[00:00:06] >> For there are today my pleasure to introduce our speaker today this. Year of California Berkeley both because she's a fantastic hero scientist and a leader in the field but also because of my post doctoral work with her 20 years ago started to pervade. 20 if there we are and no 20 whatever. [00:00:29] So I did in fact she helped me make my transition from engineering into your assignment so from that always grateful so young to her I got her bachelor's in physics from p.k.u. from Beijing University she went on to make a transition that didn't have to actually do this sorry her bachelor's in physics she went on to do her Ph d. in biology in the recliner in the university where she studied synoptic practice could be and signaling mechanisms she then they transition into. [00:01:01] Visual to the visual pathway and systems neuroscience in her post off the work at Rockefeller where she she did work there in the transition to Harvard University work in Cleveland and we were having the speaker of the year ago or so. Where she didn't have to work in the visual pathway She then went on to joining the faculty at Berkeley in molecular and cell biology in the late ninety's and that's where I met young and I remember at that time when she got hired she you know we saw many different seminars coming through isn't all the kind of warning signs talks that we've all sat through and she came and gave us just wonderful talking was really exciting and for me it was an exposure to a side near a science I didn't know about was very very exciting and she went on to start her laboratory there and in Berkeley. [00:01:49] And working on a wide array of different problems and it was really fun to watch her build a laboratory and be part of that stark culture at that time to see how different projects form and so on so she was a real knack for. Sniffing out really exciting problems if you looked at her background. [00:02:10] She's seen a lot of different areas from snap to plasticity to visual pathways both coding problems as well as mechanistic things onto brain state you know who interesting things that don't follow and then more recently in sleep she is or you have just disappeared from. The pole of a distinguished professor as well as our Howard Hughes Medical. [00:02:38] Institute investigator always a member of the National Academy of Science So it's really a pleasure to have your very exciting and I hope you all want and her alternative to thank you so much. Thank you. So I'm not wearing this microphone to you guys so the back here make this law will get someone say No Yeah Ok so I'll just outside a scream and then if you can hear me just wave your hand and then the other thing is that I know that you guys a lot of your engineers so if you know I'm not explaining things clearly you can feel free to entrap me I don't have to finish top because it doesn't matter at. [00:03:20] All. So so 1st thank you Gary for. 4 of the nice introduction was actually my very 1st post doc so without Garrett joining my lab my tenure would have been very questionable. So it's very important Ok so I like to tell you about our work. Sleep so this is something that my lab started about 9 years ago so we're still kind of new to this field. [00:03:49] Sleep is an essential innate behavior is innate because you know we know how to do this when we're born we don't need to learn how to do it it's essential because as far as we can tell all animals sleep including fly zone worms and even animals without a central nervous system including a jellyfish with only sort of spread out that. [00:04:12] So for none the male in animal models is usually detected by a lack of movement so they have to stay still for a period of time there's usually threshold duration that's how you score sleep at this all and also when you use a mild sensory stimulation you failed to evoke a motor response but everything's measured military and for mammalian eneral models we measure. [00:04:40] So when we fall asleep there's a very clear reduction of skeletal muscle to measure but you and then based on each year we can distinguish rapid eye movement or rem sleep and also Mangrum sleep which is also called the slow wave sleep because you see these are large amputees slow waves during. [00:04:58] Sleep. So for humans we measure e.g. empty and also a few other things like. Heart beat and also breathing so when we fall asleep there's a dramatic reduction of somatic motor activity measured that year and also there is a reduction of the motor all put like your heart rate goes down breathing rate goes down so the point is that falling asleep is not just about changes in the brain state as you can measure in the ichi but it also is about the reduction of the different types of motor output and so what I hope to commence you today that when you look at the neural circuits controlling sleep actually a lot of them are also closely associated with these motor circuits and that's why I have this sort of strange title for my talk. [00:05:52] Ok so a pioneer by the I'm a guy who discovered or Also system we actually know a lot about the newer circuits controlling weeks windows and arousal and so now we know that there are a lot of these well known your modular 2 systems such as the coli Nordic system and also mono I mean Nordic systems such as North trendily in Sarah Toni and histamine and opening. [00:06:17] Also some newer peptide systems the best known one is a Rex. System because if you lose that you have narcolepsy In contrast we know much less about the neural circuits controlling sleep so according to the textbook there is a single center in the end Sheria hypothalamus is called the pre optic Aereo appeal way where a subset of the neurons promotes sleep by inhibiting a lot of these well known weak promoting your systems so indeed using retrograde labeling from some of these weight promoting centers which we know a lot about we were able to identify a subset of the neurons in the peel away that actually promote sleep but it turns out this is far far from the whole story because about 4 or 5 years ago we and others have identified some sleep neurons outside of the peel away suggesting that the network