 So, welcome everybody to our Sussex Vision seminar today. I have the pleasure of introducing you our next speaker, Professor Feliz Dan from the University of California in San Francisco. Well, for the new people, let me introduce myself. I am Jose Moya Diaz. I am a postdoctoral researcher of the laboratory of Leon Lagnado and this term I will be part of the committee which is organizing this Sussex Vision series. So as I say, our speaker today is Professor Feliz Dan from the University of California in San Francisco and I will introduce briefly her academic background. So Professor Feliz Dan got her bachelor in neuroscience at the University of Brown under the supervision of Professor David Berson. Then she moved to the University of Washington for a career on her PhD in the program of neurobiology and behavior under the supervision of Professor Frederick. And later she made two postdoctoral, two postdocs. So the first postdoctoral research was carried on in the, in the laboratory of Professor Mark Stopfer when she established methods for in vivo patch plan recording of olfactory neurons in the locus. And then she moved as a postdoctoral fellow in the lab of Professor Rachel Wong in at the University of Washington. During her postdoc with Professor Wong, she demonstrated that different types of people are cells established connections with photoreceptors according to different route strategies and over barring time scales and she also developed methods for stable live live imaging of in vitro retina for over a day. So I would like to say thanks Feliz for accepting our invitation. It is a pleasure for us. And yes, we are expecting for for your great. Thanks. Can you see my slides. Yes, yes. Thank you. And doctors. Jose Moia deans and to Leanne like not over the invitation to the Sussex vision talks. I also want to acknowledge everyone who has made worldwide neuro possible. This forum is provided respite from this pandemic, and I hope that it outlasts the pandemic. It is an honor to be part of the seminar series and to share our labs work with people around the world. I'd like to tell you about work that was spearheaded by my first PhD student Rachel care. She then passed the baton on to JL do young Lee, who was a brave and resilient postdoc. And as you'll see later, there's quite a bit of extensive manual analysis on our morphology, and that was done by this great team arena de la Horta, Simon Pan, a tree coach a Allen Chen chat Chad Santo Thomas to need to know, and So together, we asked the question, if during the course of the day, our visual system can easily operate over light levels ranging from nine to 12 orders of magnitude. Could the visual system use some of these same adaptational mechanisms to withstand photoreceptor loss. One way to pose the question is, can the visual system even distinguish between changes in light level and changes in photoreceptor number at early stages of input loss. I will show you today that the retina has mechanisms to deal with photoreceptor loss that can be likened to adaptational mechanisms, and that keep it resilient. Why is this important. It's because we're all enduring some degree of photoreceptor loss. Throughout our lives, we lose photoreceptors due to injury disease and normal aging. And in this study summarized an annual loss of point two to point 4% photoreceptors in humans without any apparent retinal disease. The visual system has to deal with this photoreceptor loss, and it seems to do so surprisingly well. So in the seminal study from Ratnam and colleagues and later foot and colleagues, the colleagues including Doe Carol Jackie Duncan and Austin Royda. They combined psychophysical examination of visual sensitivity and acuity with live imaging of cones in patients with inherited retinal degenerations. What they found was that patients can lose up to 40 to 60% of their cones before their vision degrades the level of blur in this chart, which is 20 40 on the cell and acuity chart, not too bad. Now this suggests that either diagnostics are insensitive to photoreceptor loss below a threshold. Or that the visual system is somehow resilient to photoreceptor loss up to a threshold. Previous work from visual cortex has shown evidence before and against changes within the cortex that might be able to handle the changes in function at the photoreceptor level. And we want to understand the retinas contribution to this potential resilience. So in the context to photoreceptor death, traditionally, photoreceptor dysfunction has been caused by spontaneous or directed gene mutations. And studies have found evidence for gendered expounding mislocalization of glutamate receptors and aberrant electrical activity. In many of these models, these changes are occurring during the period of retinal development. And so any plasticity that's triggered by photoreceptor death is confounded by developmental plasticity. In many types of human retinal disease, photoreceptor dysfunction occurs in a mature retina. In the models, lasers have been used to ablate photoreceptors indiscriminately. In work from corn, Korean beer, Daniel Planker and Sasha share, they demonstrate repair of scatomas and ganglion cell receptive fields. And the growth and rewiring specific types of bipolar soldendrites. So