New Results

Diurnal modulation of multivesicular release controls the efficiency of information transmission at a sensory synapse

José Moya-Díaz, Ben James, Federico Esposti, Jamie Johnston, Leon Lagnado
doi: https://doi.org/10.1101/2021.09.12.459944

Summary

Sensory circuits adapt to changes in the external world or the animal’s internal state through the actions of neuromodulators1,2. Synapses are key sites at which neuromodulators act3,4 but it is not known how they alter the amount of information transmitted5. We investigated this question in the context of the diurnal regulation of visual processing in larval zebrafish, focusing on ribbon-type synapses of retinal bipolar cells6. We demonstrate that contrast-sensitivity peaks in the afternoon accompanied by an average four-fold increase in the Shannon information transmitted at individual active zones. This increase reflects higher synaptic gain, lower spontaneous “noise”, reduced variability of stimulus-evoked release and improved temporal precision. Simultaneously, an increase in the probability of multivesicular events with larger information content increases the efficiency of information transmission (bits per vesicle) by factors of 2-3. The neuromodulator dopamine contributes to all these changes in synaptic function, although ON and OFF visual channels are differentially affected. Multivesicular release is a property of synapses in many parts of the brain7-11 and this study demonstrates that it’s potentiation by neuromodulators can both increase the amount of information that is transmitted and the efficiency of transmission, revealing a previously unknown mechanism for adjusting neural processing.

It has long been understood that the flow of signals through neural circuits is adjusted by neuromodulators1. Less clear is how these alter the amount of information that is transmitted through the circuit3. Here we investigate this question in the context of visual processing in the retina.

In diurnal animals, the sensitivity of the retina to light is regulated both by the daily light-dark cycle and by intrinsic circadian clocks12-15. But the average luminance of a visual scene is not the variable driving most behaviours related to vision: navigation, finding food and avoiding predators all depend on detection of fast modulations in light intensity. We therefore investigated the diurnal control of temporal contrast processing, focusing on the visual signal transmitted by glutamatergic synapses of bipolar cells. These neurons are the bridge between the photoreceptors and ganglion cells that deliver the results of retinal processing to downstream circuits6. The synaptic compartments of bipolar cells contribute to a number of processing tasks, from adaptive gain control to temporal filtering and the coding of motion, colour, orientation and direction6,16-20. Bipolar cells, like sensory hair cells and electroreceptors, contain ribbon structures that supply vesicles to the active zone21. These sensory synapses do not always operate as Poisson machines in which vesicles are released independently but also signal through multivesicular release (MVR), where the fusion of two or more vesicles is co-ordinated as a single synaptic event8,22,23. MVR is also a common property of neocortical9,11, striatal10 and cerebellar24 synapses.

Diurnal modulation of contrast-sensitivity and synaptic gain

Using SyGCaMP225 we found that synaptic calcium responses to modulations in light intensity varied during the day and peaked in the subjective afternoon (Fig. 1A-D). At ZT = 4 hours, temporal contrasts below 50% were barely detected and the half-maximal response (C1/2) was generated by a stimulus contrast of 86 ± 2 % (red traces in Figs. 1B-C). But at ZT = 7 hours C1/2 it fell to 35 ± 2 % with responses saturated above 50%. When C1/2 was mapped during the course of the day it was relatively constant at ZT 1-5 hours and ZT 9-14 hours but increased to levels ∼2.4-fold lower around ZT = 7 hours (Fig. 1D). Notably, this peak in the contrast sensitivity of the retinal circuit occurred at a similar Zeitgeber time as the maximum contrast sensitivity measured behaviourally using the optokinetic reflex13,15. Luminance sensitivity was also under diurnal control but with a distinctive time-course, gradually increasing ∼200-fold through the day (Extended Data Figure 1).

Figure 1:
Figure 1: Diurnal modulation of contrast-sensitivity and synaptic gain

A. Left: Retina of a Ribeye::SyGCaMP2 fish with box over the IPL. Right: expansion of the boxed region of showing terminals of bipolar cells. B. Averaged responses to stimuli of different contrasts measured at Zeitgeber time 4, 7 and 13 hrs. C. Peak response amplitude as a function of contrast for terminals shown in B. The smooth lines are Hill functions used to interpolate values of C1/2, the contrast generating the half-maximal response. Note the diurnal variations. At ZT = 4 hrs: C1/2 = 86 ± 2% (dashed red arrow); h = 7.0 ± 1.2. At ZT = 7 hrs: C1/2 = 35 ± 2% (dashed black arrow); h = 2.7 ± 0.2. At ZT = 13 hrs: C1/2 = 72 ± 2%; h = 3.3 ± 0.2. D. Variations in contrast-sensitivity as a function of Zeitgeber time averaged across ON and OFF terminals. Note the peak around ZT = 7 hours which is not mirrored in the diurnal variation in luminance sensitivity (Extended Data Figure 1). The grey bars show the periods described as “morning” and “afternoon”. All error bars show ± 1 SD. E. Multiphoton section through the eye of a zebrafish larva (7 dpf) expressing iGluSnFR in a subset of bipolar cells. The red dashed line shows the typical position of a line-scan through a synaptic terminal. F. Examples of iGluSnFR signals from an individual OFF synapse elicited using a stimulus of variable contrast modulated at 5 Hz (0-100%, full field, sine wave) in the morning (ZT 1-3 hours, grey) and afternoon (ZT 6-9 hours, black). Note the high levels of spontaneous activity in the morning (black arrowheads). In each case the top trace shows the iGluSnFR signal and the lower trace the estimated number of quanta composing each event (Qe). G. Average contrast-response functions in OFF bipolar cell synapses in the morning (open circles; n = 20 synapses) and afternoon (closed; n = 59), where the response (R) was quantified as the average of quanta per cycle (Qc). The smooth lines are fits of a sigmoid used for smoothing. Note the differences in the shape of the contrast-response functions and in the levels of spontaneous activity (zero contrast). H. Average contrast-response functions in ON bipolar cell synapses in the morning (open circles; n = 12 synapses) and afternoon (closed; n = 31). There was no significant difference in in the morning relative to afternoon (Chi-square test, p = 0.9999).I. The contrast gain calculated as the derivative of the fits to the contrast-response functions in G and H. Note that the maximum contrast discrimination is increased by a factor of 2x in the OFF channel during the afternoon.

