; and (iii) a conventional synapse between a starburst amacrine and a ON-OFF directionally selective ganglion cell. Cholinergic and GABAergic monosynaptic interactions between starburst amacrine cells, including glutamatergic interactions with cone bipolar cells, are involved in the primary circuit underlying directional selectivity. Furthermore, the secondary circuit underlying directional selectivity, consists of starburst amacrine cells and cone bipolar cells arranged in a "push-pull" configuration, interacting synaptically onto ON-OFF directionally selective ganglion cells.

Synaptic organization of complex ganglion cells in rabbit retina: type and arrangement of inputs to directionally selective and local-edge-detector cells

The type and topographic distribution of synaptic inputs to a directionally selective (DS) rabbit retinal ganglion cell (GC) were examined and were compared with those received by two other complex GC types. The percentage of cone bipolar cell (BC) input, presumably an index of sustained responses and simple receptive field properties, is much higher than expected for complex GCs in reference to previous reports in other species: approximately 20% for the type 1 bistratified ON-OFF DS GC and for a multistratified GC, and approximately 40% for the small-tufted local-edge-detector GC. Consistent with a previous study (Famiglietti [1991] J. Comp. Neurol. 309:40-70), no ultrastructural evidence is found for inhibitory synapses from starburst amacrine cells to the ON-OFF DS GC. The density of inputs to the ON-OFF DS GC is high and rather evenly distributed over the dendritic tree. Clustering of inputs brings excitatory and inhibitory inputs into proximity, but the strict on-path condition of more proximal inhibitory inputs, favoring shunting inhibition, is not satisfied. Prominent BC input and its regional variation suggest that BCs play key roles in DS neural circuitry, both pre- and postsynaptic to the ON-OFF DS GC, according to a bilayer model (Famiglietti [1993] Invest. Ophthalmol. Vis. Sci. 34:S985). Asymmetry of inhibitory amacrine cell input may signify a region on the preferred side of the receptive field, the inhibition-free zone (Barlow and Levick [1965] J. Physiol. (Lond.) 178:477-504), supporting a role for postsynaptic integration in the DS mechanism. Prominent BC input to the local-edge-detector, often without accompanying amacrine cell input, indicates presynaptic integration in forming its trigger feature.

Gap-junction communication between subtypes of direction-selective ganglion cells in the developing retina

The On-Off direction-selective ganglion cells (DSGCs) in the rabbit retina comprise four distinct subtypes that respond preferentially to image motion in four orthogonal directions; each subtype forms a regular territorial array, which is overlapped by the other three arrays. In this study, ganglion cells in the developing retina were injected with Neurobiotin, a gap-junction-permeable tracer, and the DSGCs were identified by their characteristic type 1 bistratified (BiS1) morphology. The complex patterns of tracer coupling shown by the BiS1 ganglion cells changed systematically during the course of postnatal development. BiS1 cells appear to be coupled together around the time of birth, but, over the next 10 days, BiS1 cells decouple from each other, leading to the mature pattern in which only one subtype is coupled. At about postnatal day 5, before the ganglion cells become visually responsive, each of the BiS1 cells commonly showed tracer coupling both to a regular array of neighboring BiS1 cells, presumably destined to be DSGCs of the same subtype, and to a regular array of overlapping BiS1 cells, presumably destined to be DSGCs of a different subtype. The gap-junction intercellular communication between subtypes of DSGCs with different preferred directions may play an important role in the differentiation of their synaptic connectivity, with respect to either the inputs that DSGCs receive from retinal interneurons or the outputs that DSGCs make to geniculate neurons.

Attentional modulation of motion integration of individual neurons in the middle temporal visual area

We examined how spatially directed attention affected the integration of motion in neurons of the middle temporal (MT) area of visual cortex. We recorded from single MT neurons while monkeys performed a motion detection task under two attentional states. Using 0% coherent random dot motion, we estimated the optimal linear transfer function (or kernel) between the global motion and the neuronal response. This linear kernel filtered the random dot motion across direction, speed, and time. Slightly less than one-half of the neurons produced reasonably well defined kernels that also tended to account for both the directional selectivity and responses to coherent motion of different strengths. This subpopulation of cells had faster, more transient, and more robust responses to visual stimuli than neurons with kernels that did not contain well defined regions of integration. For those neurons that had large attentional modulation and produced well defined kernels, we found attention scaled the temporal profile of the transfer function with no appreciable shift in time or change in shape. Thus, for MT neurons described by a linear transfer function, attention produced a multiplicative scaling of the temporal integration window.

