 Okay, and it appears that we are officially live, so hello everybody, and welcome back to another session of our Sussex Vision Seminar Series, as always within the worldwide neuroinitiative. I'm George Cafetsis, a master's graduate from Tomas Oilers Lab, and currently a PhD student with Tom Baden. And as your host for today, I would like to once again begin by thanking Tim Vogels and Panos Bozellos for putting forward this ever-expanding initiative towards a greener and much more accessible seminar world. Having said that, allow me of course to get back to the reason we all gathered here for today, and introduce our guest from University of California, Berkeley, the Paul Leach Distinguished Professor in Biological Sciences, and member of the Helen Willes Neuroscience Institute there, and of course member of the American Association for the Advancement of Science, Professor Marla Feller. Marla is a physicist by training and after her bachelor at Berkeley, she studied surfaces and liquid crystal interfaces by means of second harmonic generation for her PhD with UN Ron Sennig. I do not know what credit is due to the Woods Hole summer course on Neural Systems and Behavior for Marla's passion in biophysics and neuroscience, but that started becoming evident through her postdoctoral years. First working with David Tank at Bell Laboratories and developing optical methods for studying the pre-sine optic terminals of the frog optic tectum, and then as a Miller postdoctoral fellow with Carla Sudge using the retina as a model system and doing seminal work on spontaneous activity. Marla started her lab at NIH in 1998, held the position of associate professor at the University of San Diego, and in 2007 she returned to Berkeley, where nowadays she is a professor at the Department of Molecular and Cell Biology. And before I move on to more research specific introductions, I consider it highly important to mention that throughout the years Marla has received many awards not only for her research, but also mentorship. In her lab they employ both optical and electrical techniques, and by focusing on mice and their retinal circuits they study spontaneous activity, how it is generated and what roles it plays in the development, organization and maturation of neural circuits and systems. And today we will have the pleasure of hearing about their latest findings in her talk entitled interplay between circuits that mediate spontaneous retinal waves and early light responses during retinal development. So without any further ado from my side, please all welcome Professor Feller. Marla, the stage is officially all yours. Great. Wow, what an introduction. Thank you George. That was great. So let me share my screen. Okay, are we, is it all working George? Yes, we are good to go. Okay, let me just hide this. Great. Okay, well George, thank you so much for that introduction for the invitation to participate in this series, and also to Tom for organizing the series it's been, it was great during the pandemic it kept us all sane. And I greatly appreciate that it's continuing to go on. So I shortened my title for today to make it a little bit more tractable. And what I'd like to do is talk to you about work that my lab focuses on having to do with neural activity, how it's generated and what role it plays in development at stages of development that are prior to the photoreceptors wiring in and vision begins. So just as an overview of what I'm going to talk about today. Oh, I want to, I want to start by showing this picture of the people who are in the lab. This is actually a picture from the lab a year ago, but it's going to be these people's work who I'm focusing on today. And so you'll be, I'll be highlighting them as we go along. This is on a beautiful island that's in the middle of San Francisco Bay on a hike that our lab took. Okay, so let me just overview the sources of activity that are that are present prior to the maturation of vision. So the first we'll be talking about is George introduced are these spontaneously generated retinal waves, and how they provide this robust source of depolarization with these very interesting propagation patterns that change across development. So I'm going to start by talking about that. I'm then going to talk about a second source of activity, which are generated by these intrinsically photosensitive retinal ganglion cells, which are functional very early in development and mediate light guided behaviors. And then I'm going to introduce some strategies that the lab is using for disentangling the role that these two sources of activity play in visual system development. But there's time at the end to expand out a little bit on how, you know, the field can move forward, making clear that one has to disentangle these two sources of activity when thinking about development. Okay, so just to get everyone on the same page, I'll briefly introduce the retina. So the retina is this neural circuit that lies at the back of the eye light enters the retina and gets focused on. Now, if we look at a cross section, the retina is not just a single plane of photodetectors, but it's actually contains many neural circuits, and I'll be talking a little bit about these circuits during development. So in the adult retina light sort of passes through all the tissue and activates the photoreceptors that are the back of the eye. And then the output of the retina are these retinal ganglion cells and it's their axons that make up the optic nerve and carry all the information about the visual scene to the brain. And that the processing that happens within the retina happens because of these intranurons that link the photoreceptors to the retinal ganglion cells. Now the periods of development I'm going to be talking about are long before this circuitry is established. And to give you a sense of when that is in development, I'm showing you a schematic here of human brain development before birth. So the way that the visual system gets wired up is the eye actually develops independently of the brain. And at this incredibly early stage of development retinal ganglion cell axons project to their targets in the brain. And humans actually open their eyes at five months gestation. So all of the work we do in our lab is in mice. And most of the work that I'll be talking about is sort of in this period of time that precedes the five month period right before eyeopening. Okay. And so at this time, clearly in humans and in mice, there isn't a lot of visual experience. And so there are other strategies that the nervous system is used to sort of activate these early circuits. So just to give you a sense of what the retina looks like at this period of development. This is again a schematic of the adult retina on the left that I showed you with ganglion cells at the bottom and photoreceptors at the top. Now in the early stages before postnatal day 10, which we call P10, it turns out there are photoreceptors that are there that are post mitotic, but they have very small outer segments. And most