 and I believe that we are officially live. Hello everybody and welcome back to another seminar of our Sussex Vision Seminar Series as always with the Worldwide Neuroinitiative. I'm George Caffetzis, I'm a graduate from Tomas Oilers Lab and currently a PhD student with Tom Baddett. And that's your thought for today. I would like to once again begin by thanking Tim Vogels and Panos Bozellos for putting forward these seven expanding initially towards a greener and much more accessible seminar board. Having said that, allow me of course to get back to the reason we all gathered here for today and to introduce our guest from University of Manchester and the director of the Center for Biological Timing that is located there, Professor Robert Lucas. For lunch he's a bachelor degree in Biological Sciences from the University of York. Rob worked with Andrew Loudon at the Zoological Society of London for his PhD on Circadian Control of the Neuroendocrine Axis. During his postdoctoral years he worked with Russell Foster at the Imperial with an ever-increasing focus on Circadian rhythms and in 2000 he started as a lecturer at Imperial before joining the University of Manchester in 2003 where he has remained ever since and nowadays holding the title of Professor of Neuroscience. For a couple of decades now in his lab Rob has been exploring the general principles that govern the biological impact of light. From the retina and the at least traditionally believed as non-image forming light responses to an organismal level and namely how light regulates mammalian behavioural physiology. Furthermore, work of theirs includes the acycoblian characterization of light's effective intensity and the study of different options with both research and clinical perspectives. And having said that I'm very happy to be leaving the stage for him for a talk entitled Melanopsis Contributions to Vision in Mice and Man. So without any further ado from my side please all welcome Professor Lucas. Rob, the stage is officially all yours. Thank you very much George. Let me share my screen. You can tell me whether that's working. And then let's do it. And now, yes, is that happening? Looking good? I can see your pointer, so we are good to go. Perfect, thanks very much. Thanks for that interaction, George, and for inviting me to present on this series. I'm really looking forward to it. I've been an attendee a lot of these. So it's a great initiative. And congratulations on scheduling this between and the hour between World Cup matches. So that's great too. Thanks very much. So as George said, we've been interested in melanopsis and kind of unconventional influences on visual and non-visual functions for a while. And it's shockingly, what, nearly 25 years since we knew that the mammalian retina had photoreceptors other than rods and cones and around about 20 years since we knew that that could be explained by these neurons here shown in an on-fast view of the mouse retina, intrinsically photosensitive retinal ganglion cells expressing the photopigment melanopsin. And the story I'm gonna talk about today is something that we've been working on over the last while nearly decade, trying to understand where the and how melanopsin might contribute directly to form vision. And just a couple of kind of caveats at the beginning. Obviously there's something other people have been working on as well, but I don't have time to review that. So it's really gonna be summarizing our own data on this topic. But I think what I'm gonna tell you the take-ons are more or less consistent with what other people have found as well. Okay, so if you think about the function of melanopsin, one of the things to be aware of is that there aren't any simple answers, right? So if we ask what melanopsin is for, it was discovered of course in attempts to understand the mechanisms underlying circadian photoentrainment. So how the circadian clock is set to local time by light. And pretty soon thereafter, additional kind of end points for melanopsin were discovered in terms of regulation of the neuroendocrine axis and regulation of pupil size. And so there's an idea developing, first of all, that might be a circadian photoreceptor or if not a circadian photoreceptor, non-image-forming photoreceptor, and both of those kinds of terminology have some value. But there's an awkward truth here, which is that melanopsin knockout mice do just fine, right? So you can take melanopsin out of the system and they still have photoentrainment. They still have a pupil light reflex. Really, there's no visual function that they lack. So we can't designate melanopsin as responsible for any particular visual end point. And that makes sense if you think about how melanopsin sits within the overall structure of the visual system. So this is just a schematic of a retina, of a versatile retina. And where melanopsin lies of course is in a subset of these output neurons of the retinal ganglion cells. And those retinal ganglion cells then are capable of responding to light because they have melanopsin, some photopigment, but they're also responding to light because they have connections with the outer retina and therefore are downstream from signals coming from rods and cones. So at the very origin of the melanopsin light response, we have to think about it in relation to what's happening also from signals from rods and cones. There's no separating it, right? It's a unity, right? So then the other really interesting bit of melanopsin biology is that if you look at the IPRGC family, turns out there are multiple classes of IPRGCs and collectively they project to every retina or recipient region of the brain. So on the one hand, melanopsin does nothing because melanopsin knock out mice do just fine. On the other hand, maybe it does everything because IPRGCs go to everywhere. And that highlights, I think, a really important principle to keep in mind here, which is that we need to ask not what melanopsin does, but how melanopsin fits into the picture of vision involving also rod and cone photoreceptors. And to get to that question, we need to start by considering what are the sensory capacities of melanopsin and how might that compliment what rods and cones are capable of doing? Okay. So in terms of describing melanopsin's fundamental sensory properties, of course we have a problem, which is that unlike rods and cones where we could separate a rod or a cone photoreceptor and record its sensory, its light responses, IPRGCs sit there in the inner retina where they're supposed to be getting input from the rods and cones all the time. So if you record an IPRGC light response, it's a composite of melanopsin and rod cone influences. So how can we separate the component of that response that comes from melanopsin? And the most straightforward way of doing it and the way that's being used most widely is to get rid of the rod and cone signal. So we can do that using retinal degenerate or transgenic animals where we're silencing or getting rid of the rod and cone