 First of all, thank you very much for considering the invitation and also for other organizers, you know, if all for this workshop. Now I will share my screen. Okay, so can you see this ppt well. It's good. Yeah. Yeah. Okay. Great. Thank you very much. Um, it's my great pleasure to introduce one of my research project to you. And the title is he may bring any energetic. Neuropea distribution. I from Fudan University. My laboratory is based on is focused on the computational study of the of the brain. Actually, we want to, you know, study the computational principles and mechanism underlying the neuro representation of the art world, art world. You know, during the evolution, I'll bring, you know, from the lower animal to the high other animal, I'll bring get bigger and bigger. And from the left, left, left, a figure, you can see that is, you know, our brain is almost the biggest compare our to our, our body weight is almost the largest. And in the right figure, figure, you can see that really it's like, like the two, two big group, one is the cold blood animal. And another one is the warm blood animal, the warm blood animals have the big dream almost 10 to 100 times larger than the cold blood animals. So, because we have the large brain, so that is enable us to have the most intelligence motion to, you know, to represent the art world and also come down, you know, better cognition, they were low animals. So that when you look at these pictures, you can see that, you know, animals are fighting for the food. During the evolution, the animals become better and better in capture them, their food. And during this process, maybe the, you know, both our brain and also our body become, become, you know, optimized to fighting for the food. And regarding our brain, we want to, you know, to ask this question. How brain structure and functions are optimized for the energy, you know, energy utilization. You know, several years ago we have a game, you know, between human champion and, and, and AlphaGo computer, you know, system. So, although the, the, the first time, for the first time the, you know, intelligence when, when the human champion group here, however, for regarding the energy consumption for the human is only cost about, you know, one coffee during the several hours competition duration. While for the AlphaGo system is cost almost half billion watts to support the systems. So, also, compared to the most powerful simple computer in the world, human brain only cost about 20 watts. How the simple computer cost, most of them cost over one, one million watts for the almost the same, you know, operation, operation, you know, times. From this way, we could say that our human brain are highly energy efficient than the, than the, you know, artificial intelligence. The new, the, in our, our brain, the cost of, you know, our brain costs the 20 watts, you know, the energy support is, is, you are wondering, you know, where this, you know, where are the sources costing a lot of energy for our brain. So if we record one neural, sorry, you will record one neural from our brain, you can see that all the time during sleep and during the waking time, the neurons firing spikes all the time. Those, those spiking activities we are traveling around the neural axons and then right to the terminal of the, of the, of the neural and then induce the post-synaptic potentials. And for one neural, for one individual neural, it's almost connected to about, you know, 10,000 of neurons surrounding with this neural. So, so this process costs a lot of energy. So that makes the, makes the brain very costful in the energy consumption. Individual, one individual neural will cost the most 10 times of the energy of our body cell. And over 70% of this energy are supporting for the synaptic transmission and actual potential generation. And almost 30% of energy are cost for the housekeeping mechanism and also the keeping the resting memory potential mechanics. So this is basically the energy, you know, distribution for one single neural. So except our, except the neurons, our brain also have almost the same amount of the glare cells. For example, or for the human brain, we have the 86 billions of the neurons while have 85 billion of the glare cells. Those glare cells are also cost energy. But most of them are how important functions support our neurons to function well. So the reason we want to ask, we want to examine the energy mechanism of the, of the human brain is we also always want to ask. And the question is, you know, for the human intelligence, whether we have a limitation on that. So now we found that since the energy is a very, very, it's a key limitation on our brain. Actually, it's constrain our, the fine rate of the individual, individual neurons in our brain. And also this fine rate is also constrain our brain intelligence because it's limited the information transmission within and the brain disease. So from the left figure, you can see that is for the human brain, the average, the neuron activity actually from our previous estimation is around one hertz. That means 86 billions of neurons within our brain, each neuron in one second is just firing one individual. Even that amount is already cost a lot of energy is 20% of our whole body energy. So if we go beyond of this fine rate, that's maybe our body couldn't support for that. But also because of this, we could say that that's a one-hertz fine rate could limit it our brain intelligence to, to have, you know, to have more creative power or other information processing, you know, speed. From this sense, we could say that maybe in the future, the artificial intelligence could, you know, go beyond us because they could have more unlimited power to support for them. But for our brain, human brain, we limited by our, by our human body's metabolic rate. So for the, for our brain, the energy supply is basically is merely, merely supported by the glucose oxidation. So when a single molecule of glucose is oxidized, it's released about 36 ATP. For a lot of times, you know, because of the oxygen, it's sometimes not enough. Sometimes it's because we need energy very fast. So sometimes it's we have the glucose of that issue. Generally, it's generated below 36 ATP is around 32 to 34 ATP for one molecule of glucose. So we have a lot of questions we want to ask the brain regarding the energy consumption is how much energy does the green mentor cost and the white mentor cost, and also how much energy costs that, you know, for different brain readings and also for, you know, we want us want to also to know is during the cognitive perception, how much energy do the neurons cost and the cells cost, and how much energy do the brain cost in the rescue state and how much is in the behavioral test. Actually, we studied this, all these questions in the last several years, you know, from experiment, also from the computational modeling combined together, we know some of the answers. So in order to understand how human brain has been its energy efficiently from a subcellular level to a critical level, we did. We did a lot of competition modeling, but this is the final goal is we want to do is to construct a three dimension energy connect to map for human brains and also for the other animals. Currently, we for this project, we are already