 Good afternoon, everybody. Welcome to Sussex Vision Series, and today we are delighted to introduce our next speaker, Professor Weiwei, from the University of Chicago. Good afternoon, everybody. Welcome to the Sussex Vision Series. Sorry. Sorry. So, Professor Weiwei, can you hear me? Professor Weiwei is actually an associate professor with Turner in the Department of Neurobiology at the University of Chicago. And in terms of academic background, she did her postdoctoral research at the University of California in Berkeley with Professor Marla Feuler. And her PhD was carried out at the Watson School of Biological Science at the Cold Spring Harbor Laboratory in New York under the supervision of Professor Robert Tomalino. So, during this stage, she studied many aspects of synaptic plasticity, and nowadays she will be presenting a really interesting talk, which is titled, Context-Dependent Motion Processing in the Retina. So, Weiwei, thanks a lot for accepting our invitation, and it is a pleasure for us to host you today. Thank you so much, Jose, for the kind introduction. I'm going to share my slides. You can share now your screen if you want. Yeah, does it work? Is it working? Okay, one second. Okay, I'll share again. You have to share. No, yes, it's perfect. Okay. It's working? Yeah, brilliant. Okay, thank you again. So, a general goal of my group is to understand the neural basis of sensory processing. And the influential conceptual framework for analyzing neural circuits is David Marl's three levels of analysis. So the first level is computation. So what does the circuit compute? What's the function of the circuit? Next is the algorithm. So what is the algorithm used for carrying out such a computation? And then lastly, how is the algorithm implemented by biophysical substrates in the nervous system, such as synapses, dendrites and neural networks? And in this framework, in a visual system, a prominent computation is the direction selectivity. And direction selectivity is first discovered by human whistle in the cat primary visual cortex. And they find that some cells in the visual cortex have a directionally tuned to moving stimulus. So they fire most strongly when the stimulus is moving in one direction, but minimally when the stimulus moves in the opposite direction. And a few years later, direction selectivity was characterized by Barton Levick in the very early stage of visual processing in the rabbit retina. So with many years of studies, we now know that in the retina, there are mainly two groups of direction selective ganglion cell types, the on and off types, that are also tuned to the direction of image motion. And also very exciting discoveries recently in a primate retina by the anesthetics group and nays group at the University of Washington, showing that these direction selective cells are also present in the primates. But in this talk, I will focus on the on off type of direction selective ganglion cells. And so in this slide, this is on off DSGC in a mouse retina. So here is the top view of the dendrites. And this is the side view. So these cells have two layers of dendrites one in the on sub lamina and the other in the off sub lamina of the inner plex from layer IPL. And here's example tuning curve plotting the number of spikes as a function of motion direction. So the arrow in the middle is the vector some pointing to the preferred direction, and the opposite direction is a null direction. And there are four sub types of off DS ganglion cells in the retina, and they their preferred directions aligned two cardinal axis vertical and horizontal axis. So these cells projects to major visual pathways centrally. So the, they, their actions leave the retina and project to the, the shell of the door so LGN, where, and then from LGN, the information is further conveyed to the visual cortex. Many studies looking at how the information from those off DS ganglion cells are used by the downstream pathways. And so another major target of those cells are the superior clipless. So here, the relationship between the retina and the clipless is a bit simpler. So in a collaborative study with Stacy Kahn's group. We're exploring this relationship. So JC group found that the retina inputs to the direction selective clicker neurons direction. So, so the direction selective clicker neurons receive tuned retina inputs. So when we use conditional knockout mouse line where the retinal direction selectivity is selectively disrupted. JC group found that the direction selectivity in the clipless is impaired, indicating that the direction selectivity in the clipless is inherited from the, from the retina. So these cells has been a classical model to study direction selectivity and the underlying neural mechanisms. So there are many multiple mechanism that that all contribute to the direction selectivity of those cells. And the key one of the key mechanism is the now direction inhibition. So in this slide, I'm, I'm, I will just briefly introduce this, this, the mechanism of the now direction inhibition. So this is an example, spiking activity from off DS ganglion cells, when the bar is moving in a preferred direction, use, we see a lot of spiking, but in the opposite direction, the self doesn't fire. And then when we measure the inhibitory current of the cell, we see that