Differences between cation-chloride co-transporter functions in the visual cortex of pigmented and albino rats

Albinism in mammals is accompanied by specific morphological and functional alterations of the visual system. To understand their cellular basis we studied the physiological characteristics and transmembrane currents of pyramidal neurons in 350-microm-thick slices of visual cortex from pigmented and albino rats using whole-cell and gramicidin perforated patch-clamp recordings. The resting membrane potential was significantly more positive and the rheobase was significantly lower in neurons of layers II/III and V in albinos as compared with pigmented rats. No significant differences were found in the input resistance, time constant and chronaxy. Whereas the alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor-mediated currents were not significantly different, the maximum gamma-aminobutyric acid (GABA)(A) receptor (GABA(A)R)-mediated currents and miniature inhibitory postsynaptic currents showed significantly lower amplitudes in neurons of layer V in visual cortex of albinos as compared with pigmented rats. The reversal potential of the GABA(A)R-mediated currents (E(GABA)) was significantly shifted to more positive values in albinos. Pharmacological experiments showed that this shift could be caused by an increased action of the inward chloride co-transporter NKCC1 and reduced action of the outward chloride co-transporter KCC2 in albino rats. This difference seems to be restricted to the visual cortex because in pyramidal neurons from frontal cortex E(GABA) was not significantly different in albinos as compared with pigmented rats. These results are discussed in relation to functional alterations in the albino visual system.

Characterization of inhibitory postsynaptic currents in rod bipolar cells of the mouse retina

The synaptic terminals of mammalian rod bipolar cells are the targets of multiple presynaptic inhibitory inputs arriving from glycinergic and GABAergic amacrine cells. To investigate the contribution of these different inhibitory receptor types, we have applied the patch-clamp technique in acutely isolated slices of the adult mouse retina. By using the whole-cell configuration, we measured and analyzed the spontaneous postsynaptic currents (PSCs) in rod bipolar cells. The spontaneous synaptic activity of rod bipolar cells was very low. However, when amacrine cells were depolarized by AMPA or kainate, the PSC frequency in rod bipolar cells increased significantly. These PSCs comprised several types that could be distinguished by pharmacological and kinetic criteria. Strychnine-sensitive, glycinergic PSCs were characterized by a mean peak amplitude of -43.5 pA and a weighted decay time constant (tauw) of 10.9 ms. PSCs that persisted in the presence of strychnine, but were completely inhibited by bicuculline, were mediated by GABAARs. They had a mean peak amplitude of -20.0 pA and a significantly faster tauw of 5.8 ms. Few PSCs remained in the presence of strychnine and bicuculline, suggesting that they were mediated by GABACRs. These PSCs were characterized by much smaller amplitudes (-6.2 pA) and a significantly slower decay kinetics (tauw=51.0 ms). We conclude that rod bipolar cells express at least three types of functionally different inhibitory receptors, namely GABAARs, GABACRs, and GlyRs that may ultimately regulate the Ca2+ influx into rod bipolar cell terminals, thereby modulating their glutamate release.

Identification of ON-OFF direction-selective ganglion cells in the mouse retina

We identified the ON-OFF direction-selective ganglion cells (DSGCs) in the mouse retina and characterized their physiological, morphological and pharmacological properties. These cells showed transient responses to the onset and termination of a stationary flashing spot, and strong directional selectivity to a moving rectangle. Application of various pharmacological reagents demonstrated that the ON-OFF DSGCs in the mouse retina utilize a similar array of transmitters and receptors to compute motion direction to their counterparts in the rabbit retina. Voltage clamp recording showed that ON-OFF DSGCs in the mouse retina receive a larger inhibitory input when the stimulus is moving in the null direction and a larger excitatory input when the stimulus is moving in the preferred direction. Finally, intracellular infusion of neurobiotin revealed a bistratified dendritic field with recursive dendrites forming loop-like structures, previously classified as RG(D2) by morphology. Overall, the ON-OFF DSGCs in the mouse retina exhibit almost identical properties to their counterparts in the rabbit retina, indicating that the mechanisms for computing motion direction are conserved from mouse to rabbit, and probably also to higher mammals. This first detailed characterization of ON-OFF DSGCs in the mouse retina provides fundamental information for further study of maturation and regulation of the neuronal circuitry underlying computation of direction.

