FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312672652> ?p ?o ?g. }

• W4312672652 abstract "Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The asymmetric summation of kinetically distinct glutamate inputs across the dendrites of retinal ‘starburst’ amacrine cells is one of the several mechanisms that have been proposed to underlie their direction-selective properties, but experimentally verifying input kinetics has been a challenge. Here, we used two-photon glutamate sensor (iGluSnFR) imaging to directly measure the input kinetics across individual starburst dendrites. We found that signals measured from proximal dendrites were relatively sustained compared to those measured from distal dendrites. These differences were observed across a range of stimulus sizes and appeared to be shaped mainly by excitatory rather than inhibitory network interactions. Temporal deconvolution analysis suggests that the steady-state vesicle release rate was ~3 times larger at proximal sites compared to distal sites. Using a connectomics-inspired computational model, we demonstrate that input kinetics play an important role in shaping direction selectivity at low stimulus velocities. Taken together, these results provide direct support for the ‘space-time wiring’ model for direction selectivity. Editor's evaluation This is an important paper that addresses a key mechanism that underlies the canonical computation of direction selectivity in the retina. By using fluorescence imaging of glutamate release from excitatory interneurons combined with a computational model of dendritic integration, the authors make a convincing case that the kinetics of glutamate release contributes to the direction-selectivity of individual neural processes in retinal neurons. This work will appeal to visual neuroscientists as well as cellular physiologists interested in dendritic computations. https://doi.org/10.7554/eLife.81533.sa0 Decision letter Reviews on Sciety eLife's review process Introduction The radiating dendrites of retinal GABAergic/cholinergic ‘starburst’ amacrine cells (starbursts) are the first points in the visual system to exhibit direction selectivity (Euler et al., 2002). Object motion away from the soma generates large calcium responses in distal starburst dendrites, from where they release GABA and acetylcholine (ACh). By contrast, motion toward the soma evokes weak responses. Starbursts play a critical role in shaping direction-selective (DS) responses of downstream ganglion cells (DSGCs), and thus, understanding how they compute direction is of principal interest. Decades of intense investigations have identified several mechanisms that underlie direction selectivity in starburst dendrites, although no single mechanism alone may be critically required (Ding et al., 2016; Hausselt et al., 2007; Kim et al., 2014; Hanson et al., 2019; reviewed by Murphy-Baum et al., 2021). A model that has garnered recent attention relies on the kinetic properties of distinct sources of glutamatergic input, which is referred to as the ‘space-time wiring’ model for direction selectivity (Greene et al., 2016; Kim et al., 2014). In this study, we sought to evaluate the kinetics of glutamatergic input to the ON starburst dendrites to understand their role in generating direction selectivity. The space-time wiring model for direction selectivity is inspired by connectomic analysis in the mouse retina showing that the proximal and distal dendritic regions of ON and OFF type starburst amacrine cells receive synaptic inputs from anatomically distinct types of glutamatergic bipolar cells (BCs) (Greene et al., 2016; Kim et al., 2014). As the axon terminals of these BCs stratify at distinct depths within the inner plexiform layer (IPL; Ding et al., 2016; Greene et al., 2016; Kim et al., 2014)—and in general BC axonal stratification patterns are linked to their kinetic properties (Awatramani and Slaughter, 2000; Baden et al., 2016; Franke et al., 2017; Gaynes et al., 2022; Strauss et al., 2022)—it has been hypothesized that different types of BCs contacting proximal and distal starburst dendrites have distinct response kinetics (Greene et al., 2016; Kim et al., 2014). Specifically, it is predicted that proximal inputs near the starburst soma are mediated by BC types that support tonic patterns of glutamate release, while distal inputs are mediated by BC types that release their vesicles more transiently. This arrangement would result in an optimal input summation along starburst dendrites during centrifugal (soma-to-dendrite) motion, as experimentally noted. The space-time wiring model for direction selectivity is algorithmically similar to classic correlation-type motion detectors described elsewhere in the visual system in both rodents and primates (Lien and Scanziani, 2018; De Valois and Cottaris, 1998), as well as the fly optic lobe (Haag et al., 2016; Leong et al., 2016; Behnia et al., 