the inhibitory inputs to the cell is directionally tuned in the now direction. They are long and fast, so that it can veto excitation and suppress spiking, but in the preferred direction, the inhibition is weak and delayed to allow max firing. And the source of this inhibitory current comes from the emicrin cell called starburst emicrin cells. So these cells form government urgic synapses to the direction selective ganglion cells. So starburst cell and ganglion cells receive light evoked glutamatergic input from this vertical for the receptor to bipolar cell pathway. When the bar is moving in the now direction of the ganglion cell, the starburst are released maximum amount of Gabba onto the ganglion cell, and this is underlies this now direction inhibition. So to summarize this in a simplified way within this Mars three levels of framework. So the computation of the circuit is, is to compute the direction of visual motion. And it use an algorithm of now direction inhibition. And this now direction inhibition is implemented by the direction selective Gabba release at the synapses between the starburst emicrin cells and the ganglion cells. But in this talk, I will move away a little bit from this implementation of directions activity. But rather, I'd like to showcase a few recent examples from our lab showing that there are many context dependent phenomenon occurring at all three levels of the circuitry at the computation level at the algorithm level and at the implementation level. And there are three scenarios that we particularly looked into one is when the background of the visual motion is noisy. So in this case, this, such as in this example the duck is moving through this glistening water. The second scenario is when the motion trajectory is interrupted by occluders in the scene what happens to the ganglion cell response and what are the underlying mechanisms that mediates those response. And then this, the third mechanism, the third scenario is how the ganglion cell response is affected by by prior visual experience. And then by three brilliant former grad students in the lab. So the first one, the noise project is done by Christian, who is now a postdoc at UW in front Ricky's lab. And then the second project was done by Jennifer Dean, who is now a postdoc at Harvard with Chris Harvey, and then the final project was led by Lindsay, who is now the postdoc at young dance lab at Berkeley. So I will just briefly summarize the, the conclusions of the first two and and focus on the last story which is just published last week at journal of neuroscience. Okay, so first, this is a summary slide for the noise study. So we found that in the retina for this direction selective circuit that is dedicated to compute motion direction. So there is a neural mechanism that that mediates the noise resilience of this motion detection. So when the background have noise, the directions activity can be preserved over a wide range of noise levels. And this noise resilience is implemented by a circuit motif is a disinhibitory motif consists of zero inhibition among Stubbers, immigrant cells before the Stubbers are inhibiting inhibits, the ganglion cell. So with the connection with Rob Smith at U Penn, we figure out that the, the, this disinhibitory motif mediates noise resilience using algorithm that is not disinhibition. So instead of disinhibition, this motif preserve, preserve motion evoked inhibition. And this surprise, this, you know, this is initially quite surprising to us, but we found that the visual noise in the background will engage in a specific network dynamics. And that interacts with synaptic plasticity mechanism between the Stubbers and ganglion cells. So basically the interaction between network dynamics and synaptic plasticity can invert the algorithm of this canonical disinhibitory motif. So that instead of disinhibition, it can mediate. It can preserve inhibition instead. And in the second project, Jen looked at how on off DS ganglion cells respond to moving stimulus, if their, if the motion trajectory is interrupted by occluders. For example, instead of this just continuous moving bars or drifting gradings. If the moving stimulus is occluded by stationary objects, or a moving object emerged from a cruder. So first of all, this project was in collaboration with my theory colleague Stephanie Palmer and her former undergrad Albert Chen, who help us with the theoretical analysis, and also with David Burson who help us with connectomic analysis on the underlying circuitry. So the take all message is that when the moving stimulus is moving in a continuous trajectory along a continuous trajectory, we have this canonical computation of the direction of the moving objects, and this is implemented by the now direction by the Stubbers DS inhibition. So this is the, the normal DS mechanisms. But when the stimulus is moving. When the trajectory is interrupted, then the population of DS ganglion cells, transiently switch from encoding the direction of motion into encoding the spatial location of the motion interruption, and this is suggested by the theoretical analysis by Stephanie's group. And this is implemented by a neural mechanism that involves a spatial displacement of the accessory separate field of those ganglion cells. And because during interrupted motion. And because