Parallel processing in the mammalian retina

Our eyes send different 'images' of the outside world to the brain - an image of contours (line drawing), a colour image (watercolour painting) or an image of moving objects (movie). This is commonly referred to as parallel processing, and starts as early as the first synapse of the retina, the cone pedicle. Here, the molecular composition of the transmitter receptors of the postsynaptic neurons defines which images are transferred to the inner retina. Within the second synaptic layer - the inner plexiform layer - circuits that involve complex inhibitory and excitatory interactions represent filters that select 'what the eye tells the brain'.

Local structural balance and functional interaction of excitatory and inhibitory synapses in hippocampal dendrites

Theoretical and experimental studies on the computation of neural networks suggest that neural computation results from a dynamic interplay of excitatory and inhibitory (E/I) synaptic inputs. Precisely how E/I synapses are organized structurally and functionally to facilitate meaningful interaction remains elusive. Here we show that E/I synapses are regulated across dendritic trees to maintain a constant ratio of inputs in cultured rat hippocampal neurons. This structural arrangement is accompanied by an E/I functional balance maintained by a 'push-pull' feedback regulatory mechanism that is capable of adjusting E/I efficacies in a coordinated fashion. We also found that during activity, inhibitory synapses can determine the impact of adjacent excitatory synapses only if they are colocalized on the same dendritic branch and are activated simultaneously. These fundamental relationships among E/I synapses provide organizational principles relevant to deciphering the structural and functional basis for neural computation within dendritic branches.

Cation--chloride cotransporters mediate neural computation in the retina

The ability of directionally selective (DS) retinal ganglion cells to respond selectively to stimulus motion in one direction is a classic unresolved example of computation in a local neural circuit. Recent evidence indicates that DS responses occur first in the retina in the dendrites of starburst amacrine cells (interneurons presynaptic to the ganglion cells). We report that the directional responses of starburst-cell dendrites and DS ganglion cells are highly sensitive to the polarity of the transmembrane chloride gradient. Reducing the transmembrane chloride gradient by ion substitution or by blocking the K-Cl cotransporter resulted in the starburst cells responding equally to light moving in opposite directions. Conversely, increasing the chloride gradient by blocking the Na-K-Cl cotransporter eliminated responses to light moving in either direction. Moreover, in each case, blocking the chloride cotransporters or reducing the transmembrane chloride gradient eliminated the directional responses of DS ganglion cells in a manner opposite that of the starburst cells. These results indicate that chloride cotransporters play a key role in the generation of direction selectivity and that the directional responses of starburst cells and DS ganglion cells are exquisitely sensitive to the chloride equilibrium potential. The findings further suggest that the directional responses of DS ganglion cells are mediated in part by the directional release of gamma-aminobutyric acid from starburst dendrites and that the asymmetric distribution of the two cotransporters along starburst-cell dendrites mediates direction selectivity. A model of direction selectivity in the retina that incorporates these and other findings is discussed.

Seeing More Clearly: Recent Advances in Understanding Retinal Circuitry

Among 10 breakthroughs that Science announced at the end of 2002 was the discovery of a photosensing (melanopsin-containing) retinal ganglion cell (RGC) and its role in entraining the circadian clock. This breakthrough exemplifies the ultimate goal of neuroscience: to understand the nervous system from molecules to behavior. Light-sensing RGCs constitute one of a dozen discrete RGC populations coding various aspects of visual scenes by virtue of their unique morphology, physiology, and coverage of the retina. Interestingly, the function of the melanopsin-containing RGCs in entraining the circadian clock need not involve much retinal processing, making it the simplest form of processing in the retina. This review focuses on recent advances in our understanding of retinal circuitry, visual processing, and retinal development demonstrated by innovative experimental techniques. It also discusses the advantages of using the retina as a model system to address some of the key questions in neuroscience.