we're also going to use a method of ablating specific photoreceptor types with temporal control to understand how partial photoreceptor loss is handled by the remaining retinal circuit. And though we're not modeling any specific retinal disease, we can still gain understanding about the effects of photoreceptor loss on the retina. This is relevant because ultimately the success of gene therapies rescue of photoreceptors relies on the functional integrity of the rest of the retinal circuit. I'm going to use a simple framework to define the possible outcomes. We can think about under normal conditions. We present a range of light levels to the retina and a ganglion cell spikes with a range of spikes. If half the photoreceptors are gone, the remaining retinal circuit could propagate this input loss. And this might show up as a decrease in response amplitudes and or a rightward shift in this intensity response relationship, in which more light is necessary to generate the same spike output. The possibility is that input loss to somehow mitigated within the retinal circuit, in which case the responses might be better than predicted by a linear propagation of input loss. We want to know which of these applies in the mature retina and where to do this, we killed half the cones in mouse retina. And then we use structural and functional analyses to measure the input and output of the bipolar cells, as well as the input and output of ganglion cells. And at each of these places will determine whether the input loss is propagated or mitigated. And as I go through the data, a purple box will highlight the cell or synapse that we're examining. To generate mature mice with partial cone loss, we selectively express the simian diphtheria toxin receptor under the cone opsin promoter. And this receptor remains inactive until we inject diphtheria toxin at postnatal day 30 after retinal development. And then we inspect the retina at multiple intervals following the diphtheria toxin injection. What I'm going to show you today is inappreciably different across these intervals. I'm going to combine them. We're going to take a look at the cone death first. We look at the pedicles of the cones in control retina. This is an on-foss view and then following diphtheria toxin injection. You can quantify the density of cones in this histogram in control retina in black and following diphtheria toxin in red. And this manipulation approximately halves the total population of cones. To understand the impact of this partial cone loss on the first synapse, we focused on the dendrites of the type 6 cone bipolar in the mouse retina because it's sparsely labeled in this transgenic line. It was generated in Rachel Wong's lab. And because its dendrites are highly stereotyped. So it's easy to tell when they've changed. The bipolar cell is in yellow and it's masked so you can see the dendrites more clearly down here. On average, this bipolar cell type contacts four cones, one, two, three, four. Following the phtheria toxin injection, this particular bipolar cell is only contacting a single cone. You can quantify how many cones each of these type 6 bipolar cells contacts, which on average is four in the control and reduces to two in following diphtheria toxin. So the amount of cones that are lost for each bipolar cell is consistent with a degree of cone loss. Unlike the type 6 bipolar cells, there's two specific types of bipolar cells that can recover the cone inputs in a mature retina. Known by the phtheria toxin center groups, the XBC can recover their cone contacts. And from the beer, polinka and share group that the S cone bipolar can also reestablish contacts. For these type 6 bipolar cells, we wanted to know whether the dendrites remained functional. So what's the glutamate receptor distribution on these remaining dendrites. And that relies on the metabotropic glutamate receptor 6, which we can label with an antibody generously given to us from Katherine Morgan's. And that's shown in cyan. And we can quantify the location of the peak and glue our six is a function of distance from the soma to the dendritic tip. And that's shown in this histogram. And that tends to peak at the last quarter of the dendrite because that's where the cone contacts are being made. So that's for control retina. In a retina and during cone loss, we can separate the dendrites into those that still contact cones like this one. And if you quantify the location of the angler stick six, it still peaks at the last quarter. So those dendrites seem to remain intact with their glutamate receptors. The dendrites that no longer contact cones. The mglour six distribution in the dark red is randomly distributed, suggesting non functional synapses. Together, these results show the stability of the first synapse of the visual system is dependent on cone contacts, and is also regulated independently at each dendrite. From the morphology when we conclude, we interpret that the cone loss is being propagated at the first synapse both by the bipolar cell dendrite, the type six dendrites, which lose their cone contacts, and by the glutamate receptors which only disappear from the dendrites that have lost their cone contacts. Next, we're going to look at the second synapse, which involves the external