importantly, those excitatory interneurons by polar cells have not formed their synapses that connect the photoreceptors to the ganglion cells. There are also amocrine cells that are present and starting to form synapses with the retinal ganglion cells. But you can essentially say the retinal ganglion cells are separated from any influence from the photoreceptors. So just because they're separated from photoreceptors doesn't mean that they're silent, but instead they're spontaneously generating activity. And this activity is what we call retinal waves. And so this is a movie actually that was taken by Alex Terriackel who was a postdoc in the lab. And we'll be talking more about Alex's work in just a second. And what Alex did here is he took out, he isolated a mouse retina from a mouse that expresses GCAMP in ganglion cells and took a movie of the spontaneous activity while focusing the two photon microscope on the ganglion cell layer. And this is what the activity looks like when he does that. And you can see that there's this spontaneous increases in calcium in these cells. And we know from a lot of work that these increases in calcium are due to these cells firing action potentials. So this is calcium that's coming through voltage gated calcium channels and some neurotransmitter receptors, but it's all mediated by depolarization. So the lab is very interested when we want to understand the circuits that generate waves and what role they play in development. We actually look at these propagation properties of the waves of gaining insight into both of these questions. And so we have been really excited about using a new technology, a new what we call macroscope that was developed in the lab by Ben Smith. And this is just a very simple epifluorescent scope where we have a nice 4x objective and a really nice CCD camera. And Ben has designed an imaging system that really very nicely projects this information gathered from this objective onto the camera. And by optimizing the system, we have a field of view that's five millimeters across and with a resolution of 1.5 microns. So we have cellular resolution, but this what this allows us to do is image these waves over the whole retina. So this is a schematic of sped up activity and the retina and this macroscope is allowing us to image this activity really to gain much more insight into spacial temporal properties. So this is important for a couple of reasons. So as I was saying, we care about these spacial temporal properties because they give us insights into the circuits that mediate these waves and what role they might play during development. And this is interesting because waves actually persist for an incredibly long time during development. So during this period of development, the retina itself is undergoing a tremendous amount of development itself. So when we first see waves, which we see embryonically, the retina is made up of a lot of dividing cells and migrating cells and the ganglion cells are mature and they sent their axons. They've just reached their targets in the brain, but there isn't a lot of synaptic circuitry going on from work that was done by Jeremy Keisler. We know that these starburst amocrine cells, which are going to be really featured in understanding circuits for retinal waves, are actually present, but they're just migrating into the position far from their adult morphology and positioning. Stage two waves are apparent later in development during the first postnatal week. And this is a circuit. At this point, we know that these starburst amocrine cells have actually started to form synapses with each other and with retinal ganglion cells. And those synapses are cholinergic. So starburst amocrine cells release acetylcholine and we know during the stage of development that they themselves have nicotinic acetylcholine receptors as do the ganglion cells. And so we think of this as a kind of cholinergic circuit. And then stage three, which happens during the third postnatal week, become present, is when those bipolar cells finally form those synaptic connections with the ganglion cells and the amocrine cells. And at this point, the circuitry that mediate waves switches. So at stage two, we can put on nicotinic acetylcholine receptor blockers. Waves go away. We do that on stage three. Waves are still there. And instead, in order to block them, we have to use ionotropic glutamate receptors. I've actually included some references down here, which there are many, many papers by quite a few labs that have identified the circuits that mediate waves, but I like to highlight the ones that are kind of my favorites. There was a nice one by Kevin Ford from our lab in 2012, Aaron Blankenship studied stage three retinal waves in our lab and actually there's a really fantastic paper that I wish came from our lab. They came from, it was done by a crew in Kirkensteiner, which is a really beautiful model that sort of has gotten at the mechanisms that underlie stage three waves. And I am going to highlight some work that is just coming out from our lab, should be published any minute, we hope, but right now it's in bioarchive and this is work that was led by Christian Vufo. Okay, so, and this is, and so the work that I'm going to be talking about was, was, as I said, led by Christian Vufo with, and, oh, sorry, I'm not going to talk about that, and contributions from Maya, Andy Chen, who is an undergraduate in the lab and Ben Smith, and this is now the microscope imaging that we've done of these waves across these three stages of development. And so what you're looking at here is this beautiful slide that will, that are microscope images of stage one, stage two and stage three waves. What's shown on top, I should say that the, that the pink, the magenta background is just a, it's a, it's a, and essentially an average, it's, it's all the activity that happens across this retina. And then hopefully in the green, you can see the delta F over F signals and you can see the difference in the imaging. What's shown in the top, what's shown at the bottom is the segmentation algorithm that Ben Smith has put together that really identifies where the rate, the waves are happening. You can really see a lot of differences. So for stage two, you can see there's this one mammoth wave that's slowly making its way across the different, the entire retina. While on stage three, you have a lot of littler waves that are actually propagating slightly more rapidly. And then the stage one waves seems to have a lot of what we think of as kind of false initiations and every so often you get one of these big propagating waves. So another way to sort of summarize what this activity looks like is in this nice representation that Ben has developed. So if you look at in this plane, this is an XT representation. This shows you the spatial extent of the waves that have propagated during about two minutes of recording. And then if, and as it rotates and finalizes here, you can see what that looks like over time. So stage two, we had one big wave