2014), and thus appears to reflect a core computational principle. While it is an attractive model for direction selectivity, there is scant evidence that the kinetics of glutamatergic input varies along the length of starburst dendrites. Direct electrophysiological measurements revealed that BC5s (including types 5i, o, and t) and BC7, which make ‘ribbon’ synapses predominantly on the distal and proximal dendrites of ON starbursts, respectively, have similar temporal properties (Ichinose et al., 2014). Regional differences in input kinetics were noted when BC output was measured postsynaptically using voltage-clamp techniques (Fransen and Borghuis, 2017; but see Stincic et al., 2016). However, these electrophysiological recordings do not provide precise information regarding the anatomical location of synaptic inputs. To this end, genetically encoded fluorescent glutamate sensors (iGluSnFRs) have provided an alternate way to measure BC output kinetics (Borghuis et al., 2013; Marvin et al., 2013; Yonehara et al., 2013). However, recent imaging studies have found that most BCs of the same polarity (including the BC5s and BC7 types) have similar temporal properties, at least to spatially restricted stimuli that are relevant to local starburst dendritic computations (Franke et al., 2017; Strauss et al., 2022). Finally, optogenetic studies have demonstrated that direction selectivity remains intact under conditions in which all BC input to starbursts is pharmacologically blocked and the starburst network is directly stimulated in relative isolation (Hanson et al., 2019; Sethuramanujam et al., 2016). Taken together, previous results provide little support for the space-time wiring model for direction selectivity. In the present study, we identify specific stimulus conditions under which stark kinetic differences can be observed in proximal and distal BC inputs. We did so by directly monitoring input kinetics across starburst dendrites using iGluSnFRs selectively expressed in these cells, across a range of stimuli and pharmacological conditions. We used temporal deconvolution to estimate vesicle release dynamics, and in a connectomics-inspired computational model, we tested how the specific spatial distributions of kinetically distinct glutamatergic inputs impact direction selectivity. We found that diverse BC kinetics play a role in shaping direction selectivity mainly in the context of relatively large objects, which move slowly across starburst’s receptive field. Results Temporal diversity of glutamate responses along single starburst dendrites We injected AAVs containing flex-iGluSnFR intravitreally into ChAT-Cre-expressing mice. In a few cases, we observed a strong but sparse expression of iGluSnFR in ON-type starburst cells (Figure 1A). This provided a unique opportunity to visualize glutamate response kinetics across the length of individual dendrites (Figure 1B). Signals from dendritic regions proximal to the starburst soma were captured in a different optical plane than the more distal dendrites, which are ~5 μm apart (Ding et al., 2016; Greene et al., 2016). Spots of light (200 μm in diameter) centered on the imaging field evoked robust iGluSnFR signals throughout the first ~60–80 μm section of starburst dendrites, where glutamatergic BCs are known to make synapses (Ding et al., 2016; Greene et al., 2016). The peak amplitudes of the iGluSnFR signals measured over small regions of interest (ROIs; 5×5 μm2) were relatively stable across the length of single starburst dendrites (Figure 1C). However, the magnitude of the sustained phase significantly decreased as a function of distance (Figure 1C; ΔF/F=0.80±0.29 at proximal sites; ΔF/F=0.29±0.16 at distal sites, measured in 20 dendrites in 4 retinas from 4 mice; *p<0.001, t-test). As a result, the sustained/transient index (STi) computed from the plateau/peak ratio, systematically decreased with distance from the soma (Figure 1D; STi=0.33±0.06 for proximal, 0.16±0.05 for distal ROIs; *p<0.001, t-test) (Note, STi=0 indicates a purely transient response with no plateau phase, and STi=1 indicates a purely sustained response where peak and plateau phases are equal). Taken together, these results provide the first piece of direct evidence that the kinetics of glutamatergic input varies along starburst dendrites, supporting the ‘space-time’ wiring model for direction selectivity (Kim et al., 2014; Greene et al., 2016). Figure 1 Download asset Open asset Temporal diversity of inputs across single starburst dendrites revealed by sparse iGluSnFR imaging. (A) Two-photon z-stack image (left) of a single ON starburst amacrine cell expressing iGluSnFR. Changes in iGluSnFR fluorescence evoked by a 200-μm spot were measured across the single starburst dendrite (yellow box; left). Proximal and distal responses were captured in separate focal planes, and the resulting images were stitched together (right; the vertical