To measure transmission of the visual signal in terms of its elementary units – synaptic vesicles - we expressed the reporter iGluSnFR sparsely in bipolar cells23,26 (Fig. 1E; see Methods and Extended Data Figure 2). Synaptic function was compared over a two-hour period beginning 1 hour after light onset (“morning”) with a two-hour period beginning 6 hours later (“afternoon”; Fig. 1F). We began by measuring the contrast-response function (CRF) simply as the average number of vesicles released per cycle of a 5 Hz stimulus. There was notably little diurnal modulation of the CRF measured at ON synapses (Fig. 1H) but in the OFF channel the maximum rate of release measured at 100% contrast increased from 3.05 ± 0.5 vesicles/s in the morning to 5.1 ± 0.3 vesicles/s in the afternoon (Fig. 1G). This increase in synaptic gain was also accompanied by an increase in contrast sensitivity (1/C1/2) and the combined effects were assessed as the derivative of the CRF (“contrast gain”; Fig. 1I). Contrasts in natural visual scenes rarely exceed 40%17 and in the morning this range was signalled best through the ON channel. But in the afternoon the OFF channel became dominant, with contrast gains increasing by factors of 2-6.

Dopamine regulates contrast gain

A key neuromodulator in the retina is dopamine released from amacrine cellsunder control of internal circadian clocks12,27 as well as external signals, such as changes in luminance14 or the appearance of food-related odours28. Dopamine levels are a minimum at night and increase throughout the day to modulate luminance-sensitivity14,29 (Extended Data Figure 3). To test whether dopamine also regulates contrast sensitivity we injected agonists or antagonists of D1 receptors into the eye. Fig. 2A shows examples of the output from a synapse imaged in the afternoon, before and after injection of the D1 antagonist SCH 23390 (∼0.1 μM). Counteracting the actions of endogenous dopamine reduced the average rate of vesicle release and shifted the CRF such that the maximum contrast gain was achieved at higher contrasts (Fig. 2B and C). Conversely, increasing activation of D1 receptors in the morning by injection of the agonist ADTN (∼0.2 μM) increased response gain.

Figure 2.
Figure 2. Diurnal changes in dopamine levels modulate of contrast-sensitivity and synaptic gain

A.Examples of iGluSnFR signals recorded in the afternoon from an individual OFF (red trace) and ON (green trace) synapses elicited using a stimulus of variable contrast before and after intravitreal injection of the D1 antagonist, SCH 23390 (black traces; 5 Hz modulation). Note that SCH 23390 abolished synaptic responses at lower contrasts in ON and OFF synapses. In each case the top trace shows the iGluSnFR signal and the lower trace the estimated Qe. B. Average contrast-response functions in OFF bipolar cell synapses after administration of D1 antagonist (black dots) in the afternoon and after administration of the D1 agonist ADTN in the morning (blue dots). Each point shows the mean ± s.e.m. (SCH 23390, n =12 synapses; ADTN, n = 12 synapses). Control responses observed in the morning and afternoon are superimposed in the graph (red lines, see Fig. 2C) C. Average contrast-response functions in ON bipolar cell synapses in three conditions: afternoon (green dots), after intravitreal injection of D1 antagonist in the afternoon (black dots) and ADTN in the morning (blue dots). Each point shows the mean ± s.e.m. (SCH 23390, n =7 synapses; ADTN, n = 5 synapses). Control responses observed in the morning and afternoon are superimposed to the graph (green lines, see Fig. 2D). D. Relative response gain by diurnal modulation and after manipulation of dopaminergic signalling (dashed lines). Note that diurnal modulation of synaptic gain is higher in OFF synapses, whereas dopamine modulates the dynamic range by ∼16 fold-change in ON and OFF synapses.

The dynamic range over which D1 receptors adjusted synaptic gain was calculated as the ratio of the CRFs in the presence of the agonist and antagonist: in both ON and OFF channels the maximum modulation was ∼16-fold, occurring at contrasts of 20-40% (Fig. 2D). Diurnal modulation of gain was narrower than this potential range: 1.7-fold in OFF synapses and 1.1-fold in ON.

Diurnal modulation of synaptic noise and variability

In the framework of information theory30, an increase in synaptic gain will tend to reduce uncertainty (and therefore increase information) by causing a larger change in the number of vesicles released when contrast changes. But information is destroyed by “noise” and synapses are a major source of such variability within neural circuits5,31,32. We therefore investigated the diurnal modulation of four aspects of synaptic noise and variability.

i) Spontaneous release

Increases in synaptic gain were accompanied by a decrease in the spontaneous release of vesicles in the absence of a visual stimulus. In the morning, spontaneous events occurred at relatively high rates (Fig. 3A) composed of both single vesicles and MVR (Fig. 3B and C). Integrating across events of all amplitudes, the average rate of spontaneous release in OFF synapses was 4.5 ± 2.5 vesicles s-1 in the morning falling to 1 ± 0.2 vesicles s-1 in the afternoon (Fig 3B). In ON synapses these values were 1.8 ± 0.8 vesicles s-1 and 0.5 ± 0.2 vesicles s-1 (Fig. 3C). In both channels, therefore, spontaneous noise was ∼4 times lower in the afternoon compared to the morning. Increased activation of D1 receptors suppressed noise in the morning to levels close to those measured in the afternoon, but only in OFF synapses (Fig. 3B and C).

Figure 3.
Figure 3. Diurnal modulation of synpatic noise and variability

A.Top: Example of iGluSnFR signals from an individual OFF synapse elicited using a stimulus of variable contrast in the morning (0-100%, 5 Hz modulation). In this example, note the high levels of spontaneous activity that were quantified as the responses elicited at zero contrast (red dashed box). Bottom. Examples of iGluSnFR signals from the same OFF synapse after intravitreal injection of ADTN. Note the increase in amplitude and frequency of events and the reduction of spontaneous activity. In each case the top trace shows the iGluSnFR signal and the lower trace the estimated Qe. B.Quantification of spontaneous events composed by different Qe in OFF synapses in the morning (n = 20 synapses), morning + ADTN (n = 12) and afternoon (n = 24 synapses). Note the suppression of spontaneous events in OFF synapses after intravitreal injection of ADTN in the morning. C. As B, but in ON synapses (n = 5-17). Spontaneous activity was not significantly changed by ADTN. D. Left: Variability in the evoked response of OFF synapses calculated as the Fano factor, where each response was quantified as the total number of vesicles released over one cycle at the contrasts shown. Comparison is made between the morning (n = 18), afternoon (n = 27) and the morning after injection of ADTN (n = 13). Right: Average Fano factor for each condition was significantly higher in the morning compared to ether the afternoon or in the morning after injection of ADTN (t-test; p<0.0001; bars ± 1 sd). E. As in D, but for ON synapses (n = 12, 15 and 6 synapses for respective conditions). Again, the average Fano factor was significantly higher in the morning (t-test; p<0.001).F. Example recordings from two OFF synapses stimulated at 60% contrast in three conditions:afternoon (top, deep red trace), morning (middle, rlight red trace) and after injection of ADTN in the morning (bottom, blue trace). The comparison with and without ADTN is from the same synapse. In each case the top trace shows the iGluSnFR signal and the lower trace the estimated Qe. The modulation in intensity is shown below. Note that events occur at different phases of the stimulus. G. Temporal jitter of events composed of different numbers of quanta in OFF synapses in the afternoon (n = 24 synapses), morning (n = 19) and morning + ADTN (n = 16). Release events occurring in the morning were significantly less phase-locked than the afternoon (p<0.001; t-test). Activation of D1 receptors reduced temporal jitter for events composed of 3 or more quanta (p < 0.004). The solid lines describing these relations in the three conditions are described by a single exponential decay function of the form y0+ Aexp((-(x-x0)/τ))) with y0= 4.23 ± 1.2 and A = 27 ± 7 in the afternoon; y0= 9.77 ± 1.4 and A= 28.64 ± 5.6 in the morning and y0=5.45 ± 1.3, A = 30. ± 6.1 after activation of D1 receptor in the morning. H. As G, but for ON synapses (n = 6-14). Activation of D1 receptors had no significant effect on temporal jitter. The relationship observed in the morning is described by a straight line with a = 34.7 ± 1.5 and a slope = -3.6 ± 0.5.