of this spatial displacement of receptive fields at the side of the motion interruption. The, so the two DS ganglion cell populations that are tuned to the opposite motion directions. But transient synchrony because the, the, the ganglion cells that are normally tuned to the opposite direction of the motion can fire now direction response at the side of motion interruption. So there is a synchronous firing between oppositely tuned ganglion cell populations but only at the, at the side of motion interruption. So this local synchrony is very beneficial is very salient is very detectable, and is really help, helpful for encoding or pinpointing the precise location in during the motion interruption. And one way to think about this is that the population response of the ganglion cells that tunes to the two directions of the motion axes will normally encode the direction of the moving object. But at the side of motion interruption, this population response will transiently switching from encoding mode direction to encoding location, which could be beneficial for the animals survival. So the for the rest of my talk, I'd like to focus on, on the recent project we, we, which look at how the DS ganglion cells response is influenced by prior visual experience. And this is again, in collaboration with Robert Smith at up and led by a former grad student Lindsay in the lab. So the question we, we ask is, how, how is DS ganglion cell response influenced by by prior visual simulation. And the initial reason we want to look at this question is by the anecdote, anecdotal observation in the lab. So when we started recording from when we start recording from a retina we always have a set protocol so we'll first test the lie responsiveness of the retina by showing a spot like a flashing spot to make sure the retina is lie responsive the tissue is healthy. And then we show we start to do a tuning curve from the DS ganglion cells by showing moving bars for example. And after the tuning curve we usually show the flashing spot again to check whether, you know, the cell is still there is whether the tissue is still a responsive. But, but, you know, everyone in the lab has has the impression that when we first show the flashing spot, the ganglion cell response moderately. But after the tuning curve was moving stimulus, then if we show the flashing spot again, we all think the flash response become more robust and stronger. But this is all anecdotal Erwin sees that, but when Lindsay joined the lab she decided to study this phenomenon more systematically. She designed an induction protocol to induce this type of sensitization. So, so what she did is she show a test spot. This is a very weak spot to elicit moderate response in the ganglion cells and the test spots is delivered every four seconds with one second on time. So after a period of testing spots, we have this baseline farm rate to the flashing response. And then we. So she showed induction stimulus. This could be drifting, drifting gratings or contrast reversing gratings or moving spots. And both all these induction stimuli can induce sensitization. So after this induction stimulus she should use the test box again to test the sensitivity of the ganglion cells to to light flashes. So here is an example response from a dorsal DS ganglion cells. So here upper panel is the farm rate, the lower panel is the spiking raw spiking trace. The black is the before the induction stimulus the red is after the induction stimulus. You can see that before the induction stimulus we have this baseline on and off response from this ganglion cells, but after this 20 seconds of induction stimulus, both the on and off response is stronger. And we also see the appearance of a sustained firing we caught we caught that sustained component after the induction. But if we check this in the ventral ganglion cells, we only see the sensitization of the off response, but we didn't see the appearance of the sustained components, and we also didn't see a sensitization of the on response. So as to here's a schematic, the ventral cells, their respective fields are in the upper visual field, and the dorsal cells have receptive fields in the lower visual field. And we quantified the sensitization with index, which is the normalized difference between the response after the sensitization and before the sensitization. So for index of zero, so there's no sensitization those there's no change in, in the sensitivity for index, more than zero, there is sensitization for index less than zero is adaptation. So here's a summary. So for the off response, we can see both doors on the ventral cells show sense that sensitized off response. And but for the on response, we only see sensitization in the dorsal cells but not in the ventral cells, and also only in the dorsal cells we see the appearance of a sustained component following sensitization. And just to summarize again, so in dorsal ganglion DS ganglion cells, both on and off responses sensitized, and there is a sustained component. And but in the ventral cells only off responses sensitize and there's no sustained component. And here is just example showing that this phenomenon is transient and reversible. So here we we did the testing spot first and then we did induction stimulus. And then we test again to to and to see the