white line separates the two planes). (B) Examples of time-varying iGluSnFR signals (ΔF/F) (average; two trials) measured in small dendritic regions of interest (5×5 μm2 ROIs; shown in (A)). The responses and ROIs are color-coded according to their sustained/transient indices (STis; color scale bar shown in (A)). The STis (mean; two trials) are indicated above each trace. (C) The amplitudes of the peak and plateau iGluSnFR responses are plotted as a function of radial distance from the soma. Each point indicates the value obtained from an individual ROI averaged over two trials; ROIs on the same dendrite share the same color (n=66 ROIs from 20 dendrites/4 retinas/4 mice). (D) STis computed from (C) plotted as a function of radial distance from the soma. ROI, region of interest. Figure 1—source data 1 Temporal diversity of inputs across single starburst dendrites revealed by sparse iGluSnFR imaging. https://cdn.elifesciences.org/articles/81533/elife-81533-fig1-data1-v2.xlsx Download elife-81533-fig1-data1-v2.xlsx In most experiments, iGluSnFR expression was more widespread across the starburst population. Individual dendrites leaving the starburst soma were easily visible, but as they dove deeper into the IPL, they merged with dendrites from other starbursts to form the intricate ‘honeycomb’ mesh that is characteristic of these cholinergic cells. By taking care to lay the retina down flat in the recording chamber, we were able to measure responses from proximal and distal starburst dendrites in separate imaging planes (Figure 2A). We found that STis were significantly lower in imaging planes that captured distal dendrites (STi=0.21±0.07; μ±s.d.); compared to those that captured proximal dendritic responses (STi=0.34±0.07; μ±s.d.) (n=242 proximal ROIs; n=563 distal ROIs; 8 retinas from 8 mice; 10 FOVs; *p<0.001, t-test; Figure 2A–C), verifying our initial findings on a larger population level. The STi for individual ROIs measured across the population was independent of response amplitude, indicating that the estimated differences in signal kinetics are not strongly compromised by signal-to-noise and/or sensor saturation issues (Figure 3—figure supplement 1). In addition, we also found that the latencies and rise times for responses measured at proximal and distal sites were similar (Figure 2D and E), indicating that the small differences in axonal path lengths between proximal and distal BCs do not result in significant transmission delays, as previously envisioned (Kim et al., 2014). Figure 2 Download asset Open asset Measuring inputs kinetics in the starburst population. (A) In the left scan field, proximal dendrites arising from the starburst soma expressing iGluSnFR can be visualized in relative isolation. Images that were taken ~5 μm deeper in the retina (right) reveal the dense ‘honeycomb’ structure formed by distal starburst dendrites. (B) Example iGluSnFR responses evoked by 200 μm static spot extracted for a few ROIs numbered in (A) with their STis indicated on the top. Yellow bands indicate stimulus duration. Black, mean responses; gray, ± s.e.m. of two trials. (C–E) Distribution of STis (C), 80–20% rise times (D) and latencies (E) in the proximal and distal field of views (FOVs) of the individual (gray) and average (black) ROIs from different recordings (n=10 FOVs, 8 retinas, *p<0.001; t-test). ROI, region of interest; STi, sustained/transient index. Figure 2—source data 1 Measuring inputs kinetics in the starburst population. https://cdn.elifesciences.org/articles/81533/elife-81533-fig2-data1-v2.xlsx Download elife-81533-fig2-data1-v2.xlsx Anatomical studies show that inputs to proximal starburst dendrites originate mainly from BC7s, while inputs to distal dendrites arise from BC5s (including types BC5i, o, and t; Greene et al., 2016; Ding et al., 2016), indicating that the kinetic differences of iGluSnFR responses may reflect the properties of distinct BC types. By using an AAV-8BP/2 vector containing a CAG promotor, we directly monitored glutamate release at BC7s axon terminals (Matsumoto et al., 2021; Figure 3). We identified axon terminals of BC7s based on the depth of ON starburst cell dendrites that were genetically labeled by tdTomato (Matsumoto et al., 2021). We found that iGluSnFR responses at BC7 terminals were sustained, regardless of their peak amplitude (Figure 3—figure supplement 1). The most appreciable changes in iGluSnFR fluorescence occurred at the axon terminals, suggesting that the sensor signals reflect the vesicle release dynamics of individual BC7s (James et al., 2019). While these results lend support to the idea that BC7s are the source of sustained proximal input, we were unable to express the sensor in BC5s and could not directly confirm that these were the sources of transient signals observed in distal starburst dendrites. Figure 3 with 2 supplements see all Download asset Open asset Expressing iGluSnFR in BC7 axon