ii) Variability in stimulus-evoked responses

This was assessed by the Fano factor (F), the ratio of the variance-to-mean of the response measured over multiple trials33,34. In the morning, F was ∼2.6 in both ON and OFF synapses when averaged over a range of contrasts, falling to ∼1.6 in the afternoon (p < 0.001, t-test; Fig.3D-E). Notably, the variability of synaptic output was higher than expected for a Poisson process, for which F = 1. Activation of D1 receptors in the morning improved the reliability of synaptic responses to levels similar to those measured in the afternoon (Fig.3 D-E).

iii) Temporal jitter

Retinal ganglion cells (RGCs) encode information not just in their spike count but also in the timing of spikes35,36. Spike times can vary by just a few milliseconds and this accuracy depends on the precision of excitatory inputs received from bipolar cells37. The standard deviation in timing of release events (“temporal jitter”) was measured relative to the phase of a 5 Hz stimulus and the larger the release event the more precise it was on average (Fig. 3F-H). In OFF synapses the temporal jitter was 5-8 ms lower in the afternoon compared to the morning for events composed of up to 8 vesicles (Fig. 3G). Diurnal modulation of temporal precision was weaker in ON synapses and only significant for events composed of 1-3 vesicles (Fig. 3H). Increasing activation of D1 receptors in the morning reduced temporal jitter in OFF synapses (p < 0.004, t-test) but not ON.

iv) Changes in the distribution of multivesicular events

Previous studies quantifying the synaptic transfer of visual information have been limited by the inability to monitor individual active zones and used the assumption that vesicles are released according to Poisson statistics38-40. We know now that bipolar cells do not employ a simple rate-code and that visual information is also contained in the amplitude of multivesicular events23. To test whether modulation of contrast gain was accompanied by changes in the number of quanta in an event (Qe) we compared the amplitude distribution in the morning and afternoon. Fig. 4A makes this comparison for a stimulus of 60% contrast. In ON synapses, 68% of release events in the morning were univesicular, falling to 40% in the afternoon (p < 10−4, chi-squared test) and this shift was fully reversed by antagonizing D1 receptors (Fig. 4B). MVR was more prevalent in OFF synapses with only 38% of release events in the morning being univesicular but again there was a significant shift towards MVR in the afternoon (Fig. 4C; p<0.02). Qualitatively similar diurnal modulation of MVR was observed over a range of contrasts from 20% to 80%.

Figure 4.
Figure 4. Diurnal changes in the efficiency of information transmission

A. Examples of iGluSnFR signals from an individual synapse elicited (60% contrast, 5 Hz) in the morning, afternoon and afternoon after injecting SCH 23390. In each case the top trace shows the iGluSnFR signal and the lower trace the estimated Qe. B. Top: Changes in Qe in ON synapses in the morning and afternoon. In the afternoon the distribution was shifted toward multiquantal events (p<0.059, Chi-squared test). Bottom: Changes in the distribution of Qe in ON synapses before and after intravitreal injection of the D1 antagonist SCH23390. The distribution was shifted toward lower Qe (p<0.001) but was not significantly different to that measured in the morning. C. As B, but for OFF synapses. In the afternoon the distribution was shifted toward multiquantal events (p<0.007). In the afternoon, injection of SCH 23390 shifted the distribution toward uniquantal events (p<0.001). D. Examples of synaptic responses over 12 different contrasts spanning ±10% around the contrast eliciting the half-maximal response (C1/2) in the morning (top, light red), afternoon (middle, dark red) and after injection of D1 antagonist SCH 23390 in the afternoon (bottom, black; note the lower frequency and amplitude of release events). In each case the top trace shows the iGluSnFR signal and the lower trace the estimated Qe. Each contrast step lasted 2 s (5 Hz) and each trace is from a different OFF synapse. E. Mutual information I (S;Q) in four conditions: (i) morning, (ii) afternoon, iii) morning after injection of ADTN, (iv) afternoon after injection of SCH 23390. F. Specific information (I2) for events of different quantal content in OFF synapses (33 synapses). The curve describing the relation is a least-squares fit of a power function of the form i = AQex, with A = 0.20, and x = 1.81. G. As F, but for ON synapses (n = 13). The curve describing the relation is almost identical (A = 0.21, and x = 1.75).

Diurnal modulation of information transmission

How do changes in synaptic gain (Figs.1-2), noise and variability (Fig. 3) and MVR (Fig. 4-C) combine to alter the amount of visual information transmitted? A larger synaptic signal relative to noise (SNR) will increase the mutual information (I) between the response (q) and the stimulus generating it (S), although the size of the increase will depend on the statistical properties of both signal and noise41. In the simple situation where both have a Gaussian distribution I = 0.5log2(1+SNR)42. But how should we quantify the synaptic signal? When analyzing the spike code, all events comprise the same symbol and the response can be described as the number of spikes in each of a series of time bins41,43. The output from bipolar cells is qualitatively different with a visual stimulus being encoded both by the timing of release events and their amplitudes23. We therefore took an approach in which MVR composed of different numbers of vesicles were considered different symbols. The mutual information between the response and stimulus was then computed as the average amount of information about the stimulus gained from observing any symbol30,44.

In the morning, the average mutual information between stimulus and response was almost the same for ON and OFF synapses (0.445 ± 0.035 bits s-1 and 0.455 ± 0.03 bits s-1, respectively; Fig. 4E). In the afternoon mutual information increased through both channels but the increase in OFF synapses (370%) was significantly larger than in ON (270%; p < 0.001 by t-test; Fig. 4E). Mutual information in the morning was increased by activation of D1 receptors and in the afternoon it was decreased by antagonizing the effects of endogenous dopamine, although not to levels measured in the morning (Fig. 4E).