output of the same type six bipolar cells. We can label for the pre synaptic release sites with the ribbon protein, the C terminal binding protein to CTVB2, which is not changed following the period toxin. So even though the bipolar cell dendrites have lost half their input sites, they still retain all of their output sites, as if the bipolar cell could somehow weather the photoreceptor loss. We also quantified the density of post synaptic sites on the major partner of these type six bipolar cells, which are the alpha onsistine gain cells. And we can quantify the density of PSD95 across the dendrite, and also we find no significant difference between control and DTR conditions. This shows that both the pre and post synaptic machinery remains intact. So the second synapse is stable morphologically. You can put the morphology in the perhaps mitigated category. And to understand the functional signal conveyed at the synapse, we next measured how the ganglion cells respond to light. So we recorded from these alpha onsistine ganglion cells. But if we look at an onfosive of the dendrites, each ganglion cell receives input from an array of cones that feed into its receptive field. And we also know that this receptive field is organized in a center surround organization. And so previously I had mentioned thinking about the responses as a function of light intensity. We can also think about the responses as a function of space. In this particular ganglion cell, there's a region over which the cell prefers increments in light, and there's a region where it prefers decrements. And the sum of the two gives us this difference of Gaussians. This is what the receptive field looks like when all the cones are present. The question is, what is the receptive field look like when only half the cones are there. The one possibility is that fewer cone inputs results in narrower centers and narrow surrounds. And we might interpret this as a propagation of co loss. Another possibility is that the retina interprets the loss of photoreceptors as less light. And we know from experimental work from Enrath Kugel and theoretical work from Attic and Redlich that at lower light levels receptive fields tend to expand their centers and lose their surrounds. So yet another possibility is that the ganglion cell changes its receptive field. However, it can to make you best use of its remaining cone inputs. This case might predict that the receptive field widen its center and widen its surround, whichever is possible. And we might interpret the latter two as mitigation of the cone loss. So to measure the England cells receptive field, we isolated dorsal nasal retina in the dark, and then we use infrared illumination to target and record from these large alpha unsustain ganglion cells. And we record the voltage in response to a stimulus that's delivered from a broadband projector directed through the pathway of the condenser. And the projector background adapts down the rods and more effectively stimulates the M option. The stimulus is comprised of bars shown here, in which the intensity at each location and at each time point is drawn from a Gaussian distribution. We then use the ganglion cells voltage response to extract a spatial temporal filter. So here you see space and time, and the response polarity is shown in color. There's a white region over which that is that is outside the cells receptive field. There's a red region that comprises the cells surround and a blue region that comprises the cell center. So you can take a representative slice in time and look at the profile of the spatial receptive field, fit it with the difference of Gaussians and measure the one standard deviation center width, which is decreased falling partial loss and it's one standard deviation surround width, which is increased falling column loss. So what does this mean, if we return to our cartoon. The receptive fields center that were accepted field centers became narrower and their surrounds became wider. And the narrower center we said might be consistent with propagation but the wider surround is more consistent with mitigation. To explore this further, we asked whether the narrower center and the wider surround might be implemented by a control retinal circuit, if only half its inputs were stimulated, rather than eliminated. So we devised a half stimulation experiment to distinguish between mechanisms that allow the retinal circuit to adapt versus those that are triggered by photoreceptor death. And I'll start by illustrating the experimental design in this cartoon. Imagine in the control case, all the cones are present, the retinal circuit is stable, meaning it hasn't changed. And when we show our stimulus 100% of the cones respond to it. In the cone DTR case, there may be changes in the retinal circuit, we're not certain of that. And when we show our stimulus. Now only 50% of the cones respond because the other half are gone. So there are two things that are different in the cone DTR case, both the possible state of the retinal circuit, as well as the number of cones that are being stimulated. To separate these two things, we devised a new case, which is half stimulation of the control retina, in which we only present half of the stimulus the other half of the bars