that propagated across the retina. Stage three, there's a lot more complexity in those waves. And stage one has a lot of little tiny initiations with some larger propagating events. Okay, and our lab is super interested in doing this full characterization because and trying to relate this to the circuits that mediate these waves as well as what role they play in development. Okay, so let's focus a little bit on how are we going to get this question about what role do these waves play in development. And so let's talk about what's happening in development at this time. So I'm first going to talk about retinal projections to the brain and this is where waves have been implicated quite clearly in playing a role in development. So during this early stage one waves is when retinal ganglion cells are actually innervating their primary targets, they've reached their targets in the brain and in some cases are starting to innervate them. And then during this extended postnatal period, actually primarily during stage two, there's a lot of refinement of those projections. So there's retinotopic refinement of retinal projections to the colliculus, there's ispecific refinement of projections to the colliculus and lateral geniculate nucleus of the thalamus, and there's also laminar refinement in these target tissues as well. So the way that we get at what role waves play in driving this refinement is to do a loss of function experiment. And the one that has been very productive for us is a mouse that's missing a particular subunit of neuronal nicotinic receptors, the beta-2 subunit. So we use this mouse a lot. It has greatly reduced and altered retinal waves between P1 and P8, and it still has stage three waves. They come on a little bit earlier, but during this period, during the first person in a week, there's very reduced waves. And so to show you how convincing this data is that in this knockout, we have a strong phenotype, I'm going to turn to some experiments that were done by Alex Tyriak. Well, he was a postdoc in the lab. Everyone should know that Alex now has his own lab at Vanderbilt University, that's just getting started, where he's continuing to work on spontaneous activity in the visual system, but also in sensory motor systems and setting up circuits that mediate twitches. It's totally cool work. Everyone should check out his website. Okay, so what Alex did in the lab was to look at some experiments where he looked at eye-specific segregation. And the way he did these experiments is he did injections of cholerotoxin and anterograde tracer, where he put green in one eye and pink in the other. And these get anterogradely transported and they allow you to look at what their projections are like in the targets. He then used this method for clearing the brain so that we could look at retinal projections throughout the brain. And when he does these experiments, these are now what you're looking at, is a projection through the lateral adreniculate nucleus. You can see that the contralateral projecting axons take up a large part of the projection to the LGN, while the ipsi is a smaller projection, which is just showing you one section in the middle. And then you can see that there is this segregation. So this segregation actually comes about by postnatal day eight. It's present by postnatal day eight. And there's a lot of overlap between these projections at postnatal day four. And so there is this segregation process. And so when we repeat these experiments in the beta two knockout, you see that this segregation has really failed. But just on these experiments and experiments like this, that we've been able to show that these retinal waves during this early period of development are very important for refinement of retinal projections to their targets in the brain. Okay, but now I'd like to focus on work that the lab has done to understand the role for waves perhaps in retinal development. And so what's happening during retinal development during these ages? Well, during stage one waves, there's a tremendous amount of neurogenesis and positioning of cell bodies. And then there's also a couple periods of cell death. But the one I want to focus on is this postnatal day between P1 and P4. A lot of ganglion cells undergo cell death with estimates of high as 50% of ganglion cells will die during that time. And so in order to sort of get at these questions, we need to come up with a way of manipulating stage one waves. And we already have a way for manipulating these stage two waves during cell death. Okay, so now I want to talk about this work that was the focus of Christian's thesis project. I would say this project was really, Alex was really kind of the mentor who guided this project along. And there were significant contributions from Andy Chen. Andy was an undergraduate in the lab at the time and he's now getting his PhD at Harvard. And of course we have the contributions from Ben on the study. Okay, so what this team did is they did two photos on calcium imaging. And what you're looking at here are different cells, individual cells as a function of time. And these are these kind of raster plots that can tell you what their participation in waves are. So this is what the control looks like. And you can see waves are happening around once a minute. And what they found by doing experiments is that stage one waves were significantly inhibited by a gap junction antagonist called MFA. And that was something that was sort of known from previous works from our lab and other labs. But there was a bigger surprise came when we found that these waves were actually significantly inhibited by a nicotinic acetylchoyne receptor antagonist. That despite the fact at this incredibly immature state of development, where I had mentioned the starboard stamercrant cells are far from their target position in the INL and in the ganglion cell layer, they are still releasing acetylcholine and that acetylcholine is contributing to waves. So the fact that we know acetylcholine is involved in mediating these waves, we look to the beta 2 knockout to see what these stage one waves look like. And what we saw is summarized here. So we see that unlike at stage two, we still see a lot of wave activity in the beta 2 knockout. But interestingly that remaining activity is not blocked by nicotinic acetylcholine receptors and is completely blocked by a gap junction antagonist. So the waves are still there but there has been this, I guess we could say compensation so that the remaining gap junctions are sufficient to propagate these waves. But what you can tell from this is that the waves are actually greatly reduced. So if you were to look at the percent of ROIs that are active during waves, and we can see that in some of the big waves in control, you can get 50 to 70% of the retina participates in those waves, but this is significantly reduced. So that's another way of seeing the waves are much, much smaller in beta 2 knockouts. And what's not shown here is they actually propagate at half the propagation