 So much into neuroinformatics. So This is one problem I have and the other problem I have is that I'm going to talk about a system that is not that much covered by The neuroinformatics community and in fact when Gordon Shepard showed his wonderful databases I realized that the cell times I'm working on Weren't even included. Well, actually the region I'm working on wasn't even included in the cell types So I will take this opportunity then today my so-called 22 minutes of fame To try to convince you that there is much more beyond the cortex and much more beyond the cortical processing of sensory information That might be Considered if we want to understand brains and behavior So I'm going to talk about the midbrain vertebrate midbrain mostly the midbrain of birds because this is what I'm working on But I will relate also to some mammalian studies Self-circuit concept. So I will first talk about the role of the midbrain in behavior Then I will shortly introduce the anatomy and connectivity of the visual midbrain and then Describe a feedback loop in the midbrain the tecto is mcnet work that has received a lot of attention recently Especially because Eric Knudtsner has picked up that Subject in the banal and I will try to convince you that it might be Relevant for a lot of the things that have are going on in the cortex as well So I will rather talk about the concepts behind it and not so much give you data on the system because I am Convinced that most of you are not familiar with the system So I don't want to flood you with a lot of anatomical and physiological data in that respect Okay, so as car already pointed out An essential thing for behaving systems is that information is not a passive thing that is impinging on them But they are actually actively looking for information and especially in the visual system It used to be that your brains were regarded as more or less passive Analyzes of what is going are coming in from the retina. Well, in fact, it's rather depicted in that image Animals and brains are constantly updating constantly searching for information constantly looking around and trying to well Analyze the most important feature an object that is around and this is important for survival as I will show you in a few seconds It is even more difficult because for many animals the The image that you see at a given point in time is not like this But it's rather like that and we all know this because we have a phobia retina So the area that is actually in good focus and it's you know can give you a lot of data Is very small much smaller even than this so what we of course do is we look around so we have to Gather the contents of our surrounding by looking around for making cicades and we all know that but in fact It's often neglected if we look at how people try to analyze visual Processing so we look around and this is easy to do this is easy to do if we know what we're looking at So for example if you look at a face and you probably all know that face if we look at a face We exactly know where to put our eyes on where to put our phobia on and this is just depicted here So it's mostly in the eyes and in the mouth and especially you know This is the area of the face where we know that information will come from which is relevant for us However, if we look at something that we do not know or we have no concept of Where do we then put our eyes in the next picture? I'm going to show you is one of these pictures that probably most of you know But if not you might experience that you will kind of frantically search around for a concept of what you're looking at. Okay, here it is Okay, well most of you probably know it, but those two do not this is a damnation dog And if you do not know what there is, okay? You will have a hard time and you will kind of frantically look around trying to make out what is there Okay, this is very relevant for survival and The question where to look Has certainly driven evolution and this is the last introductory slide. I'm going to show you So I'm going to play a game with you So I will show you that picture and as soon as you say as soon as you think you see something that is relevant for your survival Please raise your hand Three two one go One two three, okay, I see a lot of you would probably not survive in the wild Okay, um, it's interesting isn't it? It's like an arms race in nature, right? I mean, of course the predator tries to not be seen, right? Okay, but you will I think agree that it's important to see these kinds of things so in order to put our Distance senses onto that important issues on that important objects We have to answer the basic question and this is how can we look at the most important object when we do Not know where and what it is and especially if our surround and the non-phovial region of a retina is such lowly resolved And actually lacks color in all that. Okay, so how do we do that and we do it by? Making use of mechanism that extract basic information in our non-phovial surround Which is motion signals looming signals something on the collision course towards you strong color and and contrast and of course also Non-visual cues like touch and sound so for example if that wolf would have moved and something How that you know sounded it would immediately have looked there Okay, and this is what I'm actually interested in and this is something that is going on not in the cortex but at least to a great degree in the midbrain in the midbrain of animals