terminals reveals their sustained output. (A) Cross-section of an image stack showing iGluSnFR labelled BC7 (left). The intensity profiles of the BC terminals labeled with iGluSnFR (gray), and starburst dendrites labeled with tdTomato (black) across the inner plexiform layer (IPL) show that BC terminals co-stratify with dendrites of ON starbursts (right). (B) iGluSnFR expression in BC7 terminals (top) imaged at the same depth as the proximal ON starburst dendrites labeled with tdTomato (bottom). (C) Light-evoked glutamate signals (right) extracted from three ROIs shown in (B) (left). The gray band indicates the stimulus duration. (D) Heat maps of the STis for all identified ROIs. (E) A histogram of STis for the light-evoked responses for all ROIs. Top, mean (circle), and s.d. (horizontal bar) of the STis. BC, bipolar cell; ROI, region of interest; STi, sustained/transient index. When white-noise stimuli were used to characterize the temporal properties of BCs using reverse-correlation techniques, we failed to observe significant kinetic differences in proximal and distal iGluSnFR responses. We found the input impulse responses were biphasic for both proximal and distal inputs (Figure 3—figure supplement 2). Thus, the probability of glutamate release from BC terminals appears to be transiently depressed, following a burst of vesicle release, during continuous stimulus regimes. Similar biphasic kernels have been observed in recent imaging studies (Franke et al., 2017; Strauss et al., 2022). As the biphasic nature of the distal—but not proximal—input is critical to the success of models generating direction selectivity (Kim et al., 2014; Fransen and Borghuis, 2017), we conclude that under conditions where the circuit is continually stimulated at high frequencies, input kinetics are unlikely to play a role in shaping direction selectivity in starburst dendrites. BC output kinetic differences are shaped largely by excitatory network mechanisms Next, to investigate whether cell-intrinsic or network mechanisms shape BC kinetics, we examined how iGluSnFR responses were affected by stimulus size. Increasing the spot diameter systematically decreased the peak amplitude of BC responses, indicative of the recruitment of the inhibitory surround (Figure 4A; Franke et al., 2017). Importantly, the distinction in the kinetics of proximal and distal inputs remained clear across stimulus sizes, although they were generally more pronounced for stimuli that were >200 μm (Figure 4B–E, control; n=71 proximal and 431 distal ROIs, 5 FOVs, 4 retinas from 4 mice; *p<0.001, Kolmogorov-Smirnov test). Indeed, the application of a cocktail of antagonists containing both GABA and ionotropic glutamate receptor antagonists (5 µM gabazine and 100 µM TPMPA, 20 µM CNQX, respectively)—which blocks inhibitory inputs from amacrine and horizontal cells—augmented responses, especially those evoked by larger spots (Figure 4A). These effects reduced the overall STi as compared to control, but the kinetics of the iGluSnFR responses at proximal and distal ROIs remained distinct (Figure 4B–E, drug cocktail; n=71 proximal and 431 distal ROIs, 5 FOVs, 4 retinas from 4 mice; *p<0.001, Kolmogorov-Smirnov test). Thus, while the inhibitory networks modulate BC responses, they do not appear to account for the sustained/transient differences observed here. Figure 4 Download asset Open asset Kinetic differences in iGluSnFR signals are apparent across a range of stimulus sizes and persist in the presence of inhibitory receptor blockers. (A) The average iGluSnFR signals were evoked by spots of different diameters (100–800 μm). Responses were averaged across five proximal (orange) or distal (blue) FOVs. Responses measured under control (dark traces) conditions and in the presence of synaptic blockers (SR, TPMPA, and CNQX; light traces) are overlaid. Shading indicates ± s.e.m. (B–E) Cumulative distributions of STis for ROIs in the proximal and distal FOVs under control and blocker conditions for different stimulus sizes. (n=5 FOVs, 4 retinas, *p<0.001; Kolmogorov-Smirnov test). FOV, field of view; ROI, region of interest; STi, sustained/transient index. Figure 4—source data 1 Kinetic differences in iGluSnFR signals are apparent across a range of stimulus sizes and persist in the presence of inhibitory receptor blockers. https://cdn.elifesciences.org/articles/81533/elife-81533-fig4-data1-v2.xlsx Download elife-81533-fig4-data1-v2.xlsx Blocking glutamate/GABA receptor-mediated pathways using the drug cocktail also revealed a somewhat unexpected spread of lateral excitation. Under inhibitory receptor blockade, the light-evoked iGluSnFR responses continued to grow even when the spots sizes were increased significantly beyond the size of BC dendritic fields (~50 µm; Figure 4A, n=71 proximal and 431 distal ROIs, 5 FOVs, 4 retinas from 4 mice; *p<0.01). The maximal peak response was evoked for spots that were ~200 µm