speed. So waves are reduced in the beta 2 knockout during stage one, and they're significantly reduced and almost absent in stage two. So what does that do in terms of retinal development? So to get at this question about cell death and cell positioning, we turn to a particular ganglion cell cell type that we have a good marker for. And that is these intrinsically photosensitive ganglion cells. I'm going to talk about them a lot in just a moment, but just suffice it to say this is a particular cell class that is present. A subtype of retinal ganglion cells. And this is actually the work that was done by Andy Chen. And so if you look at all of the labeled cells in an opn4-cree across with the TD Tomato P1 and compare it to postnatal day 7, you can see there's a dramatic decrease in the density. Although the retina has grown a little bit. And if you look at these at higher magnification, and I want to emphasize again, these are all images taken from the microscope, including this high resolution images. You can see that there is this reduction in density. And then if you compare this between the wild type and the beta 2 knockout, you don't see a significant difference. So this was quantified in the following way. First, if you look at the number of these cells, the density of cells, I'm at postnatal day 1 to postnatal day 7, across development, you see that there is a dramatic decrease. And this is, we know, is due to cell death. And if you now compare this in the beta 2 knockout, you don't see that there's a significant difference in the total, in the density of cells at postnatal day 1. Nor do you see, and you still see a dramatic amount of cell death that happens in the first postnatal day week. So it seems that the absence of waves during stage 2 has not prevented cell death from happening. We also did a couple of measures of the organization of the somas, what the retina world likes to refer to as the mosaic organization. And here we plotted the nearest neighbor distances in a cumulative histogram plot. So postnatal day 1, you can see that there's a lot of overlap between what we see in wild type and beta 2 knockout. And that there is a significant increase in the nearest neighbor distances that you can see between the wild type and the beta 2 knockout. As a way of talking about the organization of the somas within the retina, we calculated the regularity index. And that's the nearest neighbor distance as a function divided by the standard deviation. The higher the regularity index, the more regular the mosaic is and the lower it is, it's closer to random. And so we found that actually the regularity index did not change. And actually this value of being around 3 is what is predicted for a simulation of randomly distributed somas. So the somas remain randomly distributed and there's no change in that in the absence of normal retinal waves. Okay, so we don't think that retinal waves are playing a role in these important developmental processes, although of course a more significant blockade of stage 1 waves are necessary to completely roll that out. I now want to very briefly talk about experiments that just were published last year from the lab implicating stage 2 retinal waves in the assembly of a retinal synaptic circuit and one that is very dear to our hearts and that is the direction selectivity circuit. So this is a paper that was done by, you know, Team Direction Selectivity, that includes, again, Alex Karina Beistron, who's a graduate student in the lab, Josh Tworegg, and Maya Pitcher. Okay, so what did we find? So I'm not going to be introducing direction selective sales, but for those of you out there who know that there is a population of sales within the retina that prefer one direction of motion over another. If you record with two photos on calcium imaging over all the direction selective cells in a field of view, you'll find that those prefer directions cluster along four axes. So you'll have some cells that prefer left, right, and then some that prefer up and some that prefer down. Those cells are all present right when the eyes open, so you don't need any visual experience from the establishment of this direction selectivity. And when they looked at the beta two knockout, which just to be clear is lacking waves between P1 and P8, they find they have this remarkable result, which is there seems to be no horizontal direction selectivity in these mice. So this lack of waves during this early stage of development prevented the development of horizontal direction selective cells. So those cells are still there. They just have a lot and they're still responsive to light, but they have just lost their directional preference. So that indicates a very specific circuit has been disrupted. The robustness of this phenotype and the implications of this phenotype were actually described by JC Kane's lab over 10 years ago, 15 years ago, oh my goodness. When they showed that that mice that that these beta two knockout mice actually lack a horizontal auto kinetic reflex, a behavior that's known to be mediated by direction selectivity, but their vertical reflex seems to be present. So, so something about waves is very important for setting up the circuit and sort of getting at that question is a major focus of the lab that I won't be talking about in this talk. Okay, so as an interim study about retinal waves, I just want to say again, retinal waves are this robust source of activity before vision. And they're required for the precise connectivity to emerge throughout visual system for some circuits within the retina and not others, and of course for retinal projections to their targets. Different stages of waves have different spatio temporal properties. I mean that we retinal activity is implicated in the wiring of direction selective circuits but not in the cell death, at least of IPRGCs. Okay, and now when thinking about these questions, and we'll get to this at the end. I just want to say when we talk about whether or not retinal waves are involved in a particular developmental process. We can think about it either at the level of do the spatio temporal properties of those waves matter for that developmental process, or is it just the fact that there's some cell autonomous activity in each cell that drives it. And that's a big question that we're always asking ourselves when we're looking at how retinal waves affect development. Alright, so now I want to turn to this other topic and I have to say I have been studying retinal waves since I was a postdoc for a very long time. I have, I had said forever that they are the only source of activity before vision is established. And then this paper came out, Juliet Johnson and David Copenhagen, where they showed that mice pups that are only six days old will actually turn away from the light. So these are, this is a behavior experiment we have five different mice there's an LED to the left. And at some point these mice are going to the lights going to flash you can see all these mice are turning away from the light. Okay, so