sensitize response. And then we stopped the protocol for a while and then the cell will quickly go back to the baseline for it. And then we can induce sensitization again in the second trial, and then wait a bit and then the cells for it go back to baseline, and then we can induce it again and again. And then we also look at the timing of this phenomenon. And so here we use the testing spot with increasing intervals after induction to see what's the maximum interval that we can still see the sanitize response. So, we see that without continuous testing, the sanitize response after induction stimulus will return to the baseline in in five to 15 seconds. And then the after induction stimulus if we keep showing the testing spot at, for example, the five second intervals, then the sense that the sensitization will remain as long as a testing spot is always delivered. So, so next with Lindsay looked into the synaptic mechanisms underlying the sensitization. She want to know whether the increased sensitivity after the induction stimulus is due to enhanced excitation or due to decrease inhibition. So for this she look at the both the sub threshold membrane depolarizations by looking at the PSPs. This is after the spiking was digitally removed from the trace. So we can you can see that compared to the before induction traces, the red trace which is after induction. For the dorsal cells, she see sustained depolarization in the baseline and then also more deprived response for the on and off. But for the ventral cells, she only see sustained in sanitized off response for the PSPs. And then she also recorded EPSCs, which is the gluten motorgic input from on and off bipolar cells. And she found that after sanitization, the dorsal cells have a sustained depolarized sustained increase in the gluten motorgic EPSCs and also enhanced on off transient response. But for the ventral EPSCs, she didn't see the sustained component consistent with the lack of sustained spiking. And she saw the sensitized light evoked response. So here's the quantification. Again, showing only dorsal cells have a sustained component. But, but, but it's not, which is not present in a ventral cells. So, so it seems that this sanitized response in a ganglion cells is due to enhanced excitation from the bipolar cell excitation. And then next she tested whether the sensitized response in the ganglion cells requires synaptic input, or can she just simply depolarized ganglion cells by electrical current injection to mimic the sanitization. So what she did is she first recorded DS ganglion cell activity during this induction stimulus. So this is current clamp recordings, showing that the memory potential response to the testing spot. And then during the induction stimulus, this is drifting gratings for 20 seconds. And then the following by followed by the testing spot. And then she clipped the recorded waveform during the induction stimulus. And then she used this waveform in the voltage clam experiment to directly patch onto a ganglion cell and, and, and activate the ganglion cell, according to the, to the activation activation pattern of the cell during the induction stimulus. This voltage clamped the cell to mimic the activity of the ganglion cell during induction stimulus. And then, for the rest of the time window, during the testing spot period, she hold the cell at minus 60 millivolts to record the EPSCs. So, so she can still measure the level of sensitization of the EPSC before and after this direct depolarization of the ganglion cell. So in this case, synaptic currents are blocked where there's no visual stimulation. So the ganglion cell is only activated by direct electrical stimulation. But in this case, she didn't see any sensitization of the EPSCs or the ganglion cell response, indicating that synaptic inputs is required for this phenomenon. So another question is about the sustained components that appears in the dorsal DS ganglion cells after induction. So she next asked whether the sustained component originated from on or off pathways of the bipolar cells. So what she did is she blocked the on pathway with AP4. And then she do the induction protocol again. So first of all, in the presence of AP4, the on response is blocked, as expected. But she still sees sensitized response in the off response is still sensitized, and she still see the appearance of a sustained component, indicating that the sensitization of the off response as well as the sustained component originated from the off bipolar cell pathway. And she also recorded the PSPs in AP4, again, consistent with the spiking activity. She see that although the on response is abolished, she still see sustained component and after the, and also sustained on response in AP4. And similarly, the EPSC data is consistent with the off pathway origin of this sustained component and off response. And then she look at what are the signaling mechanisms that mediate this sensitization. So she tried various pharmacological agents to block different type of signaling. So she we didn't find the effect with gabazine, the gab allergic signaling is not involved. The muscandronic and nicotinic receptor signaling doesn't seem to be required for the sensitization. And the only drug that had an impact in the sensitization is the strengthening. And in the presence