is actually much more distributed than this model suggested by you know this textbook model with a single sleep center. [00:07:27] So they get an idea of what this distributed network might look like we decided to do a whole brain screening for sleep neurons. So we have 2 basic criteria for actually maybe actually use the cursor because some people are looking at that screen so we have 2 basic right for sleep neurons 1st they need to be sleep promoting. [00:07:52] The idea is that there activation should increase sleep and there in activation should decrease sleep Secondly they should be sleep active meaning that you know they need to be active at the right time to do their job and for screening for sleep neurons we also have to correspond these strategies we can either screen for sleep promoting neurons and then see if they also sleep active or we can screen for sleep active neurons and then see if they also sleep promoting So usually the sleep neurons when we call them the sleep neurons they need to sort of satisfy both criteria. [00:08:26] So I'm going to give you one or 2 examples each strategy that we have used. So to screen for sleep active neurons we use this technique called The Fall of false trap developed by Leach in the us lab as them for university so we crossed 2 miles lines together one of them expresses Creek are under the false promoter so Fos is a immediate early chinos activity dependent usually when a neuron has been active recently they turn on the false gene and in this mouse they should also turn. [00:09:01] And then the other mouse expresses. P. in the create dependent manner so we divide these 2 as this is crossed into 2 groups one of them we call them the aesthete the sleep deprived group so we do is to inject to mock Tiffen during sleep deprivation and the idea is that Awake active neurons they should be active during sleep deprivation because we keep the mouse awake during this period so they should express creamy r. and then be injection of 2 mocks of than should cause creamy Our which is a molecule to enter the nucleus and then this creek is a d.n.a. become been a stretch so when it enters a nucleus is going to do something to this other gene and so it's going to turn. [00:09:50] G.o.p. in a permanent manner so we can label the weak active neurons because we turn on the gene and then in this other group we call into our eyes group we're 1st going to sleep deprived them for 6 hours to make the most really sleepy and then we leave them alone we basically stop harassing them so they enter recovery sleep threat and then when we inject a mug's of than that's going to you know turn on the pee in the sleep active neurons that are active during the recovery sleep. [00:10:22] So this was done by a mostly Judge John with the help from punk. Has lap my lap to start home lab in Shanghai so while we compare these 2 groups in my eyes we saw that in some areas or so these are different grey areas so in some brain areas we saw you know more chippy labeling in the r s groups so those are shown here in these other bring areas we saw more g.o.p. labeling in the which is hardly surprising but when we looked at the top candidate this is the region with the highest relative expression of g.d.p. in the r s group so these are the putative sleeping around that region is shown here is basically the ventral lateral part of the Perry grape we call it the pot says you can see there more g.o.p. labeling in the our as group than the s.t. group now on the other hand even though there are a lot more neurons expressing Gippi But this is still a very small minority of all the cells expressing in the ventral pack so we have to figure out what you know who these neurons are and in order to do that we did Gene profiling so the report of miles line that we used to express is not only chippy but also this molecule called out 10 a which is a right also more protein that's normally attached to m r ne because his normal job is for our nature translation so even we use an antibody to pull down p. we can also point out a together with the m.r. name attached to it so we can do sequencing. [00:12:00] So that's what it data looks like this is called the volcano plot the x. axis is relative expression level of the gene Reza Each dot is a single gene so this is the relative expression level of that gene in the putative sleep neurons the g.o.p. expressing neurons relative to the entire trunk of tissue that we cut out now the verge. [00:12:22] Axis is the p. value so when we zoom into this red box here this is the these are to genes that are highly significantly in rich in the putative seed neurons you see a whole bunch of genes so we decided to focus on neuropeptides for 2 reasons one of them is that we know that newer peptides are important for signaling Secondly there are just a lot of cream mouse lies associated with neuropeptides So in this box we saw 3 neuropeptides and among them we thought that this gene called Helka which is cull see Tony gene related peptide all far right and name doesn't really matter but anyway so this gene expression shows a high overlap with the labeling of the putative house so we decided to focus on this Kalka marker. [00:13:15] So 1st we wanted to make sure that these Culkin neurons are in deep sleep active even though they are labeled by c. false befalls is a convenient but imperfect label for activity because it also depends on other factors so we wanted to make sure that these are really sleep active and for that the only way to do it is to to actually record activity so for that we did something called the courting optimal consists of an optic fiber you see this little optic fiber here surrounded by several But in those of you know 10 trolls or Stuart Rose or whatever. [00:13:50] So we packed the culture neurons with Channel Dobson by injecting the create dependent. Stress in expressing buyers into the clear line generated by Richard Palmer to his lab so here you can see that we've got this nice unit right you see these queens in the unit and every time we turn on the laser briefly 5 millisecond are the little blue dot here you see a spike in with a very short latency with a high reliability so we're pretty