Together, the results in Figs. 1-4 demonstrate that dopamine is a key neuromodulator controlling diurnal changes in information transmission in the retina, but also leave open the possibility that other pathways contribute.

Diurnal changes in the efficiency of the vesicle code

The transmission of information using spikes and vesicles requires energy so an important question becomes the efficiency with which these symbols are used45. RGCs firing at low rates cells transmit ∼3.5 bits/spike while those that fire most briskly encode ∼2 bits/spike46. In comparison, the 2.7-fold increase in information transmitted through ON synapses in the afternoon (Fig. 4E) occurred without a change in the average rate of vesicle release (Fig. 2D and E). In OFF synapses, mutual information increased 3.7-fold while release rates around C1/2 increased just 2-fold (Fig. 2C and E). The diurnal increase in the average rate of vesicle release was therefore associated with a 1.4-to 2.7-fold increase in the average efficiency with which vesicles were used to encode changes in contrast. This situation is distinct to that observed in the binary spike code of post-synaptic RGCs, where an increase in spike rate is associated with a decrease in the efficiency of information transmission46.

How does the vesicle code become more efficient at higher average release rates? To understand this we need to consider changes in the distribution of vesicles: MVR events become more prevalent in the afternoon (Fig. 4A-C) and larger MVR events carry more information per vesicle23, as shown by the supralinear relation between the specific information carried by each synaptic symbol and the number of vesicles it contains (Fig. 4F-G). This relation did not, however, change in the afternoon compared to the morning, at least for events composed of 1-6 vesicles (Fig. 4F and G; chi-squared test). The amount of information carried by each synaptic symbol was therefore independent of diurnal changes in visual processing.

Discussion

Synapses allow the flow of information through circuits to be adjusted1,5 and this study provides a quantitative understanding of this idea in the context of the diurnal control of sensory processing. We find a dopamine-dependent increase in transmission of visual information that can be ascribed to four major effects: an increase in the number of vesicles released by a stimulus (Figs. 1 and 2), a decrease in spontaneous noise (Fig. 3A-C), reduced variability of evoked responses (Fig. 3D-H) and a shift from univesicular to multivesicular release (Fig. 4). The relative contributions of these various mechanisms differed between ON and OFF pathways but MVR was fundamental to both.

A recent combination of electrophysiology with correlative light-and electron-microscopy has led to the suggestion that it may be a fundamental mode of synaptic transmission throughout the nervous system11 and it is established that MVR can be modulated by activity7,47 and by presynaptic modulation, for instance through muscarinic acetylcholine receptors in the striatum10 or GABAB receptors in the cortex48. This study has demonstrated that potentiation of MVR can not only increase the amount of information that a sensory synapse transmits but also increases the efficiency with which it is transmitted. Ever-improving techniques for monitoring synaptic output in vivo will help us understand how far neuromodulators acting in other parts of the brain modulate MVR and the vesicle code that transmits information between neurons.

Methods

Zebrafish husbandry

Fish were raised and maintained under standard conditions on a 14 h light/10 h dark cycle40. To aid imaging, fish were heterozygous or homozygous for the casper mutation which results in hypopigmentation and they were additionally treated with1-phenyl-2-thiourea (200 µM final concentration; Sigma) from 10 hours post fertilization (hpf) to reduce pigmentation. All animal procedures were performed in accordance with the Animal Act 1986 and the UK Home Office guidelines and with the approval of the University of Sussex Animal Welfare and Ethical Review Board. More information about experimental design and reagents is available in the Life Sciences reporting Summary.

Transgenic fish

Experiments were carried out using the following transgenic lines of zebrafish:

  1. Tg(ribeye:;Zf-SyGCaMP2) expressing the synaptically-localized fluorescent calcium reporter SyGCaMP 2.0 in retinal bipolar cells under the ribeye-A promoter25.

  2. Tg(–1.8ctbp2:Gal4VP16_BH) fish that drive the expression of the transcriptional activator protein Gal4VP16 were generated by co-injection of I-SceI meganuclease and endofree purified plasmid into wild-type zebrafish with a mixed genetic background. A myocardium-specific promoter that drives the expression of mCherry protein was additionally cloned into the plasmid to allow for phenotypical screening of founder fish.

  3. Tg(10xUAS:iGluSnFR_MH) fish driving the expression of the glutamate sensor iGluSnFR under the regulatory control of the 10 x UAS enhancer elements were generated by co-injection of purified plasmid and tol2 transposase RNA into offspring of AB wildtype fish outcrossed to casper wildtype fish. The sequences for the myocardium-specific promoter driving the expression of enhanced green fluorescent protein (mossy heart) were added to the plasmid to facilitate the screening process.

  4. Tg(–1.8ctbp2:SyGCaMP6) fish were generated by co-injection of I-SceI meganuclease and endofree purified plasmid into wild-type zebrafish with a mixed genetic background. The GCaMP6f variant was kindly provided by L. Looger (Janelia Farm). This variant holds a T383S mutation in comparison to the commercially available GCaMP6-fast version (Addgene plasmid 40755).

  5. Tg(isl2b:nlsTrpR, tUAS:memGCaMP6f) which drives the expression of memGCaMP6f in the optic tectum was generated by co-injecting pTol2-isl2b-hlsTrpR-pA and pBH-tUAS-memGaMP6f-pA plasmids into single-cell stage eggs. Injected fish were out-crossed with wild-type fish to screen for founders.

Multiphoton Imaging In Vivo

Experiments were carried out in a total of 117 zebrafish larvae (7–9 days post-fertilization). Fish were immobilized in 3% low melting point agarose (Biogene) in E2 medium on a glass coverslip (0 thickness) and mounted in a chamber where they were superfused with E2. Imaging was carried out using a two-photon microscope (Scientifica) equipped with a mode-locked titanium-sapphire laser (Chameleon, Coherent) tuned to 915 nm and an Olympus XLUMPlanFI 20x water immersion objective (NA 0.95). To prevent eye movements, the ocular muscles were paralyzed by injection of 1 nL of α-bungarotoxin (2 mg/mL) behind the eye. Most imaging was carried out in the dorsal the retina.