diameter. Interestingly, the amplitude of the plateau phase was further increased by ~23% in the proximal and ~20% in the distal dendritic sites, between 200 µm and 800 µm (*p<0.01, t-test). Such lateral excitation is likely to be attributed to electrical coupling, which occurs between BCs as well as amacrine cells (Sigulinsky et al., 2020; Arai et al., 2010; Asari and Meister, 2014). The finding that the kinetic diversity of BC responses is maintained in the absence of inhibition, suggests that they are shaped in large part by excitatory network mechanisms. BC output kinetics contribute to direction selectivity Next, we tested how the apparent kinetic diversity in BC input impacts direction selectivity. Since the BC kinetics inferred from iGluSnFR measurements are in part dictated by the properties of the indicator, it remains unclear how they relate to the starbursts’ physiological responses mediated by endogenous AMPA receptors. To address this issue, we first used optical deconvolution methods (Awatramani et al., 2007) to estimate the time-varying vesicle release rates from individual BCs and then used a computational model to understand how trains of vesicles are transformed into AMPA receptor-driven voltage signals in starburst dendrites. Time-varying vesicle release rates from proximal and distal BCs were estimated by deconvolving the iGluSnFR response with an idealized quantal response described by an alpha function with decay kinetics ~30 ms (Figure 5B, inset), which was obtained by matching the kinetics of ‘spontaneous’ quantal responses (Figure 5A). The resulting release rates were discretized using a Poisson process, yielding an estimate of the temporal pattern of single vesicle release events (Figure 5C). For this analysis, the quantal size for each ROI was estimated using fluctuation analysis (see Materials and methods). Indeed, convolving the average quantal response with the inferred trains of vesicle release events (Figure 5C) produced a signal similar to the original iGluSnFR measurement (Figure 5D). The release profiles of proximal and distal BCs obtained in this manner support the notion that they have different capacities to release vesicles in response to light stimuli: proximal BCs release vesicles throughout the duration of the stimulus, while distal BCs release a large fraction of their vesicles at the onset of the light stimulus (Figure 5E). Figure 5 Download asset Open asset Time-varying vesicle release rates estimated using temporal deconvolution. (A) Spontaneous iGluSnFR signals measured in neighboring ROIs across a small dendritic section (color-coded to match ROIs). (B) A typical light-evoked iGluSnFR signal measured from proximal dendrites; inset: iGluSnFR quantal event fitted with an alpha function. (C) A time-varying release rate was estimated by deconvolving the iGluSnFR signal with the quantal signal (shown in (B)). (D) Convolving the estimated release rate with a unitary event recapitulates the shape of the original iGluSnFR response. (E) Example vesicle release rates for sustained (orange) and transient (blue) iGluSnFR responses. The solid line indicates the average vesicle release rates for all ROIs. n=50 each, proximal and distal ROIs. ROI, region of interest. Previous experimental and modeling studies have shown starburst dendritic sectors to be relatively electrically isolated from each other (Poleg-Polsky et al., 2018; Morrie and Feller, 2018; Ozaita et al., 2004). Since the potential for intra-dendritic signaling is likely to be minimal, we constructed a simple ball-and-stick model (NEURON), rather than a more complex network model. Using previously described properties of starburst dendrites (Tukker et al., 2004; Vlasits et al., 2016), we directly tested how vesicle release dynamics impacts starburst dendritic computations, which is hard to achieve experimentally. Sustained and transient BCs were assumed to reflect the properties of BC7 and BC5s, respectively, and their positions on the starburst dendrites were picked pseudo-randomly from the distributions described previously by Ding et al., 2016; Figure 6—source data 1. The results shown are obtained from averaging many simulations, each with resampled BC complements (Figure 6A; also see Materials and methods). A simulated moving bar (400 μm wide) ‘activated’ BCs in succession, initiating streams of postsynaptic AMPA receptor-mediated miniature-like events (Ιdecay~0.54 ms; Vlasits et al., 2016) according to their location in the dendrite. In these simulations, we assumed the AMPA receptor-mediated events summed linearly (i.e., AMPA receptors did not saturate or desensitize during trains of activity). Synapses were sequentially turned off after the bar traversed BC receptive fields (60 μm diameter). Thus, the input duration at each point in the dendrite varied linearly with stimulus velocity. Figure 6 with 1 supplement see all Download asset Open asset Input kinetics