sure that this is a unit that expresses tunnel Dobson therefore you should be a cocker expressing your own in this Miles line. [00:14:28] So once we identify unit like this we turn out a laser and just do recording of the spontaneous activity so here this is the power spectrogram So if you see a lot of broadband low frequency activity this is probably a language that speaks to you that engineers if you see a low a broadband activity a low frequency chances are the animals in slow wave sleep a lot of high I'm to toot slow wave but if you see sort of fade up and at about 8 hertz the enemy either awake or asleep and you distinguish those based on you which is the 2nd trees. [00:15:09] Groups last night trees or you have a 2nd trees here and then this is a fun read of an identifier. So you can see that this neuron fires at a high rate whenever the animal is in the run sleep this orange period and it's the fire it is low during either ram or the blue period or week or the gray period. [00:15:30] So here is a summary of more than 30 identify Calkin neurons each line is one neuron across with 3 brain states and we also show our data in this 2 dimensional plot so the horizontal axis is the far a difference between them sleep awake normalized by their sum and then the vertical axis is the fiery difference between sleep I'll wait so the identify Tulka neurons the blue dots are mostly on the right hand side right indicating that these neurons fire more during. [00:16:04] The week but they don't seem to have a consistent Fahri difference between rem sleep wake. Now if you look at the grey units so these are recorded with the same old right in the same area but these are not 2 of them by the laser so we don't know who these neurons are but at least they're in the same area but you see that many of them are on the left so these are weight active neurons so based on this we know that the majority of the Culkin neurons are active so this is our 1st criterion so it's basically confirms the. [00:16:40] Experiment. So now the question is are they promoting And so for that we did. The activation so here is sort of a typical experiment we turn our laser on for 2 minutes per trial right the blue shading period. I randomly every 15 minutes or so and so you know the same time we're recording that. [00:17:05] And the color bars just a computer automatic about a score of the steep stage so this is a summary of all the trials from all the mice that we tested basically the laser triggered average So as you see when we turn a laser 0 here. For 2 minutes the blue shading here we see a very immediate and clear increase in the probability of 9 goes from the baseline about 60 percent to about 80 percent at the same time there's a reduction of both wakefulness and run probability. [00:17:44] And when we inactivated these neurons using i.c. plus plus which is basically a mutated channel Dobson that's a chorus channel we saw a decrease in the runs the been increasing wakefulness. So in addition to this kind of acute absolutely netted activation or in activation which is just 2 minutes per trial we also did a dread experiment which is slower longer term and you can see that the g g Q mediated activation cause an increase in the sleep for about 2 hours and the g.i. mediated inactivation cause a decrease in that round sleep. [00:18:21] So based on both. Activation and you know activation we know that these call can neurons on promoting which is our 2nd criterion. So so here we're just using this neuropeptide as a marker but it turns out that in these subcortical regions that most of the neurons they do different things right there they use these newer peptides but they also have you know these traditional neurotransmitters So it turns out of the Culkin neurons about 60 percent of them are also good made her tick and the other 40 percent are coli energy so without show you the data tell you that we think that the sleep promoting the fact is mostly mediated by these. [00:19:07] Population the Conan ordered neurons they are actually the pretend. Parasympathetic neurons and their main goal and in their main job is to constrict the pupil when you activate a code in order to pass so they're kind of related to sleep because when we fall asleep the pupil size also constrict but their main job is not to promote sleep per se. [00:19:31] Ok so these are herding neurons that promotes sleep and it turns out that in this region they're not the only made her. We also found another population and these neurons actually express another peptide are called c c k So the c.c.t. neurons are almost 100 percent. And they're entirely separate the Culkin neurons so in this case the red neurons are Culkin neurons so you can see that they're sort of sitting on the 2 sides and the c.c. neurons are sitting exactly along the midline right so they're entirely separate. [00:20:07] And so when we activated the seeking neurons we also saw an increase in the Sleep and their inactivation causes a decrease in their obsolete so the seeking neurons are also promoting. So we have these 2 populations so they're sitting right next to each other and then the question is you know do they talk to each other and so we did some slice experiments and it turns out that Ok so so the way that we do this is that in this case we used. [00:20:36] C c k cream else so we can express channel Dobson in the c.