I told you that this is not photoreceptor mediated and rodent cone mediated because those photoreceptors are not connected to the ganglion cells, and therefore not connected to the brain. And so what mediates this behavior is actually as class of these retinal ganglion cells are these intrinsically photosensitive retinal ganglion cells. So if you create a transgenic mouse that doesn't have melanopsin, this behavior goes away and mouse pups. So, so this work from the Copenhagen lab really established that these cells are there and they're driving, they're driving this behavior. Okay, so IPRGCs are present and they have been implicated in many developmental processes and every time I open a journal there seems to be another process that they've been implicated in. So just to give you a sense of the development of them IPRGCs are light sensitive as early as e 15, and the different types of IPRGCs are fully differentiated by postnatal day four or five. They have innervated the supercosmatic nucleus and the LGN by postnatal day zero. They have been implicated in some retinal development and work from Jordan rena's lab, excuse me, that having to do with the lamination of cone somas, because of a transient projection that IPRGC send up to the developing photoreceptor layer. They've been implicated in the maturation of vasculature either by suppressing angiogenesis or regression of a developmental vasculature called the hyeloid. They have been implicated in the maturation of circadian driven functions, and there was a really interesting study amongst quite a few that they're very involved that this early light activation of IPRGCs is critical for synaptogenesis and cortex. Modulating release of oxytocin during this very important early developmental period. So there's been a tremendous amount of work right that says that that IPRGC mediated light responses are implicated in lots of development that's going on at this time. But there is this question, and this is actually was the focus of Franklin Kevall Holmes thesis, which was how does a mouse tell the difference between a wave and light. So as far as we knew all ganglion cells participated in waves, and, and therefore all IPRGCs participated in waves and so how would a mouse know whether or not an IPRGC is depolarized by a wave versus whether it's stimulated by light. So this is a question that Franklin was really focused on Franklin graduated this is his graduation party, and during the pandemic, which was in 2020 Franklin is currently a postdoc in Michael does lab where he continues to study IPRGCs somehow only in the adult, I don't know, but I'm sure it's I'm sure it's very interesting. I'm sure he's doing fantastic work there. Okay, so I just want to very briefly go over Franklin's work, his thesis work was published in in two papers that their references are shown there. And I just want to sort of show you what his experiments look like. So what Franklin did is he would do a retina dissection he would put his retina under one arm of this two photon microscope in order to do calcium imaging. And then he used a projector system that would project blue light in order to activate these cells and just to give you an idea of what this data look like. You can see on the right that when there are these flashes of light of blue light that you can very clearly see light responses from IPRGCs. Now in these particular experiments he's blocked waves in order to isolate these IPRGC responses. And when he did that for some of you who are looking you could see lots of cells that were responding but they actually seem to respond with very different kinetics. And so what Franklin did is he using the strategies that were developed in the Euler lab by Tom Baden and Barron's to use a sparse principal component analysis to classify these cells in the functional groups and he found that there were about six functional groups. And you can see that and if you look at the average response and so what you're looking at here are the responses of cells to long duration stimuli that are increasing intensity and you can see these some of these cells have a very, very low threshold to the response to very low levels of light and they have a sustained response while some say have a higher threshold and have a transient response. Franklin went on to try to assign these functional groups to anatomical groups and just to introduce this to people. This is a summary plot that talks about the multiple types of IPRGCs. There are six different types that have been identified. Tiffany Schmidt's lab has shown that many of these are present very early in development. And that the different subtypes are identified by transcription factors and also where their projections are in the brain. Franklin would map his functional groups onto anatomical groups. He had six functional groups or six anatomical groups, but what he found was that they did not map precisely. So for the first functional group, those mapped very well onto the M1 IPRGCs. And so that was there was a functional group that mapped onto an anatomical group, but all the other anatomical groups actually were mixed into growing these functional groups and that mixing of these groups happened because Franklin showed there was extensive gap junction coupling between them. Okay, and to give you a sense on how powerful this gap junctions are in terms of mediating the responses of IPRGCs, what you're looking at here is the impact of a gap junction blocker on the light responses of the different functional groups. So if you look at the response and control conditions, you can get to a light stimulus, you get a fractional change of fluorescence that varies. And if you compare that to the responses in the presence of MFA, almost all the cells, all the functional groups lose their light response except the M1s. So the M1s are able to keep their light response, even when gap functions are blocked. Okay, so what does this mean? So remember, we want to get at this question about how the animal tells the difference between a wave and between a light stimulus. And so the first thing we needed to, so now we wanted to get at that question to figure out what type of IPRGCs might be mediating photo aversion. So we have one class, the M1s that are kind of cell autonomous, then we have the other class that are all gap junction coupled together. And so Franklin's strategy to get at this question was to take advantage of many of the tools that have been developed to sort of do loss of function in these two groups. So we know that transduction in the M1s are mediated by Trips C6 and Trips C7 ion channels, while they have a weaker contribution to the other functional types. And that we know that if, and if you can do intraocular injections of gap junction antagonists, that that will allow you to block the effects of the other cell types. Okay, so we took advantage of this triple knockout that was developed by Lutz Bernbomber and was given to us by Tiffany Schmidt, who is a collaborator on this project. What Franklin did