and This is why the title of my talk. Oh well actually the real title of my talk is multimodal mid bottom up mechanisms in the midbrain of birds Now why do I do it in birds the reason is that in birds the system is some from a well analysis analysis point of view very Nicely laid out if we look at the midbrain here depicted in blue This is the optic tectum if we section that we have the structure like this It's very large in birds and you can even see with the naked eye that you have a lot of laminar here And if we look at greater detail, we see that we have 15 layers that can easily even a lay man can analyze I can can differentiate between these layers and in these layers. We have all different types of cells and What is even more important? We have a clear layout of the input and output structures, and this is something that is much more How much nicer than in the mammalian counterpart in the superior colliculus because in the bird the And actually in all vertebrates the input the retinal input comes in via the top layers goes down here Makes connectivity or max connections here in the outer seven layers And then it's relayed on to interneurons and then finally to projection urine that project on to with the tectofugal Pathways with the forebrain and we know a lot about that system because this system fortunately for us in the cheek Has been a model system for developmental studies So we know a lot about the molecules lot about the gradient that you know establish the visual maps and so on and we also know a lot about the cell types in there because Numerous studies have been done with both intercellular methods and Golgi methods even going back to Ramanica how? Okay, so we have a pretty good idea of what the connectivity is like what the cell types are like also what the molecules are like and The real good thing about that system is that we pretty much know what it is doing Because what it is mainly doing is to construct a map of the sensory surround This is clear for the visual system So any object in space will lead to a focus of excitation in that tactile space map And it's also true for the auditory system because a an auditory stimulus from that specific position in Space will be computed and also lead to a Focus of excitation in the space map and there are actually bimodal neurons And this is a pretty hot topic to understand how these two Senses are put together there a lot is known about the well the the alignment of the maps Eric Knudsen has done wonderful work on that in Barnolds, but less is known about the actual bimodal integration on the level of the single nearing Now it's true not only for the visual in the auditory system But also for all other systems that you may happen to have so if you're a rattlesnake you also have Infrared system and that is also projected onto that map and it's the lateral line There's some as a sensory system and so on and so on and also quite important There is also a motor map and that motor map is contained in the deeper layers so the deeper layers receive information from that spatial excitation and then Illicit orienting movements, so it's something that well actually 40 50 years ago was termed the Sensory reflex grasp or sensory grasp reflex actually it was called and it kind of you know Explains quite nicely what the system is doing Now a lot of work has been done on that Subject however, what I want to talk about today is something that is Actually making the problem more complex and this is At any given situation you are not only faced with one object But what you have is you have a multitude of objects and this multitude of objects. Oops, uh, whoops, where is it? Yes, well of course leads to many Focus on many areas of excitation within the spatial map So the motor system now is faced with a problem how to pick up the most important Object and that goes back to what I said in the beginning. So what is the most important object? How can you pick up? How can you resolve at that level what the most? Important on the most salient object actually is because for the motor system this Excitatory map of the system is bad. Okay, you definitely need a Disambiguated map you need a clear excitation focus in order to make a clear Movement, okay, the motor system cannot deal with you know multiple options. It needs one clear signal all right, so this is the question that I'm going to tackle today and A possible solution to that is a system that had been discovered Well almost hundred years ago, but has received a lot of interest recently and this is Actually studied right now all over the world Shorong Wang in Beijing was one of the first who started it Juan Carlos letellier Gonzalo Marina and how come the doors is in Santiago Ralph Bessel my colleague in St. Louis and me we have been studying that in chicks chicken and Eric Knudsen has Picked up that like eight years ago. So working in the bar now But actually most importantly the founding father of it all he was in Spain And that was again Santiago Ramonica hall I actually rather his brother Pedro Ramon who is usually neglected But he was the one who did most of the non-mammadian stuff that Ramonica hall then publicized so well and the part this is actually on the optic tecton of the bird of the chicken and The cells I'm going to talk about today are these So these are two of the elements of the network that I'm going to present you and Actually, this cell type is called the shepherd's crook cell type It's not in