-k. neurons and then we Potch So this is slice recording and when we pop a couple neuron we know that because we suck all the psych cytoplasm from the neurons that we patch and then we do our key p.c.r. to confirm that there are you know. [00:20:55] So in this case you can see that activating the c.d.k. neurons cause a big excitation and that's blocked by seeing 2 x. which is a good way to receptor blocker so we know that there is glutamate her Jake not exactly Conversely when we activated talk of neurons and records on the scene. [00:21:13] There's also excited to put so that just says that these 2 populations of neurons they both promote sleep they actually talk to each other which is kind of interesting because this seems to be a way that they can reinforce each other so that they can stay active during the interim sleep. [00:21:31] Ok so just a very quick summary so what we found so far is that we have found in the park there are these kuchen neurons and they talk to each other. And so they promote sleep and again without showing the data I'll just tell you that they actually project to the appeal way which is the textbook sleep center so the pos state neurons are urging these neurons promotes sleep by inhibiting these awake neurons and then the real pack neurons these are excited tree and they promote sleep and they do this seem to be so this effect seems to be mediated by their projection to the. [00:22:13] Right and so there is they seem to be like upstream of the. Promoting us and then also with on this other region in the in the very very poor security are these are also Gobert in your eyes that promotes sleep and neurons also project to the g. i.v. to promote sleep. [00:22:34] Ok so now I'm going to switch to this other approach that we have taken which is to screen for sleep promoting neurons so we did this based on a very simple minded logic so if you don't you know that the technical details doesn't matter but hopefully you understand the logic which is very simple so we're going to start with the weight promoting neurons we actually know quite a bit of them and what we're going to do is they use a retrograde tracing to find their inhibitory input because if you are neurons that inhibit a lot of the way promoting neurons then perhaps your activity can promote sleep because sleep wake these are like incompatible brain states and well once we find these inhibitory sleep promoting neurons we're going to choice one more step back a look for they're excited to employ it because if your neurons that excite other sleep promoting neurons then your activation could promote sleep as a very simple kind of. [00:23:31] You know sign either you keep the same side or you you decide that So that's our logic for screening for 3 promoting around. So this was done by a very talented post-op Cheney a minor law so Cheney and targeted a whole bunch of weight promoting Like I said in the beginning we know a lot of them already and so she was looking for brain regions with a relatively broad. [00:23:56] Innovation of these wake regions so this is her 1st batch of screening these are brain regions that she targeted the weight promoting populations including history. In Michigan cholinergic read basically a whole bunch of them in 7 brain regions and then these are the brain regions where she found a relatively broad projection to the Awake promoting populations so this includes the pill way which is the textbook sleep center but in addition actually you know in this limited screening the top candidate that she and found anatomically is the ca which is essential nucleus of which is normally people think of it as sort of emotional regulation. [00:24:51] So basically this is an example of each of the 7 populations that target it and in each case her rabies wretchedly chasing shows the labeling in this tiny area this little white box here and if you blow it up you see a lot of the green which is the rabies virus labeling and red is basically the marker for God. [00:25:17] So in each of these cases you see a very strong overlap between the readings labeling and the neurons. So again but that the label neurons is still a small subset of the governing us so we have to figure out. These neurons are and again we did a gene profiling where using the same technique to explain earlier in this case if we look at the highly enriched in this box you are again we saw a 3 neuropeptides and it turns out that when we compared expression of these 3 have ties 2 of them overlapped by more than 90 percent and 2 of them overlap by 50 percent or so and it turns out that the peptide N.T.'s which is north Tensen seems to be the best marker because it labels a lot of the rabies label neurons and also N.T.'s is pretty specifically expressing the ca analogy in the surrounding brain regions so we decided to focus on and Ts and again we use the and he has created. [00:26:17] So I'll just you know go through his very quickly activation tunnel Dobson cause an increase in their runs the decreasing wake and Rem inactivation cause the opposite effect. And then yes also the CIA and there are 9 run promoting We also did recording and so here is a summary and you see most of the identify new around both of them are in the upper right quadrant so these neurons are actually more active during both ram and then runs deep compared to wakefulness and identify great dolls are all over the place. [00:26:56] So he's most of that. And he has in us also sleep active. Ok so here we have a population of inhibitory sleeping in the ca So as promised we're going to trace one more step back and look for their excited to put so 10 year mostly found 2 regions with very strong growth we've heard of projection and so one of them is this region here as you see the rabies labeling and this is the clue to which is a marker you see the overlap so what is this region so this is the Allen. [00:27:32] And it turns out that this green stripes here with a lot of labeled cells is basically this sort of vertical state stripe right between the border. Which is in red region here and the rest of the brain and I think that this is the mid bring. So it's not the entire thing is like this sort of vertical strip and we don't even know what to call that region you look at Alice they give these weird names and in the literature there's very little consistent reporting of what that part of the Fatima stuff. [00:28:08] So again we have to find a marker now in this case we just made a lucky guess right because in the CIA and here seems to be such a great marker and we said Well could it could and he has be the marker here and so when we looked at the Allen. [00:28:22] Is it. Inside to have a decision database it turns out that N.T.'s is exactly express in his vertical strip and when we looked at and he has expression with the rabies labeling and again we saw a very high percentage of overlap so it seems like and she has is also a good marker for the good news heard in your arms and indeed when we activate these. [00:28:46] Glutamate