was the following to quantify photo aversion is he mapped the head, he put, he literally put an arrow on the head of each of these little mice, and as, and then when the light, when they were exposed to the light, he looked at their head position and eventually they turned their heads at some distance away from the light. And when he did this in control he looked at how much maximum head angle there was in the dark versus in the light and he's two different intensities of light. And he found that in the Trips C6 7 double knockout that actually there it was that photo aversion was was completely prevented in at the at the medium light intensities and there was a little bit of photo aversion at the highest light intensities. However, when he blocked gap junctions, he did not impact photo aversion in these cells so so intraocular injections of gap junction antagonists did not prevent photo photo aversion, really sort of arguing that it seems like it might be M ones, but to really establish that he made use of this incredible pulse line that was developed by Samarhitar's lab, where it's an opn Cree that's crossed with this brain 3b and detail. All you need to know is it kills all of the IPRGCs, except for this one subset of M ones which are brain 3b negative, which project to the super charismatic nucleus. Alright, and when he blood killed every IPRGC except for these M ones, what he found is that photo aversion was completely normal. And so I think with these experiments. It's we can safely say that the M one IPRGCs mediate photo aversion in these pups. Okay, so now we're ready to answer our question. What does the mouse tell the difference between sensory experience that we know is mediated by and M ones, and then of course is the spontaneous activity that's mediated by the retinal waves. And so these experiments so what Franklin did in collaboration with Andy Chin did the following experiments. He allowed, he looked at cells that participate where he looked at waves, and then he also looked at light responses. He looked at these are delta F over F transients from three different cells that are non IPRGCs. And where you see these red dotted lines is when there's a wave and you can see that these cells all participate in retinal waves, but none of them responds to the blue light. Then did experiments in where he did the same experiments but now he looked for the punitive IPRGCs, and those were because they had light responses. And I'm sorry they were GFP positive, right because we were using a mouse line that had GFP and IPRGCs, and you can see that they had light responses and some of them participated in late ways and have light responses. And then it turns out there was a subset of these cells which were the M ones which have this very distinct light response, but they actually don't participate in waves that M one IPRGCs appear to have a much smaller participation in ways. So in analyzing this data, if we look at say a different one of the different functional classes and you do a wave triggered average, you can see that you get a robust response and IPRGCs. But if you look at the M ones, there's no consistent response. And, and Franklin did this for a variety of different light stimuli and you can see that if you look at the delta F over F in terms of the response to light versus how strongly they're depolarized by waves, you can see that M ones are much more strongly depolarized by waves than they are by light. So that seems to argue that M one IPRGCs represent an isolated channel for mediating photo aversion during development. Okay, so to summarize on the second part of the talk, light stimulation of IPRGCs mediates behavioral responses at the same age as their rental waves, the visual system differentiates spontaneous from light about events by exclude exclusion of M one IPRGCs from waves. And that the light stimulation of IPRGCs. Oh, sorry, I just repeated something. Sorry. Okay. So, so forget that third one. And that the visual. Oh, I repeated them all because they're really important findings. Okay, so this question about how how M one IPRGCs are excluded from rental waves is actually something that I would really love to follow up. So it could be that M ones don't have nicotinic acetylcholine receptors, and that's why they don't participate in waves. And certainly the single cell RNA seek data seems to argue that they have a lower expression of nicotinic acetylcholine receptors, or it could be that they're strongly depolarized by waves and they go into a depolarizing block. And Franklin tells me that he's doing those experiments may I'm hoping I'm trying to convince Franklin to do these experiments while he's in Michael's lab Michael please let him do these experiments but hopefully we'll have an answer for you for that very soon. Okay, so now I want to take a few minutes to sort of get at this question about the interplay what we other experiments the lab has done to get at this interplay between IPRGCs and waves, because they're both going on at the same time. We know that waves depolarize IPRGCs. We know that light does not dramatically influence stage one or stage two rental waves. So when we do calcium imaging, we see no difference with what how light affects the space or temporal properties of rental waves. During RENNA found by doing MEA recordings that when you turn on the light that there are some cells that you increase the burst duration of some cells during when you turn on the lights during stage two waves. So, but the frequency of the waves and the propagation speed of the waves doesn't change in light. We acutely blocked rental waves for about an hour. Or if we look in the beta two knockout mouse, we have a dramatic increase in the number of light sensitive cells. We'll get twice as many light sensitive cells in when we blocked waves for you know the sustained period of time, and we'd love to understand how that happens. So our current thinking is that this is due to an increase in the gap junction coupling right between IPRGCs and perhaps non IPRGCs. But we're also open to the possibility that this could lead to some sort of changes in melanopsin expression that changes the sensitivity of different IPRGCs, and that's an ongoing area of research. And we're really interested in this question about whether extensive coupling of IPRGCs contribute to rental waves. So they are a very gap junction couple network, we know that gap junctions particularly during stage one are very important for rental waves. And we wonder if that coupling is what's critical for gender plays a role in these waves. There were some experiments done by Samarhatar's lab that showed when you kill the IPRGCs that that had an impact on waves and I specific segregation. And so that's kind of supportive of that particular model. And now I just want to expand out a little bit more about how this early activity influences development. We are not the only lab studying this and there's a tremendous number of labs I just kind of want to highlight some some interesting findings