are fixed at a mean. And so in this case, the retinal circuit is stable. And when we show our stimulus only 50% of the cones, see the dynamic stimulus and can respond to it. But if the receptive field looks like the cone DTR case, then we know these receptive field changes can be caused by half stimulation of the cones, since nothing else is changed about the retinal circuit. But if the receptor field looks like the control case, then we know that 50% cone stimulation alone is insufficient to cause the receptive field changes observed. Which is it. Here I'm showing you the spatial temple receptive field of the same ganglion cell under full stimulation and half stimulation. And these white stripes represent the locations where the bars are held constant. The profile the receptive field again fit it with the difference of Gaussians. And then because the data are acquired in the same cell we can look at the center with under full stimulation versus partial stimulation. And the data fall live along this line of slope unity, showing that the center with hasn't changed. The same is true of the surround with, with the data lie around this line of slope unity. Neither the center nor the surround change significantly after half stimulation, and this is different from what we saw in cone DTR retina. If we return to our schematic. Since the receptive field didn't change following half stimulation, we interpret that the changes after cone loss must be caused by mechanisms within the retinal circuit. So we determined that the mechanisms in the retinal circuit change receptive fields after cone loss. But whether these receptive fields can be considered propagation or mitigation remains ambiguous. The decrease in center was consistent with propagation, while the increase in surround was consistent with a mitigation. We also identified the source of these changes to help categorize the effect. Here JL stepped in to measure the excitatory input currents between bipolar cells to the gangrene cell and inhibitory input currents from amicron cells directly onto the gangrene cell. And here I'm showing you the metric for the center one center deviation center width. In this citation, the center with under control conditions versus DTR show that the data fall below the line of sub unity showing that there's a shrinking of the excitatory center in the surround. There's an increase in the excitatory surround for inhibitory inputs. There was no change in the center width and a significant significant increase in the surround with an inhibition. So we see a significant change in both the center width and the surround widths. And to test whether the control retina could exhibit the same changes, again JL mimic the effects of partial cone loss by blinking every other bar and presenting the stimulus to a control retina. The data are shown on the right axis for partial stimulation versus full stimulation again on the bottom axis. And for all the measurements center and surround with for excitation and center and surround with for inhibition, the data fall on the line of sub unity. So like the voltage recordings, partial stimulation could not evoke the same receptive field changes. Now, because the DTR and the partial stimulation both have their own respective control conditions. I'm going to show you the data again, but now is a ratio between DTR and its control condition and partial stimulation and its respective control condition. And that's shown here for each of the metrics. I'm now showing you the ratio, so that if the data fall along the line of sub unity, the change within DTR and partial are similar. And anything off of the line might tell us that there's a distinct mechanism between DTR and partial stimulation. For excitation DTR shows a greater change, a greater decrease in the center width. Also an excitation there's a expansion of the surround compared to partial stimulation. For inhibition, there's no significant change. And for inhibition surround there's a significant increase in the surround only in the DTR condition. So the loss of cones has evoked mechanisms that trigger narrowing of the center width and expansion of the surround widths. And such changes can't be recapitulated by partial stimulation in the control retina. I'm going to focus in on this finding here, we'll see that both excitation and inhibition have increased surrounds after partial cone loss. And I want to show you other signatures of changes within the circuit specifically within inhibitory circuits that could account for these changes. So up to this point I've been showing you the spatial linear filter in the spatial temporal linear filter, but we can also examine the temporal filter by slicing across the other dimension. Here are the linear temporal filters in response to the bar noise under the four conditions, and then we can take the integral of these temporal filters and plot the integration time under DTR versus control, which is not significantly changed, as well as partial versus full stimulation also not significantly changed. And as I showed you before, the summary is the ratio between the DTR versus control and partial versus its control, and the data lie around this line of soap unity. So no change in the integration time of the temporal filters when