of strengthening, both the on and off spiking response is no longer sensitized. And we also see a significant reduction in the sustained component. And similarly, a similar effect is seen for the EPSCs. So strengthening blocks the sensitization of the off and the sustained EPSCs as well. And to summarize these results, I presented so far, we think that glycinergic signaling is required for the sensitization of the dorsal DS ganglion cells. We think before the induction stimulus, there is a glycinergic amcarine cell that modulates the glutamatergic inputs from the off bipolar cells to the off DS ganglion cells. After the induction stimulus, the hypothesis is that this glycinergic inhibition adapts after the induction stimulus, and this leads to the disinhibited release of glutamate from the off bipolar cells, so that the off DS ganglion cells receive more glutamate from the off bipolar cells. And there's one thing that we also noted is that in the presence of strengthening so if we block glycinergic signaling with the strengthening didn't prevent the sanitization of the on EPSCs the glutamatergic response, but it does abolish the spiking response. So this suggests that the sanitized on EPSC is not sufficient to cause this the the sanitized spiking response for the on pathway. So, this is just the recapitulation so the strengthening doesn't affect on EPSC, but abolish the sanitization of the on spiking. And we also noted that the strictly impairs the sustained component, and we know that the sustained component originated is originated from the off pathway. So, our hypothesis is that the sustained component from the off pathway plays a role in the sense that in the stabilization of the on response. So the hypothesis is that after induction stimulus, the off bipolar cell release more glutamate both transiently to the off response, but also have a tonic increase to cause the sustained depolarization. And of course the off bipolar cells have the sustained depolarization depolarization in the off dendritic layer, but we hypothesized that this sustained depolarization from the off dendritic layer can spread to the on dendritic layer to boost the excitability of the on dendrites, so that the next incoming on EPSC is more likely to generate spiking response. So, so there's basically the electronic spread of depolarization from the off dendritic layer to the on dendritic layer. And, and, and so and one way we think about the by stratified on off DS dendrites is that there are two dendritic layers, and so the spread can occur, you know, from one layer to the next layer through the soma. But another interest and this, but another feature of the off DS dendrites morphology came to our attention, because we noticed, there are many crossover dendrites between the on off dendritic layers. So here is an example of a reconstruction of a DS of DS ganglion cells recorded. And you can see that the off and on dendritic layers are shown. But also, many, there are many dendrites that across between the layers. And this phenomenon has been observed very early on in the rabbit and in the mouse. And, but here we start to wonder whether those crossover dendrites actually plays a role in facilitating the electronic spread of depolarization from off to the on layers. So, a former undergrad in the lab start to analyze the this morphological feature of dendritic crossovers. So in this example, this is the same cell, and this is the off dendritic layer this is the on dendritic layer, and the red colors dendrites off layer dendrites but originates from on layer. So these are coming from a diving dendrites from the on layer and then further elaborate in the off. And then in the on layer those yellow color dendrites originates from the off layer. And so again, this is the side view. And then here is the quantification. So there are different configurations. There are layer dendrites in the off layer that come from on layer and there are on dendrites that come from off layer. And then there are some from on to off and then back to on and the other way around, but the most common crossover configuration is off dendrites originated from the on layer. And this is quite and this crossover is very consistent across all cells in the DR before in the posterior tuned cells we reconstructed and about 30% of total dendrites originates from the opposite layer. And also we also did a show analysis to analyze where are the diving points distributed. Concentrically, and we found that most of the crossover points occurs in the, in the distal half of the dendritic tree. And, and, and this seems to facilitate the spread, the signal propagation across the dendric layers and bypassing the soma. And then our collaborator Rob used a biophysically realistic model of the cell. And so she, he used the same cell that that I show in the previous slide. And, and, and simulated the electronic spread when we inject current or a synaptic inputs occurred in off layer and how effective is spread into the on layer with a crossover dendrite. And so the on layer is shown in the in the off layer is in top on layer in the bottom. And then the current injection is in here in the in in this mark in the off layer. And because of the crossover dendrite the on layer that the vertically aligned to the injection site in the off have a prominent