photoreceptors or by applying pharmacological agents which deaffer the IPRGCs in different ways. I'm gonna talk about for a little bit about the kind of the outcome of those sorts of experiments in different contexts. All right, so the first thing if you do that we can learn about the melanopsin sensory capacity, it's that it's maximally sensitive in the kind of cyan portion of the spectrum of the Lambda max about 480 nanometers. So this is really, really useful information but it doesn't really separate it with what rods and cones can do because that's right in the middle of the visible spectrum, okay. So one way in which it's different from rods and cones is if you look at rod-less comas animals or rod-less comas preparations, they're much less sensitive to light than an intact preparation. So the sort of threshold sensitivity of melanopsin is really pretty high. And so if you look at the pupil light reflex, you need about 15 lux of daylight to give the equivalent amount of photons for melanopsin to drive a pupil light reflex. So that gives you an idea of threshold. And then if you look at different responses you can have a slightly different answer. So for the circadian clock, for example, it's much less light. So 0.5 lux of equivalent daylight would give you that effect. So a couple of things from that, clearly far above threshold for rods or cone photoreceptors but equally far from ludicrous amounts of light. So most of us here as we watch the seminar will be having melanopsin activity. It's well within a kind of a dim light range. Okay. So one of the other cardinal features of it from rod-less comas preparations is that the signal has very, very poor temporal resolution. And so this is data from David Berson's description of IPRGC's way back in 2002 and showing that in response to a light step here at low light intensities it takes a long time for the IPRGC to start being activity, activated by its melanopsin signal and then it decays a long time afterwards. And once you get higher light intensities that on switches faster but you get an even slower decay thereafter. And we see that as well if you look in the brain. So this is a work from anesotized mice where we introduce a multi-electrode probe into the dorsal-electrogeniculate in rod-less comas animals. And you can find lots of light responses but again, they have this characteristic being slow to turn on and slow to turn off. So in terms of visual capacity then that implies that there's kind of a space-time averaging element to the melanopsin signal which would introduce kind of visual blur for visual images, for visual patterns. The other thing is that they have very poor contrast sensitivity. So in the rod-less comas animal we've tried quite a few times to look for light adapted responses with reasonable visual contrast and pretty much failed. So it's very hard, this is data for a sinusoidal modulation a very high amplitude sinusoidal modulation a range of frequencies. And you can see there's almost, there's really is no significant response from these rod-less comas animals. Okay. So the poor chamber resolution, poor contrast sensitivity is bad for many visual functions but of course for some things it's totally fine. And one of the functions for which it would be pretty good is measuring that change in ambient light in background irradiance. And in some ways then, if you don't need very high contrast sensitivity, in fact, maybe it's good not to have high contrast sensitivity because you need to track a very wide range of life intensities that might happen over a dawn or dusk, for example. And similarly sort of space-time averaging is quite useful for that because you can get a more accurate estimate of light intensity. And it's interesting that if you look at the threshold for menonopsin responses, they sort of start at civil twilight and it's carrying you all the way through then into that daylight range. So it's very good for that. Okay. So, but what I want to talk to you about is a manonopsin control contributions to vision and of course, that ability to measure background light intensity can be important for the visual system as well. And I'm not going to talk about this in detail, but I've already mentioned the pupil light reflex. And if you think about what manonopsin does for vision, certain pupil size is probably a really important first answer. But there's also growing evidence that manonopsin contributes to kind of a network light adaptation in the early visual system. So work from a number of labs around the world, not my own, have shown that IPRGCs make these centrifugal intra-retinal connections to send a signal out to the rest of the retina. And again, work from partly from my group, but also from others, showing that there's a manonopsin-dependent adjustment in the visual code. And I don't have time to talk about the lovely work of my friends and colleagues in Manchester contributed to this over the years. So what do I want to talk about? I'd like to talk about the work that we've done, asking the question of well and whether manonopsin itself could make a direct contribution to form vision. So why is that a question worth pursuing given what I've told you already? So first of all, we know that circadian light responses in humans can survive even in the absence of light perception. So we think that those responses come from manonopsins as work from Chuck Seiser's lab in the mid-90s. But we don't know, but we think it comes from manonopsin and yet these people have very, very rudimentary visual vision. So having manonopsin is not itself enough to have vision. On the other hand, there are a number of papers showing that visually guided behaviors can be recorded in animals that lack rod and cone function. That's either an advanced retinal degeneration or the work of Eckeren colleagues in genetically silenced rods and cones. So we've done some work with the retinal degeneration model in this. And I can tell you getting them to show visually guided behavior is very, very hard. They're pretty close to blind. So again, that fits in a way with human phenotype that manonopsin on its own is very, very bad at supporting vision. Echo got much better outcomes with her preparation, interestingly. And that maybe highlights something that we need to keep in mind when we ask questions about manonopsin function based on rod-less cone-less animals. And that is that this is a totally unphysiological preparation because manonopsin is always supposed to be working in the context of incoming signals from rods and cones. So as an example, it's perfectly tenable to think about manonopsin really as a neuromodulatory influence that's regulating those signals coming in from rods and cones rather than a source of independent visual information in itself. And obviously if you get rid of rods and cones, then it's a totally different way that the system's working. So could manonopsin sensory capacity be different if we looked at it in the intact retina? And that's a challenge that we've set out to look at