Asymmetric temporal properties in the receptive field of retinal transient amacrine cells

The speed of signal conduction is a factor determining the temporal properties of individual neurons and neuronal networks. We observed very different conduction velocities within the receptive field of fast-type On-Off transient amacrine cells in carp retina cells, which are tightly coupled to each other via gap junctions. The fastest speeds were found in the dorsal area of the receptive fields, on average five times faster than those detected within the ventral area. The asymmetry was similar in the On- and Off-part of the responses, thus being independent of the pathway, pointing to the existence of a functional mechanism within the recorded cells themselves. Nonetheless, the spatial decay of the graded-voltage photoresponse within the receptive field was found to be symmetrical, with the amplitude center of the receptive field being displaced to the faster side from the minimum-latency location. A sample of the orientation of varicosity-laden polyaxons in neurobiotin-injected cells supported the model, revealing that approximately 75% of these processes were directed dorsally from the origin cells. Based on these results, we modeled the velocity asymmetry and the displacement of amplitude center by adding a contribution of an asymmetric polyaxonal inhibition to the network. Due to the asymmetry in the conduction velocity, the time delay of a light response is proposed to depend on the origin of the photostimulus movement, a potentially important mechanism underlying direction selectivity within the inner retina.

Transcriptional control of neuronal diversification in the retina

During embryonic development, the array of vastly different neuronal types that are incorporated into the functional architecture of the mature neuroretina derives from a common population of multipotent retinal progenitor cells (RPCs). Retinogenesis proceeds in a precise chronological order, with the seven principal cell classes generated in successive phases. Cell biological experiments established that this histogenetic order, at least in part, reflects intrinsic changes within the RPC pool. In recent years a number of molecules controlling various aspects of cell fate specification from RPCs have been identified. However, few attempts have been made to integrate previous concepts that emerged from cell biological studies and more recent results based on molecular genetic experiments. This review aims at providing an overview of recent advances in our understanding of the cellular and molecular mechanisms underlying retinal neuronal diversification, with a particular focus on cell-intrinsic factors.

New directions in retinal research

Direction-selective retinal ganglion cells (DSGCs) respond to image motion in a "preferred" direction but not the opposite "null" direction. Extracellular spike recordings from rabbit DSGCs suggested that the key mechanism underlying the directional responses is spatially offset inhibition projecting in the null direction. Recent patch-clamp recordings have shown that this inhibition, which acts directly on the DSGC, is already direction selective. Dual recordings established that the inhibition arises from starburst amacrine cells (SBACs) located on the null side of the DSGC but not from those on the preferred side. Thus, for each radially symmetric SBAC, processes pointing in different directions would provide the null-direction inhibition to subtypes of DSGCs with different preferred directions. Ca2+ imaging revealed that the SBAC terminal processes respond more strongly to image motion away from the soma than towards the soma, therefore accounting for the direction selectivity of the inhibitory input to the DSGCs.

Adaptation of response transients in fly motion vision. I: Experiments

Two types of transient responses have been investigated in fly motion-sensitive neurons in the past: the impulse and the step response. In response to a brief motion pulse, cells show a sudden rise in activity followed by an exponential decay ('impulse response'). In response to the onset of a constant velocity stimulus, cells exhibit transient oscillations before settling to a steady-state value ('step response'). Since the impulse response has been shown to shorten when tested after presentation of an adapting motion stimulus, we investigated whether adaptation also occurs during the step response. We tested this hypothesis by recording extracellularly the response of the H1-cell in the lobula plate of the blowfly Calliphora vicina to gratings of varying pattern contrasts and drift velocity. We found that the transient oscillations of the step response strongly depend on the pattern contrast: at low contrasts, oscillations lasted for several seconds, whereas at high contrasts, they settled within fractions of a second. This suggests that motion adaptation occurs during the initial period of the stimulus presentation and is dependent on the contrast of the motion stimulus. Using identical stimulus parameters (contrast and temporal frequency) for the adapting stimulus and testing the impulse response afterwards, we found that the impulse response and the transient period in the step response shortened in a similar way. We then analyzed the dynamic of the transients oscillations produced by ongoing motion of a square wave pattern in the anti-preferred direction (null direction) of H1. As observed for preferred direction motion, we found that the duration and amplitude of those transients shortened as the contrast and the velocity of the pattern increased, and that the oscillations disappeared when a blank screen instead of a pattern was presented before the onset of motion. Under both stimulus conditions, i.e. grating and blank screen before motion onset, the steady-state response level showed the same dependence on the contrast and temporal frequency of the pattern. When we analyzed the responses of the cell to pattern of various sizes and contrasts moving in the preferred direction of the cell, we found that increments in the size affected the overall amplitude of both the transient oscillations and the steady-state response level, whereas the duration of the oscillations only depended on the local pattern contrast. We also tested the impulse response before and after the presentation of an adapting stimulus presented in either the same or a different location of the visual field. The response shortened only when both the adapting and the test stimuli were presented at the same location. These last experiments demonstrate a strictly local mechanism of adaptation affecting the response transients of both the impulse and the step response.