we measure the voltage outputs of these ganglion cells. This is true for the excitatory input currents coming from the bipolar cells, but we see no significant difference in the DTR and no significant difference in partial stimulation to the data fall along this line of soap unity. But now when we look at inhibition, we see that integration time is significantly larger in the DTR retina compared to its control, but the partial stimulation doesn't evoke a prolongation of the integration time. So in the ratio metric, the data sit up here. So there's a much longer integration time for DTR, and that wasn't true in the partial stimulation. So now in addition to the increase spatial integration of inhibition that I showed you previously, we have another piece of evidence for increased temporal integration of inhibition. So while the spatial and temporal integration has increased, what I haven't shown you is what's happening to the nonlinear component of this model. So in our linear nonlinear model that maps the stimulus to the cells response, I've been showing you the linear component and ignoring the nonlinear component. And now I'll show you that which takes a linearly filtered version of the stimulus and plots it against the cells actual response. You can see some example cells for control from control written on black and from partial commas in red, and the same over here for full simulation versus partial stimulation. We can then fit each of these relation input output relationships with the line and take the slope of that line as a measure of gain that tells us how stimulus translates to output. And we see a significant drop in the game, following colon loss. We also see a significant drop falling partial stimulation in control retina. And the ratio of these measures shows us that the decrease in gain for DTR exceeds that for partial stimulation. So this is for the voltage output. And the same is also true for excitation where the degree of gain decrease is greater for DTR is compared to partial stimulation and the data point that's down here in the ratio metric. And then for innovation we see something different again, which is that the gain in DTR retina compared to control is not appreciably different. The gain for partial stimulation is lower than falling than for full stimulation. And that's where the data live up here. So somehow the inhibitory pathways or input to these inhibitory pathways are being spared or they're recovering from partial colon loss. And we speculate this could be a result of preserved or recovered excitatory inputs to specific inhibitory pathways and or direct up regulation of inhibitory inputs. I'll show you evidence for the second, but it certainly doesn't rule out the first mechanism where excitation along the way is somehow preserved or bolstered. So this cartoon shows all the possible sites of up regulation in inhibitory pathways, which includes horizontal cell feedback to cones, feedback inhibition between an amicron cell and a bipolar cell. And both of these could account for the changes we saw in excitatory inputs to these ganglion cells because these are presynaptic. There will also be changes in the direct inhibition from amicron cells to ganglion cells and serial inhibition between an amicron cell to an amicron cell. And both of these could account for the changes we observed in an inhibitory currents to the ganglion cell. I'm going to start with the simplest, which is amicron cell and potentially bipolar cell input to the ganglion cell. I'm showing that input currents in response to the bar stimulus. And now I'm going to show you spontaneous release events from bipolar cells when JL measures excitatory currents, and from amicron cells when JL measures inhibitory currents. She can then isolate the single events and quantify the amplitude of these minis, which is not changed for excitatory or for inhibitory minis. When she quantified the frequency of the events, she found that the miniature excitatory post-maptic currents decreased, and this is potentially consistent with the loss of cone inputs. And then she saw that the inhibitory minis actually increased in frequency. So this is consistent with more synapses being made between an amicron cell and a ganglion cell, or the activity from that inhibitory pathway being upregulated. When JL quantified the integration time of these minis, she found no difference in the excitatory mini, but a significant prolongation of the integration time of the inhibitory minis. So both of these findings are consistent with the light evoked increase in gain, as well as the increase in integration time. So while the loss of cone inputs is reflected in the decreased synaptic events between the bipolar cell and the ganglion cell, the increased frequency of events between amicrons and ganglion cells suggested a recovery from cone loss. And we also have some anatomical evidence that cooperates these functional findings. So JL quantified the sites of inhibitory receptors by staining for jeffrin in the axon terminals of type 6 bipolar cells. And she saw a significant increase in inhibitory receptors within bipolar cells. And this could also be consistent with a decrease in the mini excitatory potential post-maptic currents. We also looked at the jeffrin