The signal-to-noise ratio of the microscope was optimized by collecting photons through both the objective and a sub-stage oil condenser (Olympus, NA 1.4). Emission was filtered through GFP filters (HQ 535/50, Chroma Technology) before detection with GaAsP photomultipliers (H7422P-40, Hamamatsu). The signal from each detector passed through a current-to-voltage converter and then the two signals were added by a summing amplifier before digitization. Scanning and image acquisition were controlled under ScanImage v.3.6 software49. In iGluSnFR recordings images were acquired at 10 Hz (128 × 100 pixels per frame, 1 ms per line) while linescans were acquired at 1 kHz. In GCaMP recordings images were acquired at 20 Hz (128 × 50 pixels per frame, 1 ms per line). Full-field light stimuli were generated by an amber LED (lmax = 590 nm, Thorlabs), filtered through a 590/10 nm BP filter (Thorlabs), and delivered through a light guide placed close to the eye of the fish. These wavelengths will most effectively stimulate red and green cones. The microscope was synchronized to visual stimulation.

Stimulation protocols

Measurements of contrast sensitivity with SyGCaMP2 were made by stimulating the fish with a series of 10 s stimuli (full-field sinusoidal modulation at 5 Hz) around a mean intensity of 55 nW mm-2. Measurements of contrast sensitivity with iGluSnFR used 2 s stimuli. To measure the distribution of events amplitudes and the temporal precision fish were continuously stimulated for 30 s at a given contrast.

Luminance sensitivity was assessed by stimulating the fish with a series of light steps (4 × 3 s) at 9 different light intensities increasing in steps of 0.5 log unit steps ranging from 11 pW mm-2 to 110 nW mm-2 (equivalent to 3.3 × 1011 photons mm-2).

Drug injections

Dopamine signalling was manipulated by injecting the antagonist of D1 receptors SCH 23390 at a final estimated concentration of 200 nM (Sigma).Finally, the long-lasting dopamine receptor ligand [3H] 2-amino-6,7-dihydroxy 1,2,3,4-tetrahydronapthalene (ADTN) (Sigma) was injected to a final estimated concentration of 200 nM. We confirmed that these drugs gained access by including 1 mM Alexa 594 in the injection needle; within 5 mins of injection the dye could be detected within the inner plexiform layer of the retina. Vehicle injection did not affect synaptic responses to varying contrast.

Quantal decomposition of iGluSnFR signals

To detect and quantize glutamatergic events from line-scans, there were six major steps, described in detail previously23. Steps 4-6 are summarized in Extended Data Figure 4.

  1. Separation of regions of interest (ROIs) by spatial decomposition using Gaussian fits to the average intensity profile across the line-scan. Typically 1-3 hot-spots of iGluSnFR fluorescence were detected in a single terminal, representing active zones within the line-scan. Active zones more than 1-1.5 μm apart could be distinguished from eah other.

  2. Time series extraction by weighted averaging. Once each spatial component had been defined, a time series for that component, F(t), was computed as the integral of the Gaussian fit to the linear intensity profile estimated in 1.

  3. Baseline correction and calculation of ΔF/F. Bleaching sometimes occurred during an observation episode and was corrected using a linear function of time. The iGluSnFR signal used for all analyses was the relative change in fluorescence, ΔF/F0, calculated from the bleach-corrected signals. The most frequent value (that is, the baseline) of the trace was used as F0.

  4. Identification of events by Wiener deconvolution. Release events within an active zone were identified by their characteristic kinetics using a Wiener filter. The time-course of iGluSnFR transients could be described by a function of the form: Embedded Image where A describes the amplitude of the event and τr and τf and are the time constants for rise and fall in the signal, respectively. Transients at most synapses could be described using a kernel with parameters of τf = 0.06 s and τr = 0.001 s. These parameters were relatively invariant for transients of different amplitudes so that iGluSnFR transients could be described as a linear time-invariant system (LTI). The filter described by equation 1 therefore allowed us to both “denoise” signals and estimate the underlying input. The result of the Wiener deconvolution was a time series in which glutamate release events were described approximately as Dirac-δ impulse.

  5. Extraction of events. Although the use of Wiener deconvolution significantly improved the signal-to-noise ratio, the baseline in the deconvolved traces was not noiseless, making it necessary to set a threshold for counting a deviation in this signal as an event. The distribution of values in the deconvolved trace was measured and the threshold set at 3-4 standard deviations above the basline. Events were then timed at the local maximum in the deconvolved trace above this threshold.

  6. Amplitude clustering to create a time series of quantized events. A histogram of event amplitudes at an individual active zone could be described as the sum of Gaussians. The quantal amplitude, was then calculated as the average inter-peak distance, which was always similar to the amplitude of the first peak, indicating that the first peak in the distribution represents vesicles released individually. Based on the evidence that glutamate transients of varying amplitude were integer multiples of a unitary event or quantum, we partitioned events into numbers of quanta using a Gaussian Mixture Model (GMM). Under this framework, the probability p of a value x is given by: Embedded Image where N is the normal probability density function and Φ represents the mixing probability and sums to one. We used the Expectation-Maximization (EM) algorithm for clustering and this was run up to 15 times on the synapse, each with a different number of components and with the mean and variance parameters initialized randomly, until acceptable convergence had been reached. Defining a time series as the number of quanta within each event allowed for computation of vesicle release rates and information theoretic measures.

Calculation of temporal jitter

In order to quantify variability in the timing of glutamatergic events, we first calculated the vector strength, rq, for events composed of q quanta: Embedded Image where tqi is the time of the ith q-quantal event, T is the stimulus period, and Nq is the total number of events of composed of q-quanta. The temporal jitter, Jq, can then be calculated as: Embedded Image where f is the stimulus frequency.

Calculations based on Information Theory

To quantify the amount of information about a visual stimulus that is contained within the sequence of release events from an active zone we first needed to convert bipolar cell outputs into a probabilistic framework from which we could evaluate the specific information (I2), a metric that quantifies how much information about one random variable is conveyed by the observation a specific symbol of another random variable41. The time series of quantal events was converted into a probability distribution by dividing into time bins of 20 ms, such that each bin contained either zero events or one event of an integer amplitude. We then counted the number of bins containing events of amplitude 1, or 2, or 3 etc. By dividing the number of bins of each type by the total number of bins for each different stimulus, we obtained the conditional distribution of Q given S, p(Q|S), where Q is the random variable representing the quanta/bin and S is the random variable representing the stimulus contrasts presented throughout the course of the experiment. In the absence of information about the distribution of contrasts normally experienced by a larval zebrafish, a uniform distribution of contrasts was used for S. Each contrast step lasted 2 s (5 Hz) and they were presented in two different pseudo-random orders, of which one is shown in Fig. 4D. The contrast sensitivity varied between synapses and between morning and afternoon (Fig. 1E-G) so to make allowance for this the stimulus set S was adjusted for each synapse to span contrasts ±10% around C1/2 measured within that synapse.