increase in in memory potential depolarization to And Rob also simulated with with known active dendric mechanisms in the model and he found that this crossover events is local. And this, so this is because the depolarization in the off layer is more effective in recruiting the on layer active mechanisms when the electronic distance between the off layers are shortest. So it's so this effect depends on distance. So the model predicts that if a sense that sensitization is locally induced in one side of the dendrite in the off layer, then the on response is more is more effectively potentiated in on layer that are vertically aligned. So, meaning that the on the same side on the same side of the of the dendrites. But if the induction or if the sentence if the sensitization is induced in the opposite side of the dendrite between the on off layers, then the off and so this this is less effective in boosting the on response in on layer. So again, this simulation is is mimicking the induction protocol we use. So in this case, local test spot is used, and then a local induction stimulus is used to induce sentence sensitization. In one side of the DS ganglion cell dendrites, followed by on on stimulus, either in the same side side or in the opposite side of the dendritic tree. And so the model predicts that when the testing spot for the on off and on the same side, we should see a more pronounced sensitization of the on response due to the crossover signal from the off dendritic layer. So we tested this simulation results with experiments. So what Lindsay did is she showed a local testing spot followed by induction stimulus a local drifting greetings but especially confined to one side of the DS ganglion cell dendrites. And then she after induction stimulus she test the spot response using a test spot either in the same side of the induction stimulus or in the opposite side of the induction stimulus. Then she found that she see more pronounced sensitization of the on response. If the testing spot is on the same side of the induction stimulus compared to the spot that was delivered to the opposite of the induction stimulus. So this is consistent with the modeling showing that this crossover. So the, the, the spread of the sustained depolarization from the off dendrite is, is distance dependent. So it's more effective in recruiting active dendritic mechanisms and to bring the on dendrites above the spiking threshold at the shorter electronic distance. So, to summarize, we found that in the, in the dorsal retina, we see that after a short period of induction stimulus, the off response of the ganglion cell is transiently enhanced. And in the dorsal retina, we also see the appearance of a sustained depolarization originating from the off layer. And this sustained depolarization in the off dendrites of the ganglion cell spread into the on layer of the dendrites. And then this and through the both the soma and the crossover dendrites, and this leads to the increased excitability of the on dendritic layer, and which analyzed the enhanced on response. And in the ventral retina, we didn't see sustained depolarization, we also didn't see send us sensitization of the on response. We weren't suggesting that the sensitized on response in the dorsal cells comes from the sustained depolarization from the off dendritic layer. And, and since this mechanism originates from the off bipolar cell pathway, we wondered whether this is this can this this mechanism also impact other retinal ganglion cell cell types. We recorded targeted several other cell types, mostly the alpha cells with big soma size. So she recorded a subset of transient on alphas sustained on alphas transient off and sustained off, and she only found the similar pattern of sensitization in the sustained off alpha cells. So again, in the dorsal, in the in the dorsal retina, she found that at least a subset of sustained off alpha cells also shows sanitization after the same induction protocol. But in the ventral sustained alpha cells she didn't see sensitization and she doesn't see the sustained firing. And also she looked into development. So when is this sensitization first appear in development. So she found that around eye opening or right before eye opening about P 12 and P 13. She didn't see sensitization in the dorsal cells. So, if anything, she she see a moderate adaptation. So it seems that this sensitive sensitization in in is developed after eye opening. And then she checked whether she can prevent the development of this phenomenon with dark rearing. So, when she dark reared animals from pH all the way to P 13. She still see sanitize response from those ganglion cells. So indicating that the sensitization of the ganglion cell is developed after eye opening but is independent of. dark rearing so dark rearing doesn't affect the development. So here are just a quantification so early in right around eye opening. We didn't see any sensitization if anything there might be a tendency to adapt for the on response. And, but we see sensitize response in the adult even in the conditions of dark rearing. So to summarize this part, we think so we're, we particularly look at a phenomenon where the off DS ganglion cells show adaptive changes in game to prior visual stimulation. So this, this commutation is implemented by