Space-time maps and two-bar interactions of different classes of direction-selective cells in macaque V-1

We used one-dimensional sparse noise stimuli to generate first-order spatiotemporal maps and second-order two-bar interaction maps for 65 simple and 124 complex direction-selective cells in alert macaque V1. Spatial and temporal phase differences between light and dark space-time maps clearly distinguished simple and complex cell populations. Complex cells usually showed similar direction preferences to light and dark bars, but many of the directional simple cells were much more direction selective to one sign of contrast than the reverse. We show that this is predicted by a simple energy model. Some of the direction-selective simple cells showed multiple space-time-slanted subregions, but others (previously described as S1 cells) had space-time maps that looked like just one subregion of an ordinary simple cell. Both simple and complex cells showed directional interactions (nonlinearities) to pairs of flashed bars (a 2-bar apparent-motion stimulus). The space-time slant of the simple cells correlated with the optimum dX/dT (velocity) of the paired-bar interactions. Some complex cells also showed a space-time slant; the direction of the slant usually correlated with the preferred direction of motion, but the degree of slant correlated with the inferred velocity tuning only when measured by a weighted-centroid calculation. Principal components analysis of the simple-cell space-time maps yielded one fast temporally biphasic component and a slower temporally monophasic component. We saw no consistent pattern for the spatial phase of the components, unlike previous reports; however, we show that principal components analysis may not distinguish between spatial offsets and phase offsets.

Modeling reverse-phi motion-selective neurons in cortex: double synaptic-veto mechanism

Reverse-phi motion is the illusory reversal of perceived direction of movement when the stimulus contrast is reversed in successive frames. Livingstone, Tsao, and Conway (2000) showed that direction-selective cells in striate cortex of the alert macaque monkey showed reversed excitatory and inhibitory regions when two different contrast bars were flashed sequentially during a two-bar interaction analysis. While correlation or motion energy models predict the reverse-phi response, it is unclear how neurons can accomplish this. We carried out detailed biophysical simulations of a direction-selective cell model implementing a synaptic shunting scheme. Our results suggest that a simple synaptic-veto mechanism with strong direction selectivity for normal motion cannot account for the observed reverse-phi motion effect. Given the nature of reverse-phi motion, a direct interaction between the ON and OFF pathway, missing in the original shunting-inhibition model, it is essential to account for the reversal of response. We here propose a double synaptic-veto mechanism in which ON excitatory synapses are gated by both delayed ON inhibition at their null side and delayed OFF inhibition at their preferred side. The converse applies to OFF excitatory synapses. Mapping this scheme onto the dendrites of a direction-selective neuron permits the model to respond best to normal motion in its preferred direction and to reverse-phi motion in its null direction. Two-bar interaction maps showed reversed excitation and inhibition regions when two different contrast bars are presented.