that are out there, which is that this early retinal activity is known to regulate the timing of corticothalamic So when cortex innervates the lateral geniculate nucleus, that's something whose timing has been has been identified as being important for the timing of which is controlled by retinal waves that there is an amazing paper that came from The Lopez Bendito lab that showed that the segregation of visual and somatosensory circuits within the superior colliculus is something that's dependent on actually early waves perhaps stage one retinal waves. And that, and that this idea that spontaneous activities important for development of retinal microcircuits has proven to be the case. In actually the Drosophila system so there's been these wonderful studies by Akin while he was in Larry Supersky lab and now in his own lab at UCLA that shows that there's spontaneous activity in the Drosophila visual system, and that blocking that activity also prevents the development of some specific circuits within that visual system. Our lab and others have really focused on the interaction between retinal waves and glia. So we know that work done by Julie Rosa while she was a postdoc in the lab, and Josh Toreg who was a graduate student in the lab and is now graduating and is now staying on a little bit longer as a postdoc has shown there's robust signaling between retinal waves and glia retinal waves cause period and cause transients in the glial cells, and we're really interested in in what role these glial this neuronal glial signaling plays and in development, and I should say this was also seen in zebrafish and work done by the do lab. There have been these amazing talks on this very series that I have been watching that have gotten us very interested in the vast and whether or not retinal waves and signaling to the glia might be interested, it might be involved in the development of the vasculature. These were talks by Excuse me that we heard here on in the series sorry and we're carrying out some experiments by a new graduate student in the lab as Samira who's going to be sort of getting at this question about whether blocking retinal waves will have some effect on development of the vasculature. Another sort of thing that we're extremely interested in is that retinal waves our lab has been able to show with an imaging using a cell based fluorescent sniffer technology that retinal waves drive release of dopamine and and we're really but we also know that in the adult certainly that activation of IPRGC's drives release of dopamine and during rena has some very good evidence that this also might be important for development of the retina itself. And so we are continuing to do experiments to develop methods for looking at live imaging of dopamine in the retina and these experiments are being carried out by a new postdoc in the lab Ron Shan. And then there could be all these things we haven't thought of yet right so there can be a lot of developmental processes that are going on. And as I had said earlier I'm really interested in this idea of cell autonomous effects like do these periodic depolarizations or the strong activation of light do something at the transcriptional level and cells. And so I'm really excited by a collaboration that we're doing with Karthik Shakar who is a new faculty member here at UC Berkeley. And a real has really novel approaches for studying single cell RNA seek data and using that as a way of looking at the development of different ganglion cell types and retinal types and this is going to be work that is being led by Rachina another new postdoc in the lab. Okay, so this is the lab. We love retinal waves. And this is the current state of the lab. The current members of the lab as I mentioned Josh has recently graduated staying out a little bit longer to finish some exciting experiments he's doing on the development of direction selective circuits. I want to call out to the recent graduates I think I've taught Alex is a professor at Vanderbilt Christian is now starting a really exciting fellowship at the World Bank. Franklin is a postdoc in Michael does lab Andy Chen is I think rotating in Michael does lab and Yixin Zhang who worked with Franklin on his projects is now a PhD student at Hopkins and for those of you who know other recent. Fellow labs graduate Matthew Summers recently graduated is now at the Allen brain Institute and molecule Kezney is at the Institute for tonics in Barcelona. All right, thank you very much. Thank you very much for this very impressive talk Marla and your meticulous attempts in trying to figure out what it might be including the something we haven't thought about just yet. So let me go ahead and post the zoom room link in case people want to start joining us already and remind to our audience that after this first initial moderation from my side, I will be continuing in a more informal zoom get together fashion. So please make sure to follow the link in case you want to be part of the discussion later on. So there are already questions appearing the chat before I go into them one very naive borderline stupid question that I have is so when comes to studying retinal waves. Do we know if there's a circadian component in them. Now that is not a stupid question and you know. Okay, I'm going to take a step back to answer this question George, it's so weird to be in a zoom room because as far as I can tell I'm only talking to George and so I'm going to tell George that as he mentioned I was a physicist and that one and people often ask me what is the thing that I miss about physics and I have to say the one thing that is frustrating about neurobiology is we don't there's so much variance. We don't know what the variance is due to and oh my god all our variance could be due to circadian rhythm right so so it is something that certain people in the lab have tried, like if you do run the wave imaging in the middle of the night versus. Okay, no one's in the middle of the night but morning versus evening, and we don't see big differences but I would say that we have not studied that rigorously enough to really make that conclusion. I will say when we put on dopamine receptor antagonists and in wild type right now we don't see an effect on waves. So if the circadian rhythms are changing dopamine levels then that it's that's not likely to be a method but we do all of our experiments and see 57s. They don't have melatonin and so it could be that we're not studying the right mouse to figure that out. So yeah, it might be important we haven't looked at it. All right, thank you very much so following your comment people are already joining us in the zoom room so I would like to ask them to keep their cameras and their microphones muted for the time being, as I moderate the conversation with the questions that appeared in the chat. So the first one is from Daniel Kershensteiner. Great talk since M1 IPRGC's do not participate in waves. Do they differ in their eye specific segregation and retinotopic refinement from other RGC types. It's