heart and he has neurons we saw increases sleep and you know activation causes decreased. Ok so now we have 2 populations one going to meet her tick and the other and yet a common denominator is nor Tonnesen as a marker and so you know I gave this talk actually the people just came out last year but I didn't given this talk for more than 2 years just to show you how long it takes for the publication process so so initially I would just present this and the question that I always get is that was the role of and Ts right and so for a long time answer was I don't really care all we need is a mark her right as I said get the right career I don't care to rest but if you ask this question like 20 times you always give this crappy answer to some point you're like Ok maybe we should do the experiment so finally we did it and I said the way that we did it was to use crisper line technique. [00:29:42] You know once people figured out this it turns out the implementation is pretty easy. So what we did is a what 10 year and it was 2 days. It is a very complicated construct the Basically it expresses the present marker and tunnel Datsun but also the Gaidar nay for whatever molecule you want to knock down. [00:30:06] So. The control is basically a. Sort of run them gene and then the other one that's interesting is N.T.'s or so of the so if you use it guide our against and to yes then cast and eyes basically going to damage the gene so that you can knock down the expression of N.T.'s So here you can see that in the control you see the cherry expression means the virus is working and he has expressed but when you express and he has Darnay you can read knockdowns and he has expressions here and so you can do this in the Ca or in the post here asylum now when we did this in the control case which is this white bar. [00:30:53] An activation still causes an increase in sleep I said this is just a repetition of the experiment I showed you earlier except except that we also expresses a control get our name. But why you not done she yes this black bar here you see that the increase is greatly reduced Not only that when we express the Gaidar that. [00:31:15] Transmission We also saw a decrease in promoting fact now in the poster a solemn oath you can knock down and yes that basically wipes out the the run increase and if you knock down the meter transmission you basically also eliminate the non ram promoting fact so the conclusion of this is that both he and his peptide and the traditional transmitter is either good we tortured or gaba are important for the night in fact of these neurons. [00:31:48] Ok so here we have and yes that already labels 2 populations. And it turns out that and he says is a good marker even broader than that because in this case we just went to the. Data Base and we found another cluster of and yes neurons in this case is again in the peri actually Dr Gray the pack rat and so when we activated these neurons that this was done by pong. [00:32:17] Ok so when we did that we also saw a huge increase in the inactivation causes decrease in so these v.l. pag. Neurons So these are also good made her tick also promoting Now we also did courting these padding around but just for a change I want to show you a different kind of technique in this case we did calcium imaging because the pack in fact most of the regions that we look at a very deep in the brain so we can't really just do the regular imaging we had to insert is green lens which is a micro endoscope so that we can bring the image from deep down here to up here and we use is miniscule developed by Mark so that the mouse can actually wear it on the head and move around and so here is you know you can see some cells and these are the Chelsea and traces the bottom line is that these Pac and he has neurons also active and so Paul also used the. [00:33:18] Develop I-Ching yen to knock down and Ts and that also really reduce the number of promoting effect. Tool for transmission also largely eliminated the fact so again in this case both the traditional transmitter and a newer peptide are important. Ok so just a quick summary here so we have in our screening for sleep we found these and she has neurons that leave all white range of populations the neurons in some cases of other cases due to her sick the all promote sleep and in all these cases both the peptide and the traditional transmitters are important. [00:33:59] So I also want to just very quickly highlight 2 things so one of them is that and here seems to be a really general marker I showed you see a solemn pack in fact there's one more population which I haven't show you the data which I'll show you later but I'm going to prematurely summarize that in fact these are also good made her tick and he has neurons that also promote that gram sleep so she has is also very interesting neuropeptide it was actually discovered in the 1070s initially as a high pole pensive peptide meaning that if you inject a peptide into the body it lowers the blood pressure. [00:34:39] And now we know that it also reduces body temperature and slows down the heart and in fact all of these things happened during the interim sleep so it's actually not just sort of a coincidence just the marker it actually is really doing something right related to alter nomic regulation and sleep the other thing is that it really seems that the venture region of the pack is a. [00:35:01] Upstream sleep center because with multipole population is of excited to be neurons that promote sleep and they all project to these. Promoting centers. Ok so I'm going to try to quickly tell you one more story again using this 2nd strategy so I told you about the peel away and the Cea and it turns out that this 3rd candidate region that came out of this anatomical screening also is interesting so this region was. [00:35:38] Was. Characterized by dental you another really really talented post doc in a lab so this region is called as an artist of. Parser which is part of the basal ganglia. Now many if you know many if you probably know that the basal ganglia is really important controlling movement right so to understand how these sort of the motor control neurons also seem to be related