relation to golden shepherd But it's simply because the accent origin is like a shepherd's crook like a shepherd's star for those in German, okay, so the good thing for us is that this network is contained in the slice preparation So I'm mostly working in slices because it's a very convenient way of analyzing cells and connectivity and stuff So the good thing is that we do have the Tectal layers and then we have the ismic nuclei here. This is the ismic nuclear area We have three of them and the connectivity with these has been well worked out And it's shown here, and I will walk you through in detail. So these are the tecton layers These are the outer layers the tecton the retinal aphorins These are the interneurons and then here this big layer. This is the projection urine that projects towards the telencephal Okay, and now from the middle layer. You have this specific cell type This is the the the shepherd's crook Shell type and it projects to all the ismic nuclei and now you have several Motives of projection first of all you have the type where the shepherd's crook project onto the nucleus and from there Projection goes back at exactly the same position where the input came from This is a cholinergic feedback. There are two nuclei the SLU and the IPC the names doesn't matter don't matter but they go back to exactly the same position homotopic and Project there with cholinergic efforts Now you have the same cell type making output to another nucleus, which is the magnesolar nucleus, and this one is GABA Ergic now from the GABA ergic nucleus magnesolar is you have two projections back to the tecton one is a broad GABA ergic feedback and That goes not to the area where the cell receives its input from but actually spares this area So it rather goes into the surround in this tecton spaceman And then to make things a little bit more complicated You have a second GABA ergic cell type in the IMC that does not project back to the tecton But projects to the other two nuclei inhibiting those okay, this is the basic network and Well, we started off by Trying to understand what the connectivity is like how the cell types look like what the you know The synapses look like but we have not really succeeded there and what the physiology of the system is like and This was the first approach that we did The good thing as it said again The whole thing is contained in the slice preparation So we have not in the entire tecton, but at least some part the entire connectivity Contained so we can you know go in stimulate the retinal afference Which is quite nice because we know that this kind of mimics input to retina Okay, like an object a moving object whatever and then we can stimulate record lesion Well do whatever we want to do with the system and try to work out what is Actually going on here the second approach is modeling and this is something that I don't do myself But I did it together with Ralph Bessel at the Washington University in St. Louis And the third that we just recently picked up is to image that system with Voltage-sensitive dyes to get a better understanding of what the network dynamics of the system are Okay, I will just give a few examples of the data or the type of data that we have gathered so far and then Come back and try to well wrap it up and and say something about the general function in the brain okay now for the Cell physiology as I said in the slice You can easily or patch onto the cells and record from the cells and well do whatever you want to so this is the Shepherd's crook cell type and you can see the accent here is Depicted in red and it's actually a very interesting cell type because the axonal origin up here Well, what does it mean for cellular computation in that cell type? It's actually a pretty interesting question isn't it so and actually Ramonica how already thought about that And I will show you a picture of that in a few seconds Okay, so what you can do you can just you know record and get the cell parameters and well you can give retinal input to the system That was done here. This was a 500 microsecond long Stimulus to the retinal input layers and you see that the cell starts of firing for several seconds and that appears to be a network Thing because you can actually hyperpolarize the cell in between and it resumes firing So these are the types of data that you can gather in the system And of course you can also get the the latency is okay Stimulating up here seeing how long does it take for a signal to come down to here? How long does it take from here to there and so on and so on so you get all the parameters that you actually know in order to model the system which is membrane parameters the neurotransmitters are known and Well the IFs relation the spontaneous activity, which is actually only here in the SLU and the response latencies between all these different elements and And well we did that and at that time then Ralph Vessel took over and modeled the system and What we are well we were lucky because just a little bit earlier. There was work by the Chilean group Gonzalo Marine in 2007 they showed something interesting If you record from the ismic system From one of these nuclei in the ismic system these cells have receptive fields Okay, so you record from the cell and has a regular receptive field which is down here position one So if you stimulate the cell by moving an object through that receptive field what you see is you see this rhythmic bursting of the cell Now this