We computed the joint probability distribution by the chain rule for probability (given the experimentally defined uniform distribution of stimuli S): Embedded Image In order to convert this distribution into the conditional distribution of S given Q, we used the definition of the conditional distribution: Embedded Image From these distributions we computed two metrics: the mutual information I(S;Q)50 and specific information I2(S;q) 44. Mutual information is defined traditionally as: Embedded Image Embedded Image The specific information, I2(S;q), is defined as the difference between the entropy of the stimulus S minus the conditional entropy of the stimulus given the observed symbol in the response q: Embedded Image Embedded Image representing the amount of information observing each quantal event type q ϵ Q carries about the stimulus distribution S. Note that mutual information can also be computed from the specific information as the dot product of the specific information vector I2 and the vector describing the probability of an event of a given quantal size p(q). This adds to the interpretability of both metrics – the specific information is the amount of information a single (specific) symbol gives about the stimulus, and the mutual information is the average amount of information about the stimulus gained from observing any symbol.

Statistics

All data are given as mean ± s.e.m. unless otherwise stated in the figure legends. All statistical tests met appropriate assumptions and were calculated using inbuilt functions in IgorPro (Wavemetrics). When data were not normally distributed we used non-parametric tests. All tests were two-sided and significance defined as P < 0.05. Data collection was not randomized because all experiments were carried out within one set of animals. Delivery of different stimuli was randomized where appropriate. Data were only excluded from the analysis if the signal-to-noise ratio (SNR) of the iGluSnFR signals elicited at a given synapse was not sufficient to detect unitary responses to visual stimuli with a SNR of at least three.

Author contributions

J.M-D. conceived, designed and executed experiments, analyzed results and prepared the manuscript. B. J. carried out analysis and wrote code. F. E. conceived and executed experiments with SyGGCaM2 and carried out analysis. J. J. conceived and executed experiments and carried out analysis. L. L. conceived the project, designed experiments, analyzed data, repaired equipment, wrote code and prepared the manuscript.

Extended Data

Extended Data Figure 1.
Extended Data Figure 1. Diurnal changes in luminance-sensitivity of bipolar cell synapses

A. Averaged responses from ON terminals to light steps of different irradiance measured at Zeitgeber time 1, 10 and 16 hours. Note large variations in amplitude and kinetics. Each light step was of 3 s (n = 535 terminals from 10 fish). B. Peak response as a function of irradiance for ON terminals in B. The smooth lines are Hill functions of the form R = Rmax*(Ih/(Ih + I1/2h)), where R is the peak response, I is the irradiance, h is the Hill coefficient and I1/2 is the irradiance generating the half-maximal response. At ZT = 16 hrs: Rmax = 0.91 ± 0.04; h = 2.0 ± 0.2; I1/2 = 0.066 ± 0.02 nW/mm2 (dashed blue arrow). At ZT = 10 hrs: Rmax = 0.85 ± 0.06; h = 0.8 ± 0.1; I1/2 = 0.65 ± 0.18 nW/mm2. At ZT = 1 hrs: Rmax = 0.853 ± 0.02; h = 0.9 ± 0.2; I1/2 = 0.88 ± 0.18 nW/mm2 (red arrow). C. Variations in luminance sensitivity as a function of Zeitgeber time averaged across both ON and OFF terminals (n=535 and 335, respectively). The lower bar shows the timing of the light-dark cycle.

Extended Data Figure 2.
Extended Data Figure 2. Decomposition of iGluSnFR signals into vesicle counts

Summary of the basic steps for quantal decomposition of iGluSnFr signals. 1. Raw trace extracted from individual active zones (linescan, 1 KHz). 2. Deconvolved trace using the estimated Wiener filter. The dashed red line shows the threshold for counting evebts above the noise. 3. Histogram of event amplitudes for a representative active zone (373 events accumulated using stimulus contrasts of 20%, 60% and 100% and a frequency of 5 Hz). The black line is a fit of eight Gaussians, identified using a Gaussian mixture model. Note that the variance of successive Gaussians did not increase in proportion to the peak number. The first peak had a value of 0.24, and the distance between peaks averaged 0.25, indicating the existence of a quantal event equivalent to ∼0.25. 4. Estimation of the number of quanta per event. For more details about analyses see James et al., 2019.

Extended Data Figure 3.
Extended Data Figure 3. Diurnal changes in luminance sensitivity

A. Left: Averaged SyGaMP2 signals at different Zeitgeber time. ON terminals green and OFF terminas red. Each step of light (10 nW mm-2) lasted 3 s. These averages are only from responsive terminals. Right: the percentage of terminals generating a significant response to the same light step (averaged across both ON and OFF). Bars show SD. B. Effects of manipulating dopamine signalling on luminance sensitivity of the ON channel. Luminance vs. response plots for ON terminals. Red circles compare this function at ZT 1 hr under control conditions (solid circle) and after injection of the non-selective dopamine receptor agonist ADTN (∼0.2 μM; open circles). ADTN caused a prompt change in the luminance-response function to forms measured at ZT 10 hrs (solid black circles), increasing Rmax from 0.53 ± 0.02 to 1.02 ± 0.07, and reducing I1/2 from 0.88 ± 0.18 nW mm-2 to 0.05 ± 0.02 nW mm-2 (± sd, as estimated from the fitted Hill function shown). The higher gain and luminance sensitivity at ZT 10 hrs could be explained as an effect of dopamine at D2 receptors, because it was completely reversed by injection of the selective D2 receptor antagonist sulpiride (∼2 μM; open black circles; Rmax = 0.57 ± 0.13, I1/2 = 1.16 ± 1.34 nW mm- 2). Results collected from n = 535 terminals from 38 fish. C. Effects of manipulating dopamine signalling on luminance sensitivity of the OFF channel. Comparing control responses at ZT 1 hr and 10 hrs showed a significant reduction in I1/2 from 3.9 ± 1.3 to 0.0128 ± 0.005, but without a significant change in Rmax (0.35 ± 0.03 vs 0.30 ± 0.02). ADTN injected at ZT 1 caused a prompt increase in luminance sensitivity, reducing I1/2 to 0.013 ± 0.005 nW mm-2. The higher luminance sensitivity at ZT 10 hrs could be partly explained as an effect of dopamine at D2 receptors, because injection of sulpiride (∼2 μM; open black circles) increased I1/2 from 0.05 ± 0.01 nW mm-2 to 0.35 ± 0.14 nW mm-2. Results collected from n = 355 terminals from 38 fish.

Acknowledgements

The authors express many thanks to all the members of Lagnado laboratory for discussions during the cours of this project. We also thank Tom Baden for his many insightful criticisms and suggestions. This work was supported by a grant to L.L. from the Wellcome Trust (102905/Z/13/Z).