Choline acetyltransferase is expressed by non-starburst amacrine cells in the ground squirrel retina

We have used immunostaining techniques to reveal a new type of amacrine cell that is immunoreactive for choline acetyltransferase (ChAT), the acetylcholine synthesizing enzyme, in the Ground Squirrel (Spermophilus beecheyi) retina. Cryostat sections and double immunostained wholemount preparations were examined by confocal microscopy. This new ChAT type III cell is distinct in morphology and neurotransmitter content from the well know 'starburst' amacrine cells (types I and II) that are so well represented in the ground squirrel retina [J. Comp. Neurol. 365 (1996) 173-216]. The type III cell colocalizes glycine with the acetylcholine and does not appear to be GABAergic or exhibit calcium-binding proteins like the well-known starburst type. As well, type III cells do not occur as a mirror-symmetric pair with normally placed and displaced varieties. The type III cell is probably a small field amacrine type branching broadly in upper sublamina b of the inner plexiform layer, and is most likely A6 of the Ground Squirrel retina [J. Comp. Neurol. 365 (1996) 173-216]. Type III cells are ideally placed in the architecture of the Ground Squirrel retina to influence ON directionally selective ganglion cell types.

Synaptic input to the on-off directionally selective ganglion cell in the rabbit retina

A physiologically identified on-off directionally selective (DS) ganglion cell with its preferred-null axis defined was stained with horseradish peroxidase (HRP) and prepared for electron microscopy. A continuous series of thin sections were used to examine the cell's synaptology. Although the DS cell dendrite received the majority of its synaptic input from a heterogeneous population of amacrine cell processes, a frequently observed synaptic profile consisted of a DS cell dendrite receiving synapses from a cluster of several amacrine cell processes. These clusters of processes were assumed to be from a fascicle of amacrine cells, most of which probably belonged to several different cholinergic starburst amacrine cells. The most frequently observed presynaptic profile within the clusters consisted of a synaptic couplet in which two processes synapsed with each other before one of them finally synapsed with the DS ganglion cell dendrite; occasionally, a chain of three serial synapses was seen. In addition, a specific microcircuit that has the potential to exert lateral feedforward inhibition was also observed. This microcircuit consisted of two cone bipolar cell terminal dyad synapses where one dyad contained an amacrine cell process making a reciprocal synapse and a DS ganglion cell dendrite receiving direct excitation; the other dyad synapse, found lateral to the first dyad, contained two amacrine cell processes that both made reciprocal synapses, but one fed forward to make a putative inhibitory synapse with the DS cell dendrite.

Starburst cells nondirectionally facilitate the responses of direction-selective retinal ganglion cells

The mechanism of direction selectivity in retinal ganglion cells remains controversial. An important issue is how the starburst amacrine cells, which are known to provide a major synaptic input to the direction-selective ganglion cells, participate in the directional discrimination. Here, we present evidence that the cholinergic outputs of the starburst cells affect the responses of the ganglion cells symmetrically; they provide a feedforward excitation that facilitates the response of the ganglion cells to movement in both the preferred and null directions. This seems to place a constraint on models of the directional discrimination in which the starburst cells participate, namely, that their cholinergic synapses be nondirectional in their effects on the ganglion cells.

Fundamental mechanisms of visual motion detection: models, cells and functions

Taking a comparative approach, data from a range of visual species are discussed in the context of ideas about mechanisms of motion detection. The cellular basis of motion detection in the vertebrate retina, sub-cortical structures and visual cortex is reviewed alongside that of the insect optic lobes. Special care is taken to relate concepts from theoretical models to the neural circuitry in biological systems. Motion detection involves spatiotemporal pre-filters, temporal delay filters and non-linear interactions. A number of different types of non-linear mechanism such as facilitation, inhibition and division have been proposed to underlie direction selectivity. The resulting direction-selective mechanisms can be combined to produce speed-tuned motion detectors. Motion detection is a dynamic process with adaptation as a fundamental property. The behavior of adaptive mechanisms in motion detection is discussed, focusing on the informational basis of motion adaptation, its phenomenology in human vision, and its cellular basis. The question of whether motion adaptation serves a function or is simply the result of neural fatigue is critically addressed.