a great question we have not looked. We have not looked at that I think that would be a really fun thing to to address now that we know that they seem to be doing this on their own right to be independent of waves. I think I would love to I would love to look at that. The next one appearing the chat is from Gotham, our Tramani, a fantastic talk I was wondering are starbastamacran cells the source of acetylcholine for stage one waves. So we haven't tested that explicitly and when, if looking at Jeremy K's work. Actually, I guess he didn't. I guess I don't know if Jeremy if you're out there you can answer in the, in the, in the chat but I don't think he did. Like Collin acetyltransferace staining or anything. He just had a marker for starbor cells. So our assumption is that they are I know that there have been some reports that maybe horizontal cells transiently express acetylcholine during development so we haven't tested explicitly whether it's starbor cells but that that's our assumption. And the next one appearing in the chat is from Chase Helmer. Is there any possible contribution of nicotinic acetylcholine receptors in bipolar cells to stage three wave propagation, like propagation speed. Yeah, I mean that was Chase with your studies revealing that those existed have is is made us realize that that's something we have to consider and we haven't looked at that explicitly but absolutely. We don't see I have, I guess I will say during stage three waves we do put on nicotinic acetylcholine receptor blockers, and we don't see an effect on the propagation properties of waves, but we haven't done that sort of at the level of the microscope imaging that we're doing now. We've done it more on like smaller fields of you and so we don't see a big effect on frequency, but there may be other like propagation speed and stuff like that that might be important for us to look at so it could be a source of modulation. We do know that starbor cells still participate in waves, right during stage three. So acetylcholine is certainly being released. They're just not contributing. As far as we can tell, they're not the major source of propagating rental waves. Right, and given that there are a lot like you don't have YouTube open so there are a lot of messages saying both thank you and greetings in the beginning, and I would like to thank the audience for you know either waking up everyone. Everyone come into the Zoom room briefly I want to say hello. I think I didn't miss any question that appeared already. So one of mine like before like if there's no more questions I will be terminating the broadcast after this. So one of mine is like so you mentioned that that stage one something that differs in terms of the patterns of rental waves is that we have like a lot of failed initiations. Is it really failed initiations or could it be like a more local function of these retinal waves, like saving the circuits at the local scale and not global. Yeah, I mean I think I failed I mean we make that stuff up right so I would say yes failed initiation is something we made up it could be that it's a little gap trick to couple network it could be it's a little. You know, just source and spontaneous depolarization we don't know what it's necessarily do to so perhaps failed initiation is not is not exactly the right thing to say, I mean one of the things that you know Ben has pointed out as he and Maya are sort of really spending a lot of time looking at the space of tempo properties across development are the transitions and so stage one waves you know as they convert to stage two waves. You know you see more of the big propagating events and few of those very small events that don't go anywhere right so there seems to be that gradual transition and the same thing sort of happens when you go from stage two to stage three now you don't have as many big waves and they start having more frequent initiations and they start to break up a little bit more but then everyone so well you'll have one of these big. What we think of as cholinergic waves so they're really so that so everything is kind of a continuum and and maybe these three periods are artificial in some sense, particularly between stage one and stage two and more work will have to be done to really identify that. Alright, thank you very much. One more question appeared in the chat and I'm sorry if I mispronouncing the name is from Amiros Han the honey amazing work is the retinal waves causing the retinal formation and retinotopic organization or is it more like a correlation if it's causation is it explained by heavy and learning. Yeah, so I'm all right. Is it causation. I don't know here I'm going to wear my physicists had we don't ever prove anything in biology so everything's correlation and and I will say we have the loss of function studies. My crown has done a really very nice gain of function studies where he's done things where he's restored activity back into the beta to knock out and shown that if you restore activity has slightly different. He's altered propagation properties and and he can show that some maps are sort of form normally and some aren't. And he's also done some really nice experiments where he's used channel adoptson to sort of stimulate like the left eye and the right eye So you have left eye versus right eye and he's been able to show that if you do that you can still you can rescue eye specific segregation as long as those left eye right eye stimuli aren't coming to frequently I think he found it had to be like a second or so different I apologize I don't remember that. And so that kind of does argue that this that the salient feature of retinal waves for driving eye specific segregation is that there's a very small likelihood that the two eyes are ever fine firing synchronously in the same retinal topic location. And so I think that that's probably pretty safe to say what mediates it on the other end I think it's it's a huge topic of what many people have been studying. There certainly seems to be you need synaptic transmission and for in order for it to occur, but it's not heavy in in the sense that a classic heavy in model because ipsy is always so much weaker. So the model seems to be if ipsy is active. It will form that it will be able to innovate that a little ipsy patch. But but it doesn't take over the whole retina if contra is silent. And so there seems to be like you need activity to compete within that little patch and so you know it's not a pure heavy in process. Right. Thank you very much for the questions that appear thank you very much marla both for the talk and for addressing them and with this general and whether it's like heavy and learning or not. Yeah, I will be like I will be terminating the broadcast I'm posting once again the zoom room link in case people haven't had the chance to save it before so they can join us and given that we are already more than 10 people who probably have like questions for you. Let's continue offline. Thank you very much. David Daniel up all my favorite people.