some years ago. And the obvious problem we have is how do we separate the manonopsin signal from that coming from rods and cones in an intact retinal system without first of all kind of removing some of those inputs. And the strategy that we've taken is to take a concept that we borrowed from psychophysics, visual psychophysics and color science that that's a receptor sign substitution. So many of you will know about this. I don't have time to go into great detail but just in broad principles, this is how it works. So imagine that you have two photoreceptors in your visual system and they differ in their spectral sensitivity. So I've just plotted the spectral response profile for photoreceptor one and photoreceptor two. You can present a narrow band light source that will very actively stimulate both photoreceptor one and photoreceptor two. And you can also present a different stimulus that will very actively excite photoreceptor two but not photoreceptor one. And if you adjust the intensity and wavelength of those two lights, you can achieve a situation where they provide exactly the same excitation for photoreceptor two, but A is always going to be much more active, much brighter for photoreceptor one. And substituting light A for light B therefore is silent for photoreceptor two but represents a big increase in effective brightness for photoreceptor one. And the outcome is that you record responses elicited by photoreceptor one. This is a very robust phenomenon. It's the basis of RGB architectural visual displays for example. And we can expand it to three or more photoreceptors by adding spectrally distinct light bands. You can use it to allow independent control of effective brightness for any target photoreceptor alone and in combination. So it's very powerful. But one limitation or one of the limitations is it works best when photoreceptors have divergent spectral sensitivity. And that's why all of the mouse data that I'm going to show you use this really, really powerful model we got from Jeremy Nathan's lab which is an animal in which the human red cone opson is expressed in place of the mouse medium wavelength sensitive cone. So it shifts their cones in sensitivity away from melanopsin spectral sensitivity. The other really important thing here is that it's all about calibration. It's all about getting those A and B lights exactly the right wavelength and intensity combination. So I'm not going to bore you with the weeks and years and months of calibration that we've done with all these stimulus stimuli. They're in the published papers, but we do a lot of it. Okay, so how does it work? So this is work from Tim Brown when he was in my lab. Please say he's now a professor here at Manchester and a former speaker on this Sussex Visions seminar series. So what Tim did was to produce a system where we only have three types of LEDs. We have a UV, a blue and a green and we can independently control their intensity. So this shows the spectral power distribution of two of those lights at two settings where we're regulating the amount of blue light here so that when we're producing a lot of blue light that has a big stimulus for melanopsin. And to make sure that it's not also a stimulus for cone we adjust the red and the UV lights to compensate for the effect on cones. So we can produce a stimulus which is only visible to melanopsin. And we can have a control condition where we get exactly the same melanopsin stimulus but now we also allow cones to see an increase in light intensity by just having a spectrally neutral increase in light intensity. So what happens if you do that and you record from the mouse visual thalamus? And the answer is that your responses fall into two different response categories which we call melanopsin responsive or non-melanopsin responsive. And the melanopsin responsive units when you present an all photoreceptor stimulus have the sharp on excitation and then maintain firing when the light is present and then it decays afterwards. And when you do the melanopsin only signal you get this quite different sort of response where you first of all don't see a response and then it builds up slowly and dissipates slowly. Okay, whereas the non-melanopsin responsive units don't respond to the melanopsin only stimulus and only respond transiently when the light goes on. And we get them at a ratio of about one to two. So about a third of the units are melanopsin responsive. And of course we can check in melanopsin knockout mice whether to make sure that our stimuli work and again in melanopsin knockout mice we don't get a melanopsin only response. So in the dorsal atrogeniculate nucleus of mice sustained units are melonic to responsive and visually intact animals. The melanopsin component responses qualitatively distinct because it's sluggish and sustained. Okay. So having established that principle we've gone on to use it quite a lot. And one of the things we wanted to do was to develop a more sophisticated iteration of it. We wanted to add additional wavelength bands so that we can control rods as well to make sure that we weren't recording responses from rods at all. And then also allow us to present patterns stimuli. And Frank Marshall who's an excellent workshop technician developed this thing based on a DMD projector system where we swapped out the light engine and now can control, can present images with five different wavelengths of light. And Annette Allen is one of the heroes of my presentation. She was involved in this work and lots of what I'm going to talk about now. And it's now a Henry Dale fellow here in Manchester. So when you do this and present melanopsin only stimulus looks the same as when you do it just with a three LED system. But what we can now do is do more complicated things. And one of the things we wanted to do early on was to ask not what melanopsin only can do but what happens if we provide a stimulus which is visible to rods and cones but invisible to melanopsin. So because we have a different wavelength sensitivity to our cone photoreceptors when we flip from two different spectral compositions they look different color to us but for mice they will look the same for melanopsin but want to look brighter for rods and cones. And when you do that you see if you like the flip of the melanopsin only response. So at early points in that step response the response is exactly the same. And the longer that the step is present the more divergent those two conditions become. So firing is lower in the mel less condition. That requires quite high contrast for melanopsin. So just keep that in mind we need rather big differences of melanopsin for in order for us to see that distinction. And that's gone on to look at the frequency preference of that melanopsin signal by using this binary noise stimulus where we're introducing steps at between 15 and 0.1 Hertz and looking at a cross-power spectral density analysis you can see that for our MR units first of all they like low frequencies very much but in the mel less condition