to sleep we decided to take a closer look at the motor behavior of the mice so in addition to each. [00:36:14] Recordings in the in the home cage and our collaborators in Beijing helped us to use deep learning to train is something called the unit for image segmentation seems like you know deep learning is all the rage the whole planet is doing the learning so we didn't want to be left out and that's why we did it. [00:36:34] So anyway so you can see that this is doing a pretty good job so this is the original image and this is word network things the mouse is right so basically once we have this we can you know have a relatively streamlined analysis of behavior. So what we did was to extract 2 prime ministers one of them is a translation would basically look at the center of mass of the mouse in 2 different time frames and see how much it moved the other is total movement which is the total number of pixels that are different between these 2 time frames and when we plotted these 2 prime ministers we basically saw 3 clusters the dark green cluster correspond to locomotion the mouse is running around it's really easy to tell the 2nd class or corresponds to all the non-local motor movement dry so they're not you know translating but they're kind of like grooming or even worse you know posture adjustment the 3rd one is basically mobility but once we look at the g. of this cluster they're actually true behavior States so one of them is quite a weakness and the other is real sleep. [00:37:45] So here I was showing you the each doubter power and the e.m.t. of these 4 behavior States so each eat out of power which is basically the low frequency stuff so the higher the power the less arouse the brain is now as you can see that. You know one thing I want to point out is that. [00:38:07] These 3 greenish stays red they're all awake states right so we tend to think of their difference in terms of their different movements and you can see a clear difference in the energy but there's also a clear difference in their Delta power there is also a corresponding difference in the brain state now quite awake when those and sleep they're both immobility so we tend to think the difference mostly in terms of brain states but in fact there's also a very clear difference in their power now the point is that each brain states have motor activity they are correlated across multiple behaviors States and that's we think it's quite important. [00:38:48] Now when we looked at de transition between these 4 stays they're not random at all so when we look at all the trials in which they start out in local motion they always transition into non-local motor movement we have never seen the case were demolished was like a running in one moment and then they fall asleep the next moment it just doesn't happen and the Allies you're narcoleptic you've seen these You Tube videos like the dog was a running fan they fall asleep that only happens during narcolepsy it doesn't happen in normal animals like we don't do that either. [00:39:23] So if they start out in. The local motor movement either they go back to running or they go into a quiet week on a say don't fall asleep right away in other words if we rank these different states you know we put them in a single chain the most of the transitions are between neighboring states so this is the transition matrix that Bagnall means staying the same state and most of the transitions are either you need it Lee above the no meaning going down the chain one step at a time or going above the change one step at a time the only exception we see is that sometimes they wake up immediately going to. [00:40:02] The local motor movement and skip quite weak and that's. Ok so this is just behavior and now let's get back to s.n.r. So it turns out that are actually true subpopulations of Dobber doing around us now while we injected the create dependent virus into the pvp Parvati increase lies we mostly label the latter as an r.p.g. positive neurons but when we injected a virus into the 2 cream eyes we mostly labelled the media or as in our neurons that r.p.v. negative so we called this was a gap to neurons. [00:40:37] So well we use Opera recording to require phone the p.v. cells so this is just an example most of them are actually more active doing these movement States and then during sleep they really have the lowest Feyerick So these neurons are actually not sleep active but when we record it from the to in around us they tend to be most active during sleep and less active doing these motors States and so this is just I'm just superimposing the population average of these 2 populations as of the media was our sleep active the latter ones are not now when we often genetically activated the 2 neurons we saw an immediate decrease in locomotion on the local motor movement and clear increase in sleep but when we activated the pv neurons we saw no effect on sleep the main effect we saw was to decrease in movement and increasing quiet wakefulness and while we inactivated the got 2 neurons we saw a huge increase in movement and decrease in sleep and p.v. neuron. [00:41:45] Inactivation has a much weaker effect. So we think what's going on is that the pv neurons their movement activated but their activation many maybe terminate movement so we think what's going on is that this is mostly based on the basal ganglia literature we think that the pv neurons are in porn for action selection because every political or movement we make we also need to suppress these other and these are movements so the pv neurons and that's why when you make each movement the pv neurons are activated to suppress these other ones but if you activate entire Peavy population you basically suppress all movement but that got to a neuron seem to be they just generally suppress all movement and they're active during the Lol movement and sleep state. [00:42:37] So even though got to a neuron activation has a very strong effect decreasing movement and increased sleep they don't do this by causing direct movement to sleep transition you have decreased while you increase the other you transition but that's not what was going on so in fact when we look at the change in the