References

  1. ↵
    Bargmann, C. I. & Marder, E. From the connectome to brain function. Nat Methods 10, 483–490, doi:nmeth.2451 [pii] 10.1038/nmeth.2451 (2013).
    OpenUrlCrossRefPubMedWeb of Science
  2. ↵
    McCormick, D. A., Nestvogel, D. B. & He, B. J. Neuromodulation of brain state and behavior. Annual review of neuroscience 43, 391–415 (2020).
    OpenUrlCrossRefPubMed
  3. ↵
    Nadim, F. & Bucher, D. Neuromodulation of neurons and synapses. Current opinion in neurobiology 29, 48–56, doi:10.1016/j.conb.2014.05.003 (2014).
    OpenUrlCrossRefPubMed
  4. ↵
    Patzke, C. et al. Neuromodulator signaling bidirectionally controls vesicle numbers in human synapses. Cell 179, 498-513.e422 (2019).
    OpenUrl
  5. ↵
    Rusakov, D. A., Savtchenko, L. P. & Latham, P. E. Noisy Synaptic Conductance: Bug or a Feature? Trends Neurosci 43, 363–372, doi:10.1016/j.tins.2020.03.009 (2020).
    OpenUrlCrossRef
  6. ↵
    Euler, T., Haverkamp, S., Schubert, T. & Baden, T. Retinal bipolar cells: elementary building blocks of vision. Nat Rev Neurosci 15, 507–519, doi:10.1038/nrn3783 (2014).
    OpenUrlCrossRefPubMed
  7. ↵
    Wadiche, J. I. & Jahr, C. E. Multivesicular release at climbing fiber-Purkinje cell synapses. Neuron 32, 301–313 (2001).
    OpenUrlCrossRefPubMedWeb of Science
  8. ↵
    Glowatzki, E. & Fuchs, P. A. Transmitter release at the hair cell ribbon synapse. Nat Neurosci 5, 147–154, doi:10.1038/nn796 (2002).
    OpenUrlCrossRefPubMedWeb of Science
  9. ↵
    Christie, J. M. & Jahr, C. E. Multivesicular release at Schaffer collateral-CA1 hippocampal synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience 26, 210–216, doi:10.1523/JNEUROSCI.4307-05.2006 (2006).
    OpenUrlAbstract/FREE Full Text
  10. ↵
    Higley, M. J., Soler-Llavina, G. J. & Sabatini, B. L. Cholinergic modulation of multivesicular release regulates striatal synaptic potency and integration. Nature neuroscience 12, 1121–1128 (2009).
    OpenUrlCrossRefPubMedWeb of Science
  11. ↵
    Holler, S., Köstinger, G., Martin, K. A. C., Schuhknecht, G. F. P. & Stratford, K. J. Structure and function of a neocortical synapse. Nature 591, 111–116, doi:10.1038/s41586-020-03134-2 (2021).
    OpenUrlCrossRef
  12. ↵
    Ribelayga, C., Cao, Y. & Mangel, S. C. The circadian clock in the retina controls rod-cone coupling. Neuron 59, 790–801, doi:10.1016/j.neuron.2008.07.017 (2008).
    OpenUrlCrossRefPubMedWeb of Science
  13. ↵
    Nie, K. et al. Effects of circadian clock protein Per1b on zebrafish visual functions. Chronobiology international 35, 160–168 (2018).
    OpenUrl
  14. ↵
    Li, L. Circadian Vision in Zebrafish: From Molecule to Cell and from Neural Network to Behavior. J Biol Rhythms 34, 451–462, doi:10.1177/0748730419863917 (2019).
    OpenUrlCrossRef
  15. ↵
    Zang, J. et al. Circadian Regulation of Vertebrate Cone Photoreceptor Function. bioRxiv, 2021.2005.2005.442714, doi:10.1101/2021.05.05.442714 (2021).
    OpenUrlAbstract/FREE Full Text
  16. ↵
    Nikolaev, A., Leung, K. M., Odermatt, B. & Lagnado, L. Synaptic mechanisms of adaptation and sensitization in the retina. Nat Neurosci 16, 934–941, doi:10.1038/nn.3408 (2013).
    OpenUrlCrossRefPubMed
  17. ↵
    Zimmermann, M. J. Y. et al. Zebrafish Differentially Process Color across Visual Space to Match Natural Scenes. Curr Biol 28, 2018-2032.e2015, doi:10.1016/j.cub.2018.04.075 (2018).
    OpenUrlCrossRef
  18. Franke, K. et al. Inhibition decorrelates visual feature representations in the inner retina. Nature 542, 439–444, doi:10.1038/nature21394 (2017).
    OpenUrlCrossRefPubMed
  19. Johnston, J. et al. A Retinal Circuit Generating a Dynamic Predictive Code for Oriented Features. Neuron 102, 1211–1222 e1213, doi:10.1016/j.neuron.2019.04.002 (2019).
    OpenUrlCrossRef
  20. ↵
    Matsumoto, A. et al. Direction selectivity in retinal bipolar cell axon terminals. Neuron (2021).
  21. ↵
    Lagnado, L. & Schmitz, F. Ribbon Synapses and Visual Processing in the Retina. Annu Rev Vis Sci 1, 235–262, doi:10.1146/annurev-vision-082114-035709 (2015).
    OpenUrlCrossRefPubMed
  22. ↵
    Singer, J. H., Lassova, L., Vardi, N. & Diamond, J. S. Coordinated multivesicular release at a mammalian ribbon synapse. Nat Neurosci 7, 826–833, doi:10.1038/nn1280 (2004).
    OpenUrlCrossRefPubMedWeb of Science
  23. ↵
    James, B., Darnet, L., Moya-Diaz, J., Seibel, S. H. & Lagnado, L. An amplitude code transmits information at a visual synapse. Nat Neurosci 22, 1140–1147, doi:10.1038/s41593-019-0403-6 (2019).
    OpenUrlCrossRefPubMed
  24. ↵
    Auger, C., Kondo, S. & Marty, A. Multivesicular release at single functional synaptic sites in cerebellar stellate and basket cells. The Journal of neuroscience : the official journal of the Society for Neuroscience 18, 4532–4547 (1998).
    OpenUrlAbstract/FREE Full Text
  25. ↵
    Dreosti, E., Odermatt, B., Dorostkar, M. M. & Lagnado, L. A genetically encoded reporter of synaptic activity in vivo. Nat Methods 6, 883–889, doi:10.1038/nmeth.1399 (2009).
    OpenUrlCrossRefPubMedWeb of Science
  26. ↵
    Marvin, J. S. et al. An optimized fluorescent probe for visualizing glutamate neurotransmission. Nat Methods 10, 162–170, doi:10.1038/nmeth.2333 (2013).
    OpenUrlCrossRefPubMedWeb of Science
  27. ↵