puncto within the dendrites of the alpha-oncestine ganglion cells, which is also increased following cone loss. And that's consistent with this increase in the inhibitory post-maptic currents. Together with the functional recordings, after cone loss, the inhibitory synapses have increased between an amicron cell and a ganglion cell. And from an amicron cell onto a ganglion cell, supporting the result that cone loss causes a recovery of inhibitory pathways to the ganglion cell. So if inhibitory inputs are recovering from, or if they're especially resilient to cone loss, then we wondered what the impact of these receptive field changes could potentially be on perception, because we still haven't answered this question. So Dale and Luca tested this by convolving a library of natural images by the spatial receptor fields measured directly from control retina versus DTR retina. And then they compared the similarity between the reference image and the convolved image using two metrics, either a basic assessment or a perceptually relevant metric. And they did this across a range of image sizes. With the basic metric that calculates the mean squared error between every pair of pixels, they found that the DTR retina has greater error. So the way the image is significantly different across image sizes. But with a perceptual metric that the cement jelly group modeled off the human visual system. What they found was that there was no appreciable difference between the images whether they were filtered by the control retina or the DTR retina filter. Some analysis suggests that the way the DTR spatial filter transforms the image could be perceptually similar to the way the control retina performs. And this perceptual metric accounts for luminance contrast and structure with an image, things that seem generally important to any visual system but we fully acknowledge this metric may not reflect the performance of the mouse visual system. And we want to explore this further. The field changes we observed with 50% colon loss may not cause any appreciable difference in perception. And as pure speculation perhaps in the humans visual acuity can remain constant until more than half the cones are dead because of mitigation in spatial and temple processing within the ganglion cell receptive fields. So let's go back to the original question about whether the retina could withstand photoreceptor loss through adipation mechanisms. From our partial stimulation experiments we concluded that the cone DTR retina was behaving distinctly from control retina, and that we couldn't recapitulate the receptive field changes with partial stimulation. We considered what the adipational state of the retina could be in partial versus partial stimulation versus partial cone loss, we came up with another potential explanation. In control retina, the ganglion cells receptive field narrows its center and widen its surround when it's light adapted. And here's a cartoon of two example cones that might be sitting under two different bars of our stimulus. And this cone sees a stimulus that goes from dark to light between two time points and so its glutamate release would change dynamically from high to low. And it's neighboring cone that sees the bar go from light to dark would change its glutamate release in the opposite direction. In our partial stimulation experiment, we presume that the cone on the left is behaving the same. But now the cone under on the right is sitting under a constant mean so it's glutamate release if it doesn't change over time remains constant. So for these two conditions, the average total glutamate release is probably similar. Now in the cone DTR case, let's assume that the surviving cones can behave similarly so the surviving cone signals the stimulus dynamically. The bladed cone is not contributing any glutamate, and if there's no adjustment in glutamate release, following cone death, and the total glutamate release by photoreceptors would be less in the DTR case. In other words, the retina might be in a more light adapted state. If this is true, then a different interpretation of our results is that partial cone loss causes the retina to be light adapted. The profile of the spatial receptive field, indeed our half stimulation, did not cause a change in the center or surround widths. But in the cone DTR retina, we saw a narrowing of the center and a widening of the surround. Both features are consistent with the light adapted retina. In other words, a retina that would have less glutamate release from photoreceptors overall. Our lab is continuing to explore this relationship between adaptation and photoreceptor loss. And I want to close by summarizing what I showed you today. I told you about a model for specific cone loss in the mature retina. Half cone loss caused a commensurate loss in cone contacts on the dendrites of type 6 bipolar cells. And the glutamate receptors were lost only at sites where cones were lost while the MGR6 was retained at dendrites that still had cone contacts. At the axon terminals, the same bipolar cells, but the second synapse showed structural preservation of excitatory synapses. The main partner of these bipolar cells, the on-alpha sustained ganglion cells, preserved their spatial receptive fields with an