how finish the human brain energy connection connect to map for other animals we are still on the way. So basically, we are present the topic from the two, you know, two sections. So first is to build up the energy equations for electronic and chemical activities in individual mirrors and clear cells. The second is to construct the human brain energy map. Okay, the second is, you know, that's not I talked about that is for the brain is cost a lot of energy for the actual conditions and snap to snap to transmission. So in order to quantify clearly the energy consumption for our individual neural and the individual actual generation and propagation. So we generally we build up how to can actually style table energy function or capture the actual potential generation and and the propagation. So basically is the way we, we, you know, model the neural with Android and extra actions to to, you know, multiple compartment model, and then for each compartment we use this equations and the equation to capture them. And for each compartment we have the, you know, sodium and potassium and calcium. And the leaky conductance and we also have the capacitance within that security and then we solve the equation so we can get the keyboard energy derived from this. I don't want to go into detail. We lost you for 10 seconds. Could you please rewind back just a few sentences. Thank you. Okay, okay, sorry. So you mean the last second for this PBT right just a couple of sentences. Thank you. Okay, so because of the because of, you know, you know, lots of several say lots of several we, we talked about that is for a individual neural the most of the entities cost of all the supported actual potential generation and propagation and also for the snap to transmission. So, in, in order to quantify this energy cost for those for those actual potential generation and the propagation clearly a pure, more a pure reader way we build up hard can actually keep on any function for that. So, basically we, you know, describe a new individual neural is excellence and then right because excellence and then right. Communicate it with for this neural to communicate it with other neurons. So we capture this, you know, for structure of the neural into a multiple compartment model for each compartment we have the, you know, sodium and potassium and the calcium conductance and those conductance actually is a team changes with the voltage deposition. So basically this is a for each compartment that we have the sodium potassium channel and also we have the leaky and the calcium channel and accounting for the actual generation. We also have the capacitance within each compartment compartment. So, for each compartment that we could use this equations to capture the actual generation process and also for each, you know, boundary support for the for the ionic channels and also the So the conductance for that channels, we could use this, you know, a power power equations to describe them. So, in the, in the later we use this last, you know, equations to this to calculate the energy supported for the actual generation and propagation. Actually, we also could use this equation to to explain it for the synaptic transmission that is in a later studies. So we could, you know, The last of the ppt is construct the equation for the for the people, you know, but for this ppt we construct the full model of a neural, not only with the table equations of the excellence and then right by the we also have the cell body. So, this is a typical neural. So you can say that's the West actually generated generated in the axon in the axon compartment close to the cell body. It's about 30 to 50 microns away from the cell body in that location, the actual potential generated and then it's probably to the to the Exons and also it's back probability to the thing right. All those processes cost energy a lot. Okay, and also based on the equations and the equations we could do for the analysis and for example for for the axon has a long distance and also have a short, you know, short distance. We could calculate the energy for the is seems that if a neural have the long longer excellent. Sorry, I will remove the time because it's keep jumping. Sorry, let's see. Sorry about that. Okay, so that means for a longer exam, even neural have longer exam or longer thing, right. It's actually the power will be increased. Actually, sometimes increase to the person and then and also for the for the exons, if they have that if you have branches and also have the different diameters of the branches, the energy cost, we also could capture by this equation. Also, we could use this equation for capture the energy consumption for the neural nerves in the white matter. It's different because in the white matter, most of the exons are powered by the man on the nation. And then we calculated for for those are my little exons and also my little exons, you know, for the white matter and and I don't want to go to detail, but I want to just say something is for the young animals, sometimes most of the exons in the white matter are on my little. So we need to continue counting for that. But for the adulted animals, most of the exons in the white matter, my little for my little exon, the energy cost about 10% of the 10% of the exons on my little exons, even less than 10% but based on the specific situation. You know, whether, you know, the diameter and also the length, it's deep enough that the detail, but basically the, the melanated exons are saving energy for the actual potential propagation. Except for the actual generation and publication, the synaptic transmission is cost a lot of energy. So basically it's a pre synaptic has been released and the post the synaptic for the, the receptor. After the transmitter and then, you know, open the sodium and potassium channels, let's the ions to to go through the membrane and generate the generate the post the synaptic potentials those process. Actually, it's a small for a single synapse, but it's considered that one neural has almost 10,000 synapse. So those two and together that will cost a lot. It's much more than, you know, the actual potential generation and publication. If, if we, if we, you know, calculate all the energy for each component for the individual neural, for example, Exactly neural. So the synaptic transmission occupy almost 50% of the whole process. If we, if we add together of the case, you know, related to the transmission and the blue, blue to me, release and reception. So those synaptic transmission process would reach the 260% of the, the whole neural process. So actual potential only, you know, okay, you know, cost about 10% 10% of the four neurons energy consumption, while the housekeeping mechanism for keeping the cell survive and resting potential to maintenance of the negative, negative memory resting potential around minus 70s. Those together is cost about 30% of the energy. And that's it for you that renewal energy consumption for the gap of logic, you know, in history neural, almost the same, same distribution of the energy body. And it's basically synaptic transmission, actually it's cost occupy even higher ratio of the energy consumption because of the, the fun read is, is a lot for this in history neural. Basically, you know, we also need to consider the glare