    Jackson, C. R. et al. Retinal dopamine mediates multiple dimensions of light-adapted vision. The Journal of neuroscience : the official journal of the Society for Neuroscience 32, 9359–9368, doi:10.1523/JNEUROSCI.0711-12.2012 (2012).
    OpenUrlAbstract/FREE Full Text
  28. ↵
    Esposti, F., Johnston, J., Rosa, J. M., Leung, K. M. & Lagnado, L. Olfactory stimulation selectively modulates the OFF pathway in the retina of zebrafish. Neuron 79, 97–110, doi:10.1016/j.neuron.2013.05.001 (2013).
    OpenUrlCrossRefPubMed
  29. ↵
    Roy, S. & Field, G. D. Dopaminergic modulation of retinal processing from starlight to sunlight. J Pharmacol Sci 140, 86–93, doi:10.1016/j.jphs.2019.03.006 (2019).
    OpenUrlCrossRefPubMed
  30. ↵
    Shannon, C. A Mathematical Theory of Communication. Bell System Technical Journal 27, 379–423 (1948).
    OpenUrl
  31. ↵
    Faisal, A. A., Selen, L. P. & Wolpert, D. M. Noise in the nervous system. Nat Rev Neurosci 9, 292–303, doi:10.1038/nrn2258 (2008).
    OpenUrlCrossRefPubMedWeb of Science
  32. ↵
    London, M., Roth, A., Beeren, L., Hausser, M. & Latham, P. E. Sensitivity to perturbations in vivo implies high noise and suggests rate coding in cortex. Nature 466, 123–127, doi:10.1038/nature09086 (2010).
    OpenUrlCrossRefPubMedWeb of Science
  33. ↵
    Churchland, M. M. et al. Stimulus onset quenches neural variability: a widespread cortical phenomenon. Nature Neuroscience 13, 369–378, doi:10.1038/nn.2501 (2010).
    OpenUrlCrossRefPubMedWeb of Science
  34. ↵
    Fairhall, A. L. Whither variability? Nature Neuroscience 22, 329–330, doi:10.1038/s41593-019-0344-0 (2019).
    OpenUrlCrossRef
  35. ↵
    Berry, M. J., Warland, D. K. & Meister, M. The structure and precision of retinal spike trains. Proceedings of the National Academy of Sciences 94, 5411, doi:10.1073/pnas.94.10.5411 (1997).
    OpenUrlAbstract/FREE Full Text
  36. ↵
    Rathbun, D. L., Warland, D. K. & Usrey, W. M. Spike timing and information transmission at retinogeniculate synapses. Journal of Neuroscience 30, 13558–13566 (2010).
    OpenUrlAbstract/FREE Full Text
  37. ↵
    Cui, Y., Wang, Y. V., Park, S. J. H., Demb, J. B. & Butts, D. A. Divisive suppression explains high-precision firing and contrast adaptation in retinal ganglion cells. eLife 5, e19460, doi:10.7554/eLife.19460 (2016).
    OpenUrlCrossRefPubMed
  38. ↵
    Freed, M. A. Quantal encoding of information in a retinal ganglion cell. J Neurophysiol 94, 1048–1056, doi:10.1152/jn.01276.2004 (2005).
    OpenUrlCrossRefPubMedWeb of Science
  39. Choi, S. Y. et al. Encoding light intensity by the cone photoreceptor synapse. Neuron 48, 555–562, doi:10.1016/j.neuron.2005.09.011 (2005).
    OpenUrlCrossRefPubMedWeb of Science
  40. ↵
    Odermatt, B., Nikolaev, A. & Lagnado, L. Encoding of luminance and contrast by linear and nonlinear synapses in the retina. Neuron 73, 758–773, doi:10.1016/j.neuron.2011.12.023 (2012).
    OpenUrlCrossRefPubMed
  41. ↵
    Stone, J. V. Principles of Neural Information Theory: Computational Neuroscience and Metabolic Efficiency (Sebtel Press, 2018).
  42. ↵
    Schultz, S. R. Signal-to-noise ratio in neuroscience. Scholarpedia 2, 2046 (2007).
    OpenUrl
  43. ↵
    Borst, A. & Theunissen, F. E. Information theory and neural coding. Nat Neurosci 2, 947–957, doi:10.1038/14731 (1999).
    OpenUrlCrossRefPubMedWeb of Science
  44. ↵
    DeWeese, M. R. & Meister, M. How to measure the information gained from one symbol. Network 10, 325–340 (1999).
    OpenUrlCrossRefPubMedWeb of Science
  45. ↵
    Laughlin, S. B. Energy as a constraint on the coding and processing of sensory information. Curr Opin Neurobiol 11, 475–480, doi:10.1016/s0959-4388(00)00237-3 (2001).
    OpenUrlCrossRefPubMedWeb of Science
  46. ↵
    Koch, K. et al. How much the eye tells the brain. Curr Biol 16, 1428–1434, doi:10.1016/j.cub.2006.05.056 (2006).
    OpenUrlCrossRefPubMedWeb of Science
  47. ↵
    Oertner, T. G., Sabatini, B. L., Nimchinsky, E. A. & Svoboda, K. Facilitation at single synapses probed with optical quantal analysis. Nature neuroscience 5, 657–664 (2002).
    OpenUrlCrossRefPubMedWeb of Science
  48. ↵
    Chalifoux, J. R. & Carter, A. G. GABAB receptors modulate NMDA receptor calcium signals in dendritic spines. Neuron 66, 101–113 (2010).
    OpenUrlCrossRefPubMedWeb of Science
  49. ↵
    Pologruto, T. A., Sabatini, B. L. & Svoboda, K. ScanImage: flexible software for operating laser scanning microscopes. Biomed Eng Online 2, 13, doi:10.1186/1475-925X-2-13 (2003).
    OpenUrlCrossRefPubMed
  50. ↵
    de Ruyter van Steveninck, R. R., Lewen, G. D., Strong, S. P., Koberle, R. & Bialek, W. Reproducibility and variability in neural spike trains. Science 275, 1805–1808, doi:10.1126/science.275.5307.1805 (1997).
    OpenUrlAbstract/FREE Full Text
  51. Pola, G., Schultz, S. R, Petersen, R. S & Panzeri, S. in Neuroscience Databases: A Practical Guide (Springer, Boston, MA, 2003).
Back to top
Posted September 14, 2021.
Download PDF
Email
Diurnal modulation of multivesicular release controls the efficiency of information transmission at a sensory synapse
José Moya-Díaz, Ben James, Federico Esposti, Jamie Johnston, Leon Lagnado
bioRxiv 2021.09.12.459944; doi: https://doi.org/10.1101/2021.09.12.459944
Reddit logo Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area