glycinogenic signaling in off bipolar cell pathway, but we only see that in a dorsal retina. And then these glycinogenic mechanism implement short term disinhibition of bipolar cell excitation, and that analyze this adaptive change. So our phenomenon add to the increasing findings of sensitize sensitization in in the retina and across species. So previously, there are multiple studies looking at contrast, sorry, not contrast, both contrast adaptation and sensitization. So Steve bucket's group have shown that if you show high contrast a period of high contrast stimuli, and followed by low contrast, a subset of ganglion cells in a salamander mouse and rabbit ganglion, ganglion cells show adapting response after the high contrast. And also, Leon like not those group showed in that in the zebrafish. There are subpopulations of retina neurons that show sensitizing response after high contrast. And recently, Manukin's group have shown that in the on measured ganglion cells when presented with a low temporal frequency greetings also show sensitize response. So, you know, these to get so in in our protocol we didn't use. So the induction stimulus is with the same contrast, but we had, I think the temporal kinetics is different between the testing spot and the induction stimulus. But regardless of the specific induction protocols for sensitization. I think our studies points to a common theme of sensitization mechanism, which is the adaptation of inhibitory inputs to the bipolar cells. So it's a dis inhibition of the excitation that underlies those sensitization phenomena. So just to summarize. So, with so these recent examples using a bit more complex visual stimuli, really highlights the context contextual modulation of visual circuitry even at the very early stage in the retina. So, depending on the visual context, or depending on the visual stimulus, the commutation of the, of the retina neurons can be flexible it can, there's flexibility in what type of information, the ganglion cell can encode. And there's also flexible algorithms. So, that, for example, the, the same circuit motif can function very with different algorithms, algorithms, depending on the visual stimulus. And also different visual stimuli will recruit different subset of neuronal substrates to implement those context dependent processing. And that I like to thank my lab. So the three projects are done by Chris, Jen and Lindsay in the lab, and also a former student Hector has done really interesting very nice work at the dendritic level but I didn't have time to talk about today. And the rest of the lab, and I like to also thank my wonderful collaborators, Rob has been collaborating with us on multiple projects on the modeling. And Stephanie and Albert Chen has been helping us with theoretical analysis and and look at the coding of the ganglion cell response and David has been helping us to mining the connectomic data to support, to, to, to provide insight into the implementation of those of those computations. And I also like to thank my funding source. Thanks. Okay. Thanks a lot. Really, really interesting talk. Okay, thanks. A bunch of data, really interested findings. And yes, so we can proceed with the questions that are coming from, from the chat. And Anna, and I bless it is, is saying and asking if these are very interesting results. I'm curious about the intensities and colors of the of the visual stimuli you are using. Are you using a UV wavelength stimulus in the ventral retina. That's the first question. Yeah, that's a great question. So, initially, when we look at the dorsal ventral difference, we see that the ventral cells doesn't sanitize for for the sustained component and for the off response. And of course, the, the dorsal and ventral retina, one prominent difference is the is the their absence right so the ventral cells have a very high density of s options, but the dorsal cells have much lower as option content. And so our old led we use organic old led led that so our visual stimulus cannot activate s options. So that so to and so we then did more experiment to see whether the lack of sensitization in the ventral cells is due to our stimulus right if we cannot activate as option and actually maybe we couldn't. It's not the red stimulus to activate to sanitize the dorsal cells. So then we use the UV led to perform the same induction protocol in the ventral retina. The UV led is more effective in activating s options in a venture retina, but we still couldn't induce sensitization using a UV led. So, so I think, so there's definitely, you know, the dorsal and ventral retina already taking inputs because of the photoreceptor difference. But it seems that even when we sufficiently activate s options, we still cannot induce the same sustained depolarization and the on response in the ventral retina so our hypothesis is that something in the IPL in the off bipolar cell circuitry in the inner prex form layer is responsible for the dorsal ventral difference so they might be differentially modulated by glass energetic signaling for example. So, I have another question from Tomomi Ichinos, which says, I missed the reason why the sensitization happens only in the dorsal retina. Is the glass energetic cells differentially expressed between the two sides? Yeah, yes, great question. So whether, yeah, because