Mechanisms and circuitry underlying directional selectivity in the retina

In the retina, directionally selective ganglion cells respond with robust spiking to movement in their preferred direction, but show minimal response to movement in the opposite, or null, direction. The mechanisms and circuitry underlying this computation have remained controversial. Here we show, by isolating the excitatory and inhibitory inputs to directionally selective cells and measuring direct connections between these cells and presynaptic neurons, that a presynaptic interneuron, the starburst amacrine cell, delivers direct inhibition to directionally selective cells. The processes of starburst cells are connected asymmetrically to directionally selective cells: those pointing in the null direction deliver inhibition; those pointing in the preferred direction do not. Starburst cells project inhibition laterally ahead of a stimulus moving in the null direction. In addition, starburst inhibition is itself directionally selective: it is stronger for movement in the null direction. Excitation in response to null direction movement is reduced by an inhibitory signal acting at a site that is presynaptic to the directionally selective cell. The interplay of these components generates reduced excitation and enhanced inhibition in the null direction, thereby ensuring robust directional selectivity.

Diverse synaptic mechanisms generate direction selectivity in the rabbit retina

The synaptic conductance of the On-Off direction-selective ganglion cells was measured during visual stimulation to determine whether the direction selectivity is a property of the circuitry presynaptic to the ganglion cells or is generated by postsynaptic interaction of excitatory and inhibitory inputs. Three synaptic asymmetries were identified that contribute to the generation of direction-selective responses: (1) a presynaptic mechanism producing stronger excitation in the preferred direction, (2) a presynaptic mechanism producing stronger inhibition in the opposite direction, and (3) postsynaptic interaction of excitation with spatially offset inhibition. Although the on- and off-responses showed the same directional tuning, the off-response was generated by all three mechanisms, whereas the on-response was generated primarily by the two presynaptic mechanisms. The results indicate that, within a single neuron, different strategies are used within distinct dendritic arbors to accomplish the same neural computation.

A spiking neuron model: applications and learning

This paper presents a biologically inspired, hardware-realisable spiking neuron model, which we call the Temporal Noisy-Leaky Integrator (TNLI). The dynamic applications of the model as well as its applications in Computational Neuroscience are demonstrated and a learning algorithm based on postsynaptic delays is proposed. The TNLI incorporates temporal dynamics at the neuron level by modelling both the temporal summation of dendritic postsynaptic currents which have controlled delay and duration and the decay of the somatic potential due to its membrane leak. Moreover, the TNLI models the stochastic neurotransmitter release by real neuron synapses (with probabilistic RAMs at each input) and the firing times including the refractory period and action potential repolarisation. The temporal features of the TNLI make it suitable for use in dynamic time-dependent tasks like its application as a motion and velocity detector system presented in this paper. This is done by modelling the experimental velocity selectivity curve of the motion sensitive H1 neuron of the visual system of the fly. This application of the TNLI indicates its potential applications in artificial vision systems for robots. It is also demonstrated that Hebbian-based learning can be applied in the TNLI for postsynaptic delay training based on coincidence detection, in such a way that an arbitrary temporal pattern can be detected and recognised. The paper also demonstrates that the TNLI can be used to control the firing variability through inhibition; with 80% inhibition to concurrent excitation, firing at high rates is nearly consistent with a Poisson-type firing variability observed in cortical neurons. It is also shown with the TNLI, that the gain of the neuron (slope of its transfer function) can be controlled by the balance between inhibition and excitation, the gain being a decreasing function of the proportion of inhibitory inputs. Finally, in the case of perfect balance between inhibition and excitation, i.e. where the average input current is zero, the neuron can still fire as a result of membrane potential fluctuations. The firing rate is then determined by the average input firing rate. Overall this work illustrates how a hardware-realisable neuron model can capitalise on the unique computational capabilities of biological neurons.

Directionally selective calcium signals in dendrites of starburst amacrine cells

The detection of image motion is fundamental to vision. In many species, unique classes of retinal ganglion cells selectively respond to visual stimuli that move in specific directions. It is not known which retinal cell first performs the neural computations that give rise to directional selectivity in the ganglion cell. A prominent candidate has been an interneuron called the 'starburst amacrine cell'. Using two-photon optical recordings of intracellular calcium concentration, here we find that individual dendritic branches of starburst cells act as independent computation modules. Dendritic calcium signals, but not somatic membrane voltage, are directionally selective for stimuli that move centrifugally from the cell soma. This demonstrates that direction selectivity is computed locally in dendritic branches at a stage before ganglion cells.