expansion of the surrounds and the shrinking of the centers. And the expansion of the surround was apparent both in the excitatory and inhibitory inputs to these ganglion cells with a larger effect in the inhibitory pathway. We found an increase in the spatial and temporal processing of inhibitory inputs to the ganglion cell, including an upregulation of synapses both at the bipolar cell and at the level of the ganglion cell. And then I showed you how these modifications in the spatial receptive field result in inappreciable differences in natural images filtered with these receptive fields, only when we used a perceptually based metric. And we speculate this is because the way the receptive field changed is consistent with receptive field changes that would be observed in a light adaptive retina. In other words, this is a relatively minor adjustment in that adjustment to the visual system that is well familiar with handling these changes in receptive field during normal adaptation. I want to emphasize that this work was accomplished by a great team. All of the contributing members are written in italics. We also obtained reagents from Rachel Wong and Katherine Morgan's. And we want to acknowledge our lab manager, Jeremiah John, who keeps everything running. And hopefully one day I can tell you about work from the current graduate student Scott Harris. Our current lab members are these three involved. I want to thank the generous funding sources that our lab has been fortunate to have. I appreciate all of you for your presence and your interest. And I am eager to hear any of your suggestions and questions. Okay, Feliz. Thank you very much. It was a really nice talk. Thanks for sharing our last discoveries with us. And yes, so we have a, at the moment, we have a question in the chat from Leon. He's asking if this the impulse response in the center also get longer after con loss. If the, if the excitatory response gets longer. Yes, exactly. So in the excitatory, that's a good question in the excitatory inputs, the integration time does not get longer. But when we measure time to peak we did see an extension. We did see a prolongation of the time to peak. So for different temporal signatures, we did see an increase in the center temporal into temporal time to peak. Okay. All right. So, from Micheal Papa, he's asking if the you pull the tier samples across the observed time bones across the observed time points, two weeks, four weeks, 24 weeks after, you know, several timescales. That's right we pulled the sample so originally we did the analysis by categorizing according to interval. And we saw no differences across the intervals and so we pulled them. And for jails recent work, and jail has been looking at the different staining as a function of interval. And there is one metric that does seem to be dependent on time but we don't have a lot of samples. But the, there's this maybe a slight significant difference across time intervals with everything else that I showed you today was equivalent across intervals. So I have a question. So probably I missed something but I got a bit intrigued about your assumption about when you are losing the counts. It is resembling, it is resembling a live adapted effects in the processing so so again, would you explain me a bit more about. Sure. We also didn't think about it this way. Originally we thought that absence of coins would would be interpreted as less light because there's less inputs, there's lower signal to noise ratio that matches a low light level condition. And that's how we were originally thinking about it. And it's pure speculation that the retinas and a light adapted state, because we don't have the glutamate measurement that that seems possible to do. But the missing cones aren't contributing glutamate. And so overall there's less glutamate, and maybe the rest of the retinal circuit reads that as less light. And the spatial receptor field changes do seem to reflect a retina that is light adapted and that the center's narrower and the surround is wider. But what does it match is the kinetics. So, in the light adapted retina. The expectation is that the kinetics of the impulse response are set up to. In this case, we see kind of a combination where the receptive spatial receptor field reflects a light adapted retina, but then conversely, the, the kinetic features resemble a dark adapted retina there's greater temporal integration. In the DTR case. So, still, still remains to be determined. And then it's, it's just a suggestion that that might be happening. Okay. All right. So, at the moment, we don't have more questions in the chat. I have a second question, but this probably is very naive because I'm not very familiar with the methods you are using for ablating the cones. This method. So, over a certain time, time scale, it is possible to see a regeneration of the cause by using these methods. And if have you tested over certain time windows, if there is any recovery at the level of the circuitry connection and the processing. That's a great question. So you're asking about whether there's regeneration of cones over time. And if we were doing this in zebrafish. Yes, there would be but if we're doing this in mouse, we don't see a recovery in the columns. In terms, the