they like them less. And by contrast the non MR units don't particularly like low frequencies and there's no difference when we present the mel less condition. Okay. So melanopsin makes a unique contribution to DRGN activity and maintain response under extended presentation. So now we want to use the pattern stimulus element of the projector and we can do that now by introducing patterned images. So here we have a bar that's only visible to melanopsin and we can use that to map spatial receptive fields. So when we did this we were relatively agnostic given the time frames it's totally possible that melanopsin signal just appeared diffusely over all LGN neurons or a large fraction of LGN neurons and didn't really have any spatial was not conveying any spatial information but actually it turns out that you can map spatial receptive fields very well for the melanopsin only stimulus and they match the size and location of receptive fields to a simple bar stimulus for that unit. So the spatial information the spatial receptive field for melanopsin is exactly the same as it is for rodent cone photoreceptors for LGN units. Then we can ask what happens with those bars not visible to melanopsin. And again, if we compare the responses this is a representative unit for all photoreceptors where we get a nice when the bar appears here we get a nice on response which is then retained as long as the bar is present that's much less clear when you have the melanopsin less condition where this sustained activity is much reduced. And that means that if you map receptive fields you get different answers depending on well, you get different amplitudes depending on where you are in that 10 seconds of bar presentation. So at early points, the melanopsin less and the all photoreceptors response are superimposable whereas if you look at later time points then the amplitude of that receptive field is much smaller in the mel less condition. So melanopsin enhances spatial information under extended presentation. Okay. So what our electrophysiology has told us so far is that it's met you can there are measurable melanopic spatial receptive fields so it's feasible that melanopsin might contribute to the representation of spatial images in the LGN. And indeed it looks like melanopsin makes a distinct contribution to the visual response properties of the of neurons in the visual thalamus by enhancing response amplitude under extended view of higher contrast spatial patterns. So let's consider then why that might be useful for something more natural an active, you know, a natural viewing situation. And for that, I think we need to keep two things in mind about vision in about natural vision. The first is that natural scenes show strong correlations in local radiance. And the second is that there's an inverse relationship between the frequency and magnitude of head and eye movements across all species at least that we've looked at from the literature, including rodents. So what a net did then was to look at a simulation of what that might mean from the perspective of the receptor field of an individual DLGN unit viewing any random scenes. So this is a scene obviously of sky and ground with bushes in between. And what she's done here is superimpose the receptive field over time, the multiple placements of that receptor field across the image starting with the simulation where you've got small eye movements here or head or eye movements here which move the receptor field around this part of the visual scene. And then you have a big change and you have now the receptor field falls in the sky. But again, you have a more frequent lower amplitude changes in the direction of you which move the receptor field there. If you look then at what happens to the radiance of light within that receptive field what you find of course is that you've got through this phase here when you're mostly located here you've got lots of small amplitude changes in radiance. And then you get when you change your direction of view you get a big change in radiance and then you have an epoch here where there are smaller changes. So maybe then what melanopsin might be doing is not really tracking these small events here but helping keep track of what happens when you have these big changes here. In other words, the big patterns here between the ground and the sky in between eye movements. And that makes sense also when you think about the contrast sensitivity of melanopsin, of course the contrast when you look at analyzing the receptor field over time when you make big eye movements is much more biased towards high contrast than when you look at the between shifts and gaze epoch. Okay, so that's how it could work. Do we have any evidence that it does work that way? And to get at that what we did is ran a simple experiment where again we're recording from the LGN in mice and we envisage scenario like this, right? So this is our neighboring building here in Manchester but you can see how it's got patterns throughout but they differ across the scene in how bright the overall region is. So you have an area of high brightness, mid brightness and low brightness. So this is a very typical thing that might happen in visual scenes. So what we can do then is present to ironetitize mice the simulation of the eye moving to look at different parts of that scene by taking an image that has this characteristic and just changing its overall brightness. And we do that every one, five or 10 seconds. So to present a stimulus then that looks for the photoreceptiles like this. So all the time there are these high spatial frequency elements and throughout the recording Annette's shifting these a little bit to stimulate small eye movements. And then on top of that she's putting these changes, these shifts and changes of direction of view which are shown here then which give you big changes in radiance. And in this case because it's a spectral neutral change they're visible to all the photoreceptors. But of course we can do that. We can change the spectrum rather than the brightness and we can produce a vision that's not visible to melanopsin. So in terms of rods and cones that got these changes in background but melanopsin thinks it's exactly the same or conversely how something that's visible only to melanopsin but not visible to rods and cones. So the whole time there's this visual image and it's changing a little bit to simulate small eye movements. And at the same time we're changing it as if it was as if the eye was shifting to look at different scenes that were differing in brightness. Okay. So what happens when you look at the melanopsin responsive unit population in response to this stimulus. And if you look at the all photoreceptors trace here you can see that they're clearly responding to elements of the stimulus. And that's true also for the melanopsin less and for the melanopsin only but I think you can immediately see that neither of these conditions do as well as the all photoreceptors. And so what Annette did was simply