transitions are we never see any case where these white ones are the transitions that we don't normally see that because they're skipping with age we also don't see that with the activation Instead we see the increased transitions in these right above the transitions meaning going down the chain was that the time and. [00:43:18] The reduction of these other transitions so we think that this is important because. You know when we look at this 1st transition going from locomotion to non-local motor movement we say that's motor control. And then you look at a 2nd transition with Ses motor control but the 3rd transition from quite a wakefulness is sleep we say brains they control so we give them different names but these neurons they really don't care what we call these all they do is that you activate these neurons they basically push the animal down the chain one step at a time so I think that this is sort of an explanation of how these seemingly motor neurons can actually also promote sleep. [00:43:59] Now when we inactivated these got 2 neurons we also don't cause any artificial transition we basically bias the natural transitions in the opposite direction Ok so we also looked at the downstream targets of these neurons the pv neurons project to all these usual suspects motor control regions that got 2 neurons project to these motor regions but in addition they also project to the dorsal raff a where you see that in this region the got 2 neurons project there but the pv neurons do not so the dorsal ruff a contains the nurturing neurons and also the controlling brain state and also the locus surely is within the wagon urging us for brains to control. [00:44:43] And when we label the subset of the. 2 neurons that project to the motor part of the Fallon was we also saw the axon collaterals in the dorsal ruff a and the locus Julius and while we activated the subset the neurons projecting to the We also saw there are collaterals in these motor regions so it really seems like it's the same population that got 2 neurons that send x. on collaterals to the motor and brain state control regions. [00:45:12] Ok so we have got Birgit population now we're going to trace back and look for the excited tree input so for the isin our neurons the major excitedly input came from the. Nucleus which I prematurely summarized right so these go to a total neurons also express your tensor and when we activated these newer tensor neurons we also saw increases sleep and decreasing wakefulness Ok so just the sub at the beginning I said that falling asleep is associated with changes in the brain state and also a reduction of somatic and autonomic motor output so it kind of makes sense for the sleep control mechanism to inhibit the regions promoting arousal and these motor activity so I told you that for the so my **** motor movement we have found the s t n s n r neurons that promote sleep these are ganglia important part of the somatic motor control system now I also told you in the. [00:46:14] Hypothalamus. And I didn't show the data but also in the nucleus of the solitary tract we have found these neurons promoting sleep so all of these are important knows of the central autonomic network. So instead of being separate as implied by this circuit diagram it seems like the freak control mechanism deeply if you trace these 2 types of motor control systems so I move these yellow sleek neurons into these motor boxes. [00:46:45] On the other hand there seem to be a separation between a general state neurons and a specific action or what I mean is that the pv neurons in the as an arguer specific action neurons but the 2 neurons as a general statement or else now the general state of seem to send collaterals to the arousal system and the motor control system so that they can cordon 8 movement and brain state. [00:47:13] So we think that this kind of view of how sleep is controlled also has implications for this other question which is why do we sleep now we tend to think that sleep is mostly about changing the brain state because we lose consciousness of the environment so we're humans were soul you know like is mostly about conscious experience so the question is why is this think control mechanism sold tightly linked to and why do we need to shut down movement for a period of time each day and to me this suggests that the fundamental function one fundamental function of sleep is probably for subtler repair. [00:47:55] Why do you need to rest right probably for some kind of restoration and repair but what I don't mean is only for muscle repair I mean we know that for Astley's do in sleep they really we've heard the muscles and it's really important for their performance but we think that it's broader than that right for people in this room were probably many of us are not real Astley's we're kind of like mental athletes. [00:48:20] But I think that probably I tend to think that the a rousing system is kind of like the central muscles and they need to be repaired as well so that's something that we're really looking into right now so these are the people in the back in my lab actually did the work so the 1st part is just work and she has left my lab running her own lab in Shanghai Cheney and invention and Palm. [00:48:45] Did the 2nd part the anatomical screening and these 3 people I didn't even show their data but really there are the early pioneers in my lab they joined me about 8 or 9 years ago before I knew how to study sleep there are really the ones who figured out how to do it well so many collaborators who helped us a lot to thank you. [00:49:24] Yet it's a very good question so I didn't talk about sleep so Mark so much right I mean you know most of the neurons it talked about today are mostly promoting neurons sleep. So the ram control circuit the the 2 circus are probably all in the in the brain stem more posts here in the in the palms and the much. [00:49:44] So I didn't talk about the data but. About 5 years ago we published people are showing that there is a population around in the central region of the model of. The very very powerful e. promote run sleep pretty much if you opt to genetically active as you turn on the laser if the enemy is already