other question was about when the injection is made and whether there's increased plasticity. And the so to and Christian center groups have done exactly that so they use the same system of diphtheria toxin receptor expression, and they induced us either in the developing retina or the mature retina. And they found that there is increased plasticity when they cause cone depth in the developing retina. And so three of the four cell types they looked at continue to and maybe I'll check myself on the numbers, but more, more, more bipolar cells are able to reestablish cone contacts when the cones are bladed early, but in the mature retina, only this XBC mouse bipolar cells able to reestablish cone contacts with the other ones that were previous the other bipolar cells that were previously plastic or no longer plastic at the older age. Okay, okay, I see. So Leon is asking, if are you thinking that the global glutamate levels are somehow send it to adjust the gain of inhibitory inputs. If so, what might the sensor be. No idea. I have no idea what the sensor might be. So we don't know, we can't really distinguish between the, the changes in glutamate release being sensed by a bipolar cell that's providing input to the amicron cell, or the amicron cell itself, which is also going to have glutamate receptors. We're starting to get at this issue a couple ways. One is the question of, if there is some excitatory pathway that might be preserved or upregulated following the diphtheria toxin injection so is there a bipolar cell with that's already sensing the glutamate at the photoreceptor that can tell the amicron cell. We're in a light adapted state now change however you need to change. And we don't see direct evidence for it. I can just tell you that when we look at a global measurement of the on bipolar on on comb bipolar mediated electro retina Graham, we see a decrease in the, in that component of the wave that is a greater decrement than we see in the cone photoreceptor mediated. So across the population of on comb bipolar. There doesn't seem to be some rescue. And then the, the work from the XBC and from the so to encourage the center group. And then from the beer, the linker and share group that showed that as comb bipolar is in ground squirrel can reestablish their cone context, both of those are morphological studies, but they give us some hope that maybe only there's a bipolar cell that can upregulate its upregulate following the comas. So, sorry, the short answer is we don't know yet, but we're trying to gather clues to see who might be the sensor. Okay. All right. This is okay. I like to say thanks again for accepting for invitation. It was a really nice talk. So now I am posting the, the zoom link. So we can. Sorry, we can continue our discussion. In a private room in zoom. So I will leave the link here. And we can discuss more in detail about you, your findings and potentially future projects so everybody's work is welcome to join us. Yes, so thanks again for this. The, the link is, is already posted so everyone will be welcome to, to join us I will, I will keep the stream on for some few minutes for starting our discussion, but again, thanks a lot for, for your beautiful talk and data. Thank you. I appreciate it. Beautiful talk. Seriously. It's very intriguing how, how you can, you know, manipulate the ablation and then study, you know, in particular sites of the circuit, how these different properties are changing. And yes, as you say, we are, I was intrigued because as you say, and cons in the writing of the zebrafish seems to be regenerating quite quickly during the larval stage but well I was not aware about using mice this is also a recurrent, you know, feature. Yeah, I think, I think a similar manipulation in zebra, live as zebra fish would be interesting to see the rewiring, and if there is also some sort of a functional recovery. Yeah, yeah. Let me ask you about the spatial temporal receptive field reconstructions. They look quite nice and you're using a technique I'm not very familiar with so these were flashing bars. And the bar was, was it only ever flashed at one location at any point in time or multiple location. It was simultaneous. So, at each point in time, each bar is the intensity each bar is drawn from a Gaussian distribution so the whole field is illuminated bar just flickers so it's, it's like white noise, except the reason we use bars is because we get a much stronger around in the, when we, when we map the spatial receptive field, this is white noise. Yeah. And we're assuming some symmetry but of course we can just change the orientation of the bar to see the spatial receptive field in the other dimension. Sure. And how long, how long do you need to so what's the rate at which the bars update and how long do you expose the cell. How long it takes to reconstruct a field with, with that kind of, because they are very nice the surrounds were indeed very clear. Yeah, it's surprising how, how nice they are, considering I think traditionally when the checkerboard white noises use the surrounds are sometimes there but you know under specific conditions they can be enhanced. So we use, we're updating around 30 Hertz. And then we can show the bar for probably a minimum of two minutes to get a nice.