shape direction selectivity at low stimulus velocities. (A) Schematic representation of locations of somatic voltage and terminal Ca2+ recordings from a model SAC under simulated conditions (top). Bipolar cell conductances, somatic voltage, and terminal Ca2+ responses (bottom) were measured in the preferred and null direction from the model SAC when simulated using moving bars. (B) Direction selectivity index (DSi) of peak Ca2+ (terminal) responses versus velocity for different BC input distributions—(i) sustained and transient inputs; when proximal (sustained) and distal (transient) inputs are distributed based on connectomics data (original model); (ii) transient-sustained inputs; when sustained and transient BC inputs are reversed at their locations; (iii) all transient inputs; when all proximal inputs are replaced by transient BCs; and (iv) all sustained inputs; when all distal inputs are replaced by sustained BCs. Shading indicates ± s.e.m. (C) DSi of peak Ca2+ (terminal) responses versus number of inputs from sustained BCs at a velocity of 0.15 mm/s. BC, bipolar cell. Figure 6—source data 1 Model parameters. https://cdn.elifesciences.org/articles/81533/elife-81533-fig6-data1-v2.xlsx Download elife-81533-fig6-data1-v2.xlsx The voltage responses measured from the model cell soma were qualitatively similar to those measured experimentally. For example, responses measured in the soma rose rapidly in the preferred direction compared to those evoked in the null direction, owing to the asymmetric distribution of inputs (Figure 6A, middle panel; Ankri et al., 2020). The model also accurately recapitulated direction selectivity measured in the distal intracellular Ca2+ signals (Figure 6A, bottom traces) similar to results from two-photon Ca2+ imaging experiments (Euler et al., 2002). As expected for a mechanism that relies on a fixed delay, this direction selectivity was strongly dependent on stimulus velocity. In our model based on the experimentally determined input kinetics, anisotropic summation occurs most robustly for slow-moving stimuli (<0.5 mm/s). Four lines of evidence highlight the importance of BC kinetics in enhancing direction selectivity. First, when the kinetics of proximal and distal BCs were exchanged (i.e., proximal BCs were made transient and the distal BCs were made sustained), the directional preference was reversed from preferring centrifugal to centripetal motion (resulting in a negative DSi; Figure 6B). Notably, the strength of the directional preference was stronger for the native compared to the switched BC distributions, indicating that the kinetic mechanisms work in concert with intrinsic dendritic mechanisms (Hausselt et al., 2007; Tukker et al., 2004). Second, DSi was strongly decreased when all BCs were made either sustained or transient (Figure 6B). In all cases, the direction selectivity was most robust for slow-moving stimuli, and we found that these BC distribution manipulations had a statistically significant effect on DSi up to 1 mm/s (p<0.05; one-way analysis of variance, ANOVA; Figure 6B). DSi was significantly higher in the condition when BC distribution followed the native sustained-transient distribution over when all BCs were sustained (p<0.0005; t-test), for velocities up to 0.5 mm/s. Third, when we fixed the BC input locations and incrementally converted sustained input into transient ones (starting from the proximal site furthest from the soma), direction selectivity decreased linearly (Figure 6C; velocity 0.15 mm/s). This is presumably because the amount of asymmetric temporal summation decreases as inputs with weaker plateau phases are swapped in. Finally, as the BC populations for each repeat were randomly distributed, we asked how much the average distance between the proximal and distal populations on a particular trial impacts the resulting DSi (Figure 6—figure supplement 1). We found that for velocities with the strongest DS response (positive for the control condition, and negative for the swapped release profile condition), there were indeed linear relationships between the magnitude of DSi and the mean distance between the sustained and transient BC input (R2≈0.208 and 0.269, respectively, Figure 6—figure supplement 1). Thus, we conclude that in the context of realistic estimates of glutamate input kinetics, the ‘space-time’ wiring model for direction selectivity holds merit under specific stimulus conditions. Discussion Recent models of direction selectivity in the primary sensing neurons in the fly visual system and mammalian retina incorporate key aspects of two classic models of direction selectivity reviewed by Borst et al., 2020, Murphy-Baum et al., 2021: one relying on multiplicative interactions between excitatory inputs (Hassenstein and Reichardt, 1956) and the other relying on divisive inhibition (Barlow and Levick, 1965). While the neural substrates underlying inhibition appear" @default.