look at the correlation coefficient between stimulus radiance or nominal radiance for the melanopsin less and melanopsin only condition and the firing way of neurons. And if you look at the rate look at the melanopsin responsive units you can see that that's much better for the all photoreceptors and pretty much equal for the melanopsin less and melanopsin only. So melanopsins required to have this good tracking of that stimulus. And even without rods and cones you can still get reasonable tracking. Conversely, if you look at the non-melanopsin responsive units first of all, they don't track these large sustained changes in radiance very well. And so you can compare this to this but also there's no effect of melanopsin. Okay. So we're gonna leave the mouse electrophysiology here but just to summarize what I've told you melanopsin compensates for relaxation of rod cone signals allowing a better representation of spatial patterns in brightness in the mouse DLGN within a reasonable framework for pattern vision. So what might we predict then if we wanted to move to recording visually evoked behaviors and to try and work out how melanopsin might contribute to visual behaviors? We expect that melanopsin could contribute to discriminating core spatial patterns and during periods of visual fixation. And so those are the predictions that we want to try and test by doing some assessments of vision. So how could we do that? So it is possible to record a melanopsin contribution to vision in visually intact mice. And we did this back in 2012 based on this paradigm where we had mice in a swim maze and trained them to swim towards a brighter it's just a trapezoid swim maze. There's a false choice here where they can swim to get to the escape platform and we train them to swim towards a brighter window to find that escape platform. So that's the training epoch. And then what we can do is change the wavelength of those two lights. So in our probe sessions, one lane is always red. So that's gonna be very bright for cones and moderately bright for melanopsin. And then we present as the alternative choice a green panel where we can alter its intensity. So the bias for melanopsin is always defined green, brighter than red. So we'll expect if melanopsin is relevant we'll expect them to have a bias towards swimming towards the green panel, right? And if we then plot the percentage at which the mice choose the green channel as a function of the brightness of that green panel no surprises that if the green panel is very bright then they reliably swim there. And if the green panel is very dim then they reliably swim towards the red. What's interesting is the point at which they judge those two panels to be equally bright is different in our red cone knock-in mice compared to red cone knock-in mice that lack the melanopsin gene. So that loss of melanopsin gene is changing the spectral sensitivity of their preference of their judgment of brightness which we think is pretty good evidence that the melanopsin is helping those mice to judge brightness in this paradigm. So it is possible then to find visual responses in mice that have a melanopsin signature it is also very, very hard work. It's hard work knowing how asking a mouse what it could see and we were lucky early on to establish a collaboration with Sichi Sugimura who's a psychophysicist who was doing similar work in humans. And so what Sichi has done here is generated as kind of a spot stimulus to spot a large spot stimulus. And he's asking people to judge relative brightness against a reference and a test stimulus and the test stimulus he can either add in melanopsin effect of brightness or not. And what he finds is that the point at which people judge those two stimuli as being equally bright. So that's that 50% point here is dependent on how much melanopsin content that the reference stimulus has. So evidence that in humans they judge large field light fields with higher melanopsin stimulation is brighter. And this is something that's been reproduced now many other labs and proper psychophysicists are looking at this and have looked at the relationship with luminance and exactly how it might fit into models of brightness perception. But what we wanted to do in view of what we had from the mouse data that was coming through was to move towards being able to present spatially structured stimuli to human subjects to look at how melanopsin might contribute to spatially structured vision. And generating a version of our melanopsin projector system for humans is difficult. And the strategy has to be slightly different for various reasons. But what we've done is again start with these data projectors, these projector systems. But in this case, we've taken two projectors and we superimpose their image and we filter the RGB channels of each of them so that instead of now having two images each of which have three channels RGB we now have two projectors each of which have three channels but of different wavelengths. And so we can end up with a system like this where we've got a band here in the deep blue something at cyan, something at green something at yellow, something at red. And that gives us the degrees of freedom then to modulate it selectively the melanopsin brightness of any pixel within that image. So here are two spectral power densities that are melanopsin high and melanopsin low and they appear identical in terms of hue, saturation and luminance. That's because they identical for cone photoreceptors. And again, this is to remind me that lots and lots and lots of calibration to make sure that this works particularly in humans not only between individuals but also between parts of the retina. It's very important that you keep that in mind. Okay, so one of the first things we'd like to do with that is just to present some melanopsin images and see what we see, right? And the answer is it's pretty weird but there's definitely something there. So this is a kind of a documentation a documentary description of our sort of subjective experience. So what we've done here is present two versions of a standard image here. And in all cases we ask people to fixate here and we fill out this kind of central region of the image. Happy to explain why questions. And then image one doesn't, it's not just has the sort of standard melanopic contrast between elements of the image. And image two is augmented for melanopsin contrast. So the bright bits of the image look really bright for melanopsin and the dim bits look very dim. But for combs, they're exactly identical. So photometrically, these are identical images. And when you present them to people, this is like free-form descriptions of what people say. And it's a sort of a sense of them being brighter but more distinct, more sort of stand out. And then if you ask people to rate which of those versions is more distinct, they reliably report that the high melanopic image is more distinct. So consistent with our view then that there's something about patterns with this stimulus that