in the run sleep in 95 percent of the cases you can push them into rem sleep within like 30 seconds of activation. [00:50:14] So so yes those you know and this is have to based on earlier lesion studies if you sort of make a cut in different parts people have shown that the run control the key circuit is mostly more posterior so that's consistent with what we found as well and he seems that the NIH run sorry the run circuit is probably simpler it's not as stupid as the Nairobi control circuit. [00:50:43] We'll. Good for you. So I am sort of using that term deliberately in the in the big size because I don't know. So you know if you ask me this question was a function of sleep in fact you know everything I talked about is ha we sleep not while we sleep in my mind these are related to big questions right. [00:51:16] So you know if you asked me like a year ago say why do we sleep I get asked this a lot now my answer is that I don't care because I do is to figure out how we sleep but sort of the story of my life very often I would say I don't care about something and then band one day that's the only thing I care about right I have 10 have these big transitions so I only became I only started caring about this big time in the past 6 months so it's way too early to tell I have a feel sort of general ideas I think that protein synthesis is going to be a big part of it so one of the things that we already know from human studies is that growth hormones you know they're actually you know girls formal is mostly released during sleep so we spend about a 3rd of the time sleeping if you sleep 8 hours per day you're lucky person so it's a sort of the time but in fact probably 70 percent of girls hormone release happens during sleep. [00:52:18] And so goes hormone we know precisely stimulates muscle growth and bone grows it all has to do with the sort of motor activity but then I think there's also a central effect as well because you know for people who have growth hormone deficiency you know obviously if you have that during development you know development is is affected big time but there are these cases where people develop as adults if you have that go to more. [00:52:46] Something they not only have you know weakness of muscle and stuff like that they also have you know depression and sort of cognitive impairment so I think that there is that there's odd angle in the central nervous system this wall but there's very little in the literature so that's one thing that we're looking at now and there are also these you know general things like you know metabolic stress. [00:53:12] So I think that that's also a big part of repair. Yeah yeah. Yeah. Yes Yes No I mean it's a very good question in fact you know how you important is run sleepy very controversial I don't know because I only started thinking about function recently. So you know when I asked people some people would say the rem sleep is probably not important and they say that one reason is that for people with depression if you take antidepressants the big class of that drug is you know before ketamine. [00:54:06] Is mon ami in Riyadh taking So you basically increase the mold on means that So it turns out the more not means system is completely shut down to sleep so there wake up here and I run down here run this completely down and if you want to have sleep you really have to shut down the bone I'm a Nordic system so if you use the mon ami nervous system you really reduce the time so so people's argument is that you know these people when they start taking antidepressants you know they seem normal they seem to be doing better than before they take their drug and yet run sleep is greatly reduced that must mean that rum sleep is not important although when I talk to some clinicians they say initially gradually it comes back so then there's a question of you know long term. [00:54:56] Deprivation this out have some effect clearly seems like short term probably not as bad as sleep on the other hand I've talked to a psychiatrist and when I asked him I said you know is it true that run sleep is not important he said Who told you that if you lose runs the vehicles psychotic so then I think that the confusion is part of that is that you know sometimes people reach a conclusion but with different kinds of quality of data you know when people say the patient completely lose rem sleep how rigorous is that measurement and unfortunately you know I'm not from the human side so I'm not qualified to evaluate that. [00:55:39] But I do know that there is a recent study from a friend of my near Israel so she actually tracked down this patient a famous patient almost like him in the sea field so she had a injury like there was a shop now or something during a war that locked in his brain so you know if you measure each he she has no runs. [00:56:01] Very little. And that happened like 20 years ago he's a lawyer he has a successful practice so you can use that to say maybe he doesn't have you know maybe ram is not important on the other hand you can say well how do you know he doesn't have to read me all you know is e.g. it doesn't look like I'm so unfortunately you know you get into these are very complicated nuanced argument. [00:56:37] Yeah yeah yeah yeah. Yeah. Yeah. Yeah so it's a very good question right so which is why I actually sleep I'm so Kadian you know it's very closely related I when I 1st tell people is that you see this you study circadian rhythm there's separate so Katyn rhythm it regulates everything not just sleep but also feeding body temperature everything regs that sleep is something that's the most obvious threat sleep is separate so when I say the growth hormone is dependent on sleep there are actually people are showing that I mean the way you separate those 2 is you know you look at normal sleep and then you deprive sleep and so what people have found is that you know by the time you're supposed to go to that but if they sleep deprived during that day actually your goals hormone release is reduced until you allow the subject to go to sleep so they are ways to separate them but you can't just sort of say normally happens at night therefore you must be sleep not necessarily.