• W4312672652 created "2023-01-05" @default.
• W4312672652 creator A5018834013 @default.
• W4312672652 date "2022-08-08" @default.
• W4312672652 modified "2023-09-27" @default.
• W4312672652 title "Editor's evaluation: Spatiotemporal properties of glutamate input support direction selectivity in the dendrites of retinal starburst amacrine cells" @default.
• W4312672652 doi "https://doi.org/10.7554/elife.81533.sa0" @default.
• W4312672652 hasPublicationYear "2022" @default.
• W4312672652 type Work @default.
• W4312672652 citedByCount "0" @default.
• W4312672652 crossrefType "peer-review" @default.
• W4312672652 hasAuthorship W4312672652A5018834013 @default.
• W4312672652 hasBestOaLocation W43126726521 @default.
• W4312672652 hasConcept C12554922 @default.
• W4312672652 hasConcept C169760540 @default.
• W4312672652 hasConcept C170493617 @default.
• W4312672652 hasConcept C185592680 @default.
• W4312672652 hasConcept C2778658801 @default.
• W4312672652 hasConcept C2780827179 @default.
• W4312672652 hasConcept C41008148 @default.
• W4312672652 hasConcept C55493867 @default.
• W4312672652 hasConcept C61174792 @default.
• W4312672652 hasConcept C86803240 @default.
• W4312672652 hasConceptScore W4312672652C12554922 @default.
• W4312672652 hasConceptScore W4312672652C169760540 @default.
• W4312672652 hasConceptScore W4312672652C170493617 @default.
• W4312672652 hasConceptScore W4312672652C185592680 @default.
• W4312672652 hasConceptScore W4312672652C2778658801 @default.
• W4312672652 hasConceptScore W4312672652C2780827179 @default.
• W4312672652 hasConceptScore W4312672652C41008148 @default.
• W4312672652 hasConceptScore W4312672652C55493867 @default.
• W4312672652 hasConceptScore W4312672652C61174792 @default.
• W4312672652 hasConceptScore W4312672652C86803240 @default.
• W4312672652 hasLocation W43126726521 @default.
• W4312672652 hasOpenAccess W4312672652 @default.
• W4312672652 hasPrimaryLocation W43126726521 @default.
• W4312672652 hasRelatedWork W2038334426 @default.
• W4312672652 hasRelatedWork W2055206412 @default.
• W4312672652 hasRelatedWork W2086688350 @default.
• W4312672652 hasRelatedWork W2402345182 @default.
• W4312672652 hasRelatedWork W2507213358 @default.
• W4312672652 hasRelatedWork W2748952813 @default.
• W4312672652 hasRelatedWork W2899084033 @default.
• W4312672652 hasRelatedWork W2914785353 @default.
• W4312672652 hasRelatedWork W3163823935 @default.
• W4312672652 hasRelatedWork W4290366609 @default.
• W4312672652 isParatext "false" @default.
• W4312672652 isRetracted "false" @default.
• W4312672652 workType "peer-review" @default.