Ichinos lab has has a very elegant study on the on the cholinergic signaling at the bipolar cells. We, we, we don't know for sure, we only tried DH beta E, which is a nicotinic blocker that block alpha, alpha two, three, I think doesn't block alpha seven very well. Nicotinic, nicotinic receptors. So DH beta E cannot prevent sensitization. So it's not those nicotinic receptors. We also use atropine, which is the muscarinic receptors, and they also failed to prevent sensitization so it's not muscarinic, but we don't know whether it's alpha seven or it's other, yeah, cholinergic signaling could be involved. Okay, so yes, moving forward. Yes, I, we, at the moment, we don't have more questions from the chat. And remember, if you're interested in discussing more in details, I already posted the zoom link so we can go to a private room. I was, yeah, I, I also had the same question about the topography of the distribution of glycinergic amarkating cells. But yes, I was wondering if, if you have done any morphological study of the distribution of these so what's the question was in the same way as Ichinosis question so, but you already answered it. And so this is just by curiosity, you, you apply this, this is stimuli which generate the sensitization. Have you, have you seen by chance, any stimuli which generate the opposite effect, the depression in terms of the sensitization over your studies, or if you detected the opposite effect in any population of direction selectivity RGC. Not really. Oh, yes, that's a great question. So we, whether question is whether we can induce adaptation. Yes. With any other stimuli. Yeah, play with the size or, you know, duration of the preceding stimuli. So we haven't, we haven't using, so, you know, if we show this very brief period of, like, gradings or moving motion stimulus, we always see sensitization. But, but, but I think the, the contextual modulation of the game of those of the S ganglion cells is is definitely there's a more rich phenomenon behind it. So for example, previously, Michael reference group have shown that if you show a slightly quite a different type of induction stimulus, she can, she can flip the tuning curve to the opposite direction. So, you know, the tuning can be flipped to the now direction. And that involves a different form of adaptation in the inner circuitry in the retina circuitry. So there are definitely, I think they're definitely different contextual effects, depending on the stimulus that we show. But in terms of our induction protocol, it's the sensation is robust. And it doesn't happen in all other in other ganglion in many other ganglion cell types for the four type of alpha cells we tested. The other three doesn't show any change in this protocol, but the sustained alpha shows the sensation. Yeah. Okay, thanks for, for your password. We have more questions from Leon. So he's asking on the mechanism of sustained response after induction. Is this simply due to reduce inhibition of off bipolar cell synapse. This will imply that baseline activity that the before the induction is set by by high levels of inhibition. Yeah, so, so the question is before induction because we don't see sustained spiking. We also don't see sustained EPA sustained glutamatergic inputs. So when we record the baseline activity of the ganglion cells is very quiet. So the spontaneous found rates in our prep is very low. We barely see spontaneous spiking before we show so the visual stimulus. So I think that at the baseline level maybe the excitation is already the tonic excitation is very is low in at the level that low enough to to minimize a spontaneous spiking. And then after induction, the bipolar cell release more glutamate so there's more glutamatergic EPSCs, but actually, I didn't show it in the slides, but we also show that we also see an increased tonic IPSEs. So the inhibitory inputs to the ganglion cells through immigrant cells is also enhanced. But overall, that increase inhibition is not sufficient to counteract the increased depolarization from excitation. So we see this sustained spiking depolarization and spiking after induction. And we think that because of is the because it's the sensitization of the bipolar cell excitation, whatever the immigrant cell target of those bipolar cells are also receiving more excitation. That's why those cells will also release more IP a Gabba onto the ganglion cells. So there's a concomitant enhancement of EPSC and IPSE after induction. Okay. Thank you so much. So, yeah, so, and so I would like to say, again, thanks for accepting our invitation to the SAS exhibition series. And now, if you're interested, if the audience is interested, they can join the the private zoom room in which we will be discussing more in detail, or the results you have shown now. So thanks again. And we will keep the link online for a few minutes. So we can keep talking about your search and future directions if you want. Okay, thank you so much, Jose. Okay, no problem. Thanks a lot, everyone. All right. So people is start to join started to join. I'm not blessed. Just joined the conversation. So, yeah, I, well, when we are offline from the YouTube. Yeah, yeah, I will, I will end this now. So the link is already posted. So I will start, I will end the stream right now. Hi, how are you? Hi, are you