melanopsin is revealing. Okay. I think I'm gonna whiz through this just to say that one of the other things you can do with this melanopic stimulus is you can ask whether it regulates sort of circadian and non-visual functions as well. And so we worked with Christian Kajoken at the University of Basel and had people going into his lab in the evenings just watching movies either in melanopsin, lower melanopsin, high condition. And as over the course of the evening people tend to become more sleepy but they became less sleepy when the movie was rendered in melanopsin high. So augmenting melanopsin was keeping them more awake and also it was tended to suppress their melatonin production. So the device can show that the control of melanopsin can be used to adjust the impact of long-term light exposure on sleep alertness and melaton expression. But let's return to our question of form vision and what Annette went to go on to do next was to present more kind of calibrated stimuli to ask what sort of thing that humans might be able to see if we're in melanopsin. So for these first experiment, what she's done is again had a fixation point here and then somewhere in the periphery, we put this grating and that grating can exist in one of four directions, four orientations. And we asked the subjects to tell us what orientation that grating was, okay? This grating is visible only for melanopsin. So as far as the cones are concerned, this is just a continuation of the background field here. So it's a melanopsin only stimulus. And when you do that and you ask people whether they can see it, people get significantly better than chance at judging the angle here of this grating. On the other hand, you can see here that at no point are they getting at 100% correct. So it's not easy to see these things. So it's more of this feeling that there's something there. Okay, so the next thing we can do is we can change the grating frequency and ask what happens to their ability to see that. And when you do that, you can see that really we need pretty fat bars, pretty thick gratings in order for people to be able to resolve that melanopsin only stimulus. So it's only when we have very low spatial frequencies that we can see that stimulus. As a control, we can do exactly the same thing for an all photoreceptor stimulus and that has a much more familiar spatial frequency preference where you get optimal performance here about one or two cycles per degree. So one last thing we can do here or one of the other things we can do here is adjust the contrast, the melanopic contrast of this grating. So not just have it at maximum contrast but start to try to break it down. And what you see is that you need relatively high contrast stimuli in order for this to be visible. Whereas in the orange here for an all photoreceptor stimulus, I mean, basically we run out of resolution for our projector. Any all photoreceptor stimulus at the lowest contrast we can present is visible to the subject. So it's much less contrast sensitive than the rodent cone system. And that if you remember makes sense, right? Because if we looked at our simulations of natural view of visual scenes, we're expecting first of all that melanopsin is gonna be doing is gonna be particularly useful for low spatial frequency patterns and also that it can be biased towards higher contrasts. Sorry, lower contrasts. Okay, so melanopsin only patterns are discernible at low spatial frequency and high contrast. So the last work we've done on this is a work from Tom Boulders in the lab. So Tom's been working on this troxylaphating paradigm which relates to the question of how melanopsin works under extended view. So some of you will be already familiar, I'm sure with this visual illusion. So what we need to do here is focus is stare at this man in the boat but keep attending to the sun here. And this is a painting from Claude Monet. And this is a phenomenon that was first described actually or first documented by Erasmus Darwin. And what it is is you should start seeing that the sun start fading away, so you stop seeing it. And this is the troxylphating. So the fading of elements of the scene which are under steady view. So if we think about what we asked, what we thought melanopsin might do based on the electrophysiology, this is the type of visual illusion where we might start seeing a melanopsin contribution, right? We might think that having more melanopsin would delay that troxylphating event. And so Tom's looked at that by presenting a stimuli such as this. So again, we have a fixation point here and this time in the periphery, we get this kind of fuzzy blob appearing in the periphery. Yeah, it's a bright blob. And we just ask people to fixate on that and tell us when that blob disappears. And the answer is that the rate at which that blob disappears is dependent on how bright the spot is, right? So if you increase the contrast, the luminance contrast on the spot takes longer for the spot to disappear. Now we can introduce the melanopsin dimension and take a spot which has the same luminance contrast but introduce melanopsin contrast. And we find reliably that the spot takes longer to disappear. There's quite a lot of inter-individual variations. So in some ways it's easier to see it as a Z-score and you can see here that for each individual there's quite a substantial difference in the duration that spot lasts if it has a melanopsin component. So it works for bright spots, it also works for dark spots. So this is the converse stimulus where we're asking people to tell us when this dim spot disappears. And again, that depends on the contrast between the background and the dark spot and adding melanopsin contrast to that has the expected effect. So again, it takes longer for the spot to disappear if we have a melanopsin contrast component. Okay. So introducing melanopsin contrast slows image fade. So that's really what I wanted to tell you and to put it together in terms of like linking I think what we see in terms of the mouse electrophysiology and the human visual test, visual experiments. I think we can type think about a visual scene such as this as having patterns over multiple spatial scales, right? So from the very fine detail here to ultimately to just the overall amount of light in the environment with everything in between. And what we find is that once the further you move into the low spatial temporal frequency range the more that it looks like melanopsin is important. And that melanopsin does seem to make a distinct and uncompensatable contribution to the representation of these lower spatial frequency elements of the image. So you can almost think about it as melanopsin allowing cones to focus on the very high frequency elements here by taking care of the lower frequency components and allowing us to have visual fixation by allowing us to by keeping track of the course patterns while we're fixating on small elements here, okay. So what else is interesting to do here? I mean, there's so many things but just a couple of things that I think are really cool to be thinking about. One of which is I think that we still need to test the limits of this model in different scenarios looking at different species, for example but also in different types of view. So I think that's really important. The other thing I think to keep in mind here is that we're really asking melanopsin to do a lot of heavy lifting, right? Because if we talk about temporal scales we think that melanopsin is important in telling the difference between day and night. So that's a really long timeframe. Maybe even times of year, a really, really long timeframe. At the same time, it's also contributing to those differences in brightness that occur within the timeframe of shifts in visual, in gaze, in gaze across the visual field which is much shorter. It's also asking it to cover a massive differences in brightness between day and night time or across that dust to daytime transition and also the much smaller things that happen within a scene. So it's really, it's doing a lot and it'd be really nice to start relating that to the cell physiology and emerging evidence of signaling complexity for melanopsin. So this is a summary of what we know about the IPRGC population from Contreras et al in a very recent review and the upshot is that at least in mice there's probably six types of melanopsin methaml ganglion cell and they had distinct projections and distinct physiology. Really nice to start relating that more directly into their role to covering this range the spacia temple range of vision to which melanopsin might contribute. We also have surprising complexity in the signaling of melanopsin within the cell. So melanopsin is a G protein coupled receptor, right? Light activated GPCR like rodent conosins and it was thought for a long time that it was activating a G alpha Q signaling cascade much as happens in invertebrate receptor receptors. And that certainly is the case but we've shown in cell culture that melanopsin can also very efficiently activate G alpha I and G alpha S. And we know from the work from Tiffany Schmitz group and from King Waiow's group that actually you can get melanopsin signaling without G alpha Q pathways. So it'd be really interesting to think about how this signaling complexity also fits into this question of how melanopsin can cover this big spacia temple range. Returning to melanopsin modulatory influence on vision I think is also a really nice thing. So how is the visual code changing according to that melanopsin signal? And finally be nice to see this applied in the real world. So just disclosure of interest. We have some patterns about designing melanopsin displays but leaving that aside, all of our visual displays and image capture architecture at the moment assumes that everything about vision can be encompassed in our three cone photoreceptor system. That's why we have RGB displays. If you're building a new type of visual display architecture it seems crazy not to also include melanopsin, right? So I'll finish with acknowledgements. We've been very lucky to get funding from the ERC the Welcome Trust and the BBSRC. I've mentioned some of the contributions from people in my group in Manchester while I've been speaking but it's really a massive team effort and all of these people have made big contributions as have our collaborators around the world. So thank you very much for your attention. Thank you very much Rob for this very interesting talk and your attempts to elucidate what is the contribution of melanopsin in vision in natural retina impact-wise when it comes to the input and naturalistic settings with what you try to take it like experimental-wise. There are already some questions appearing in the chat and I would like to remind to our audience that they can either post their question there that will moderate and communicate it to Rob or you can join us in the Zoom room that we are currently using and I will be posting the link right now in the chat because in five to 10 minutes depending on the question I will be terminating the broadcast and we will continue offline. So the first question is from Gregor Belusic. A high and great word. Maybe I've missed it, but how can you distinguish between human melanopsin vision and the possible effect of rod vision? Do you assume rods are saturated that they use light levels? Yeah, so that's a really good point. So in this, whereas in the mouse for our electrophysiology we have enough capacity to have stimuli which are also rod silent. For the human work we have we also, we're working at a light level where we hope the rods are approaching saturation and we realize that this is imperfect at least for the visual display work that we've done you can of course have stimuli which would also be rod silent but then you have very little contrast. So I'm sorry, melanopsin contrast. So it's almost like we put it together as the way that we think about it I suppose is that the work from the mouse poses hypotheses and those hypotheses we can then, we can then test for the human experiments but it's a sort of a conceptual justification that rods will make little contribution rather than a direct demonstration of that. So that's totally a fair question. Thank you. And the other question is from Wei Li is M4 the main conduit for melanosis contribution to form vision? So I think that's a great question, right? And I think that, you know, I think what's really important is to keep in mind is that what we're doing is looking at a systems level. So for us, we're totally agnostic about where this comes from in terms of the, in terms of the, in terms of the melanopsin response. What we can say is that, you know, we're seeing a melanopsin contribution in the mouse LGN to maybe a third of the LGN units. And obviously the M4 IPRGCs are a much smaller fraction of the total ganglion cell population. So at some level, we need to explain that discrepancy, right? And we don't have an explanation for that yet. So does that mean that those M4 cells have lots of projections that it's a massive expansion or that maybe some of the other IPRGC types of contributing and how that whole thing fits together? In terms of the phenomena that we've described, we don't yet have a clear link to an individual IPRGC type. I mean, in terms of projections to the LGN, it's not only the M4s that go there. So it's definitely possible that other melanopsin classes could contribute. And before I move on with the cross species questions and what expectations do they generate, like your findings generate or justify for different diets, some questions that I have, like if I really apologize, my ignorance, but when it comes to primate vision and retina, do we know anything about the density of IPRGCs with increased eccentricity? Yeah, so there's really nice work done by several groups. And from memory, the upshot is that the outside the actual fovea itself, the fraction of melanopsin positive retinal ganglion cells sort of remains relatively constant. So that means that there are many more melanopsin retinal ganglion cells in the kind of periphovial region and then as you move outside, but that reflects the ovial number of ganglion cells in those different parts of the retinal. And the other question I had like stimulus wise, because you present this creating that this invisible to rods and cones, but melanopsin cells can see, is it possible to like slightly move this creating, like to introduce a temporal aspect to the stimulus without the rods and cones seeing it or this is impossible? Sure, no, no, definitely, absolutely. It's possible to do that. Yeah, you can do that. Yeah, you can do that. Have you tried to see how the responses of the melanopsin cells might change? So let me think. So you're looking for a temporal modulation. I mean, obviously with the receptive field mapping stimulus, you know, this is the bar moving in space over time along the azimuth. So that sort of does that answer your question or are you looking for something? It could be, I'm just interested because like, I mean, and these are questions that I will try to ask later, but like because it looks like the melanopsin cells have like very long glantensis compared to what you would expect from rod and cone input. I was wondering what temporal frequencies they can encode when they are the only ones taking it. Yeah, so that would be, I guess the way that we approached that was with that binary modulation stimulus where we varied the frequency of that and it looked like and from that we estimate that frequencies like less than one hertz is when you start seeing a melanopsin contribution. But I think actually there's more to do without because that's not the same thing as a kind of a sinusoidal modulation, for example, modulation and that's something that we're kind of working on at the moment. So I can't give you an answer to that. Great, thank you very much. And like now I will proceed with the questions that Tom posted and if there are no more questions appearing from the audience after these two and the answers from you, I will terminate the live broadcast as we will continue offline and some people are already here. So the first question that Tom asks is, so in that view doesn't this imply that melanopsin should be relatively more important, the smaller the eye and that's the word, the special resolution. What about animals with tiny eyes? Right, well, so there's a whole visual ecology question built in here as well, isn't there, in terms of what's important about vision. And the reason that I say that is the extent to which we looked at cross species with this in terms of electrophysiology is we've been working with a raptomus which is a diurnal rodent that has a cone rich retina really massively expanded visual system. And for them, melanopsin seems to first blush much less important for vision. And that's probably because they're performing different types of visual tasks to lab mouths. So I think that would probably be the bigger determinant of how important melanopsin is. But and then so it's to do with how probably to do as much in terms of the spatial domain, the temporal domain, right? How much having that image stabilization vision is important for them, I would guess. So but really, really interesting things about how it varies. You know, that's one of the things I wanted to get across with is like probing this hypothesis with different species, I think is really, it's really a cool thing. And the other question of his, which takes into like takes it into more evolutionary context. So many normal mammals have had melanopsin in earlier retinal neurons, including before the on of split. Would one therefore expect of melanopsin signals and what would they mean? Um, I mean, in principle, well, blimey, that's that's a first of all, that's a very difficult question because it's pure speculation, right? But I think, you know, if you look at melanopsin outside mammals and you go into into number million vertebrates, first of all, it fits in the picture of lots of different types of of options, right? Not just melanopsin, not just one type of melanopsins, multiple types of malopsins. When you get into fish, I don't need to tell you guys. And then pinopsins and para-pinopsins and VA options and like to pick the bones out of that because presumably they're all doing something slightly different in terms of the sensory capacity of the fish. I would imagine, given that, you know, why does what is it about melanopsin that means that it imposes a relatively low sensitivity? So it's pigment density is much lower than rods and cones. So we're all going to have relatively low sensitivity and or low spatial temporal resolution. So you think that melanopsin is going to be doing that wherever it is and if that's in off bipolar cells, it's doing that for off bipolar cells, you know? So like it's an unsatisfactory answer, isn't it? I don't know. I mean, we will see Thomas here. So hopefully she will elaborate and continue down that road soon. So the last question that I have before we terminate the broadcast and continue offline is if I remember correctly, from my retinal lecture, so the melanopsin is a bi-stable option, right? So it can. OK. And that is, anyway, carry on. This helps with evading saturation, I guess, but it doesn't help with the latency, right? So the latency is dependent on the later biochemical pathway. Yeah, the fact that we have like this maintained response of an extended stimulus presentation could be just that the melanopsin signaling kicks in later. So this is why it gives the impression of. I say made like maintained in that sense. So Michael Doe, of course, and Harvard has done lots of really interesting work on this. So I'll just try and channel that. And I hope I don't mess it up. But the I think the the maintained activity then is to do with. Well, we know that melanopsin can be switched on and off with light. The rate of that happens at the light intensities we're talking about, I would argue, is relatively small because there's another thing that definitely happens, which is melanopsin has a dark regeneration. So that's definitely. OK. And it also has a biochemical signal termination. So first, Robinson has shown very nicely that you get phosphorylation of the melanopsin C terminus, and that's important for switching it off. So I think this ability, this bi-stability, tri-stability is a really important phenomenon and feature of melanopsin. But probably the things that we're talking about, it's you can think about it much more like a conventional photoreceptor that the rate at which it switches on and switches off is to do with the buildup of the second messenger systems and the rate at which the the the the the signaling cascade switches turns off and turns down again, scales up and down. I see. Thank you very much, Rob. And thank you in general for giving this this talk. And at this point, I will stop the live broadcast so we can continue offline with people that are interested. Thank you very much. My pleasure. And we are a few.