cell receptive field is just down here. It's not up here But if you move another second stimulus Through the receptor through the other area up here What happens is that the first response that you know to to stimulus one is suddenly abolished and the cell goes into a high Frequency firing mode, which is quite different from what it did before So the idea was is that the effect of the ismic network and can we reproduce that if we put everything that we know About the network in the model and this is what Ralph did. So we did a rather simple model which Contains the the shepherd crook cells up here with a retinal ganglion cell input the two different types of Gaba-ergic magma cellular cells and the feedback from the one Collinurgic feedback nucleus the other we did not put into that model and What we did is we so this is the blue cell type here in the attend cells These are the two Gaba-ergic and this is the IPC neuron. So we gave input at a specific position here Okay, and after having done that well when we do that the system goes into a rhythmic Feedback mode, but if we then introduce a second stimulus at another position here in the 300 simulated input layers Then we get a switch from activity at position one to position two and that switch Is exactly when we introduce the second stimulus and the second stimulus Has to be stronger. It has to have a specific strength in order to remove the reverberating excitation from the one locus to the other locus, okay, and this we found in in our network and this is exactly what Eric Knudsen found just recently are Actually was 2010 on nine even I think when he was looking at tactile neurons and looking at the response of Neurons in the optic tectum and then having a distractor Next to that receptive field of the cell he was recording from and the distractor actually had oops The distractor actually had to have a specific strength in that case a looming speed in order to make something that he called a Switch like response in the tactile cell So before the cell would simply ignore that there is something else in the surround But when the distractor had a specific strength the cell was suddenly cease firing, okay? And that actually fits very well with everything that has been discussed in that system so far So that there is in fact a system that is weighing all the different inputs that coming on to the tectum simultaneously and well You know voting out the strongest stimulus in order to get a clear focus of excitation in the map Okay, Eric has actually done quite a lot on that But well he published that recently in a series of excellent papers. So I don't want to go into detail here Well, what we try to do is to go beyond the single cell level or actually to go beyond the physiological level and look at the Network dynamics and what we did therefore was to go in into the slice again with voltage-sensitive dice and Simply first to see what the response of the network is like when we stimulate at a given retinal or tactile position With an electric stimulus and we did that here so this is the control situation this is the situation with a GABA block and Well, the only thing that I want to show you here It's just that first of all we can monitor the network activity in response to an input at a specific position And second that apparently the entire system is under a strong ever-urgic block most of the time Which might not be that interesting on the other hand the same type of experiment has just recently been done by a Vocunet L in the recipe recalculus and there it looks exactly the same So apparently a stronger biurgic block even in the slice seems to be something that is Conserved between vertebrates now, of course the the question that we really want to answer is what happens if we have to competing stimuli Okay, and of course we did that but of course we're not done so this is something that a student of mine has just been doing and He is currently just analyzing the data. We're just not done yet. I'm sorry about that The only thing I can tell you is it's much more complicated than we thought not only to get you know decent and reliable data, but also Well removing the ISMIC nuclei does not have the strong effect that we thought you should it would have Okay, so we are currently trying to make sense of that It's there's another level to the story that apparently we are not grasping at the moment Okay, well, this is um, well, I can't tell you a lot about that, but the other thing That I want to to make a point of is that there is even though somebody said that well You do not have to look at all the details. That's of course true And you always have to try to you know get an understanding on a specific level. However, there are so many I Would say devilish details that at least I want to hint at some so for example the feedback From the IPC to the tectum here this sign at this is one sign at this is one Terminal field from one IPC neuron so that neuron has a very thick axon that bifurcates and trifurcates And so on and so on and makes zillions of synaptic inputs in well Properly not onto a specific target neuron, but rather into cartridge flooding that cartridge with as it's a co-line And how to model that and how to understand what this is actually doing we have no clear idea Well the cells that do make these efferents they look like this and well I come also from the auditory system and there this is something that is often associated with Coincidence detection because you have a bit of the dendrite dendritic field so you can now compute between inputs on the two different sites of the Neurin Well timing seems to be important in the system Okay, so maybe well, this is something to be Analyzed and so on and so on and so on The shepherds crook cell for example is interesting This is a drawing from Aramani Kahl and he at that time he wasn't possible this guy He actually speculated upon information flow from here to the axon down here and the reason behind that is these might be the cell that actually Combine auditory and visual information and visual information reaches the brain much You know later in the auditory information So that might be a delay element in order to get coincidence detection here Okay, which is interesting in its way, but well in order to study that and to go to the cellular level we have taken to Transfection and well, I won't go into detail. This is what we can achieve right now But this is not the cell type that we want to work on so we have to improve our vectors to to get a well cell type specific Transfection and then we can try to Look at these cells. Okay Well, the mechanism is not restricted to the visual system Okay, the mechanism is not restricted to the visual system It's not a visual subsystem and in general if one thing you should take home from this lecture The optic Taktim is not a visual system. The optic Taktim is a multimodal system It's only involved and only you know cares about spatial spatial relation to the animal Okay, it's not visual. It's dominated by vision, but it's not visual. Okay, so everything coming from there is multimodal Okay, and the last point I want to make is well These are all bottom-up things So this is good if you walk around into a room and in the woods and have no idea what to look at However, there are of course often situations where it's very important that you do not lose your focus and in order to do that you need to Modulate these mechanisms. Okay, and there's also a beautiful study from the Knudsen group and they have looked at another Spatial map in the forebrain the AGF map and To make a long story short if you stimulate in that AGF map in the forebrain map What happens is not that the animal is now attending that position But what happens is that the tactile circuitry is now modulated in a way that the sensitivity of the cells at the corresponding locus is increased Okay, and only if you stimulate in the forebrain very strongly then you get a direct Movement, okay So the idea is the forebrain is actually not overruling the circuitry The forebrain is tuning the circuitry and that to me makes a lot of sense because that is you know the midbrain That's the old circuitry and the forebrain was the you know circuitry discovered afterwards Developed afterwards. Okay, so that's well it. Yes, it does apply to real animals as well So it's also there in the mammals. It's been neglected a lot It's the parabageminal nucleus and everything that has been studied so far is actually comparable And this is the last thing I want to show you so you do have a world You do have topographic sender input from the world It's projected onto the optic tecton where you have a map formation You have that system that actually does something like winner takes all novelty preference stimulus competition and stuff like that and This system is then tuned by the forebrain in order to get a well most Survival appropriate response out of the system. Thanks for your patience We can take one or two quick questions. Yeah Thank you for a very nice talk What I want to ask is that whether you found of course, I'm sure you did find but I didn't I Couldn't you know understand that aspect if you found connections, you know As you explained in the in the first few in the first few slides about how one tries to to identify the Significant aspects in the environment and I'm sure there is some very strong correlation with either memory or with you Know emotions so if there is a fearful stimuli versus a happier stimuli So could you find? Connections with the hypothalamus or with the you know amygdala back to the tecton and no we haven't looked at that and of course I well, I do not want to to Elute to say is that this is like, you know the major part of the story. I mean, it's just one automatically It's an automatic mechanism going on in a situation where the brain or the animal has no clear Idea of what it is doing of how to weigh information and so on and so on so this is just the you know Automatic part of what the brain is doing how emotions and so on are integrated here is Definitely beyond what what I can talk about because I'm sure you know experience will of course matter the animals Prior experience to you know particular kind of stimuli when I would see a wolf in in an you know image I should think of a wolf as being a dangerous Thing for me and that is why I would I you know, I would identify it right in the environment This is definitely not something that you would you know Locate in that mechanism so that mechanism that mechanism is kind of dumb Okay, so this certainly has to involve cortical circuit This certainly has to involve a mygdala circuit and so on and so on this is not something that can be done with that System this is like I said it's like the default situation. You have no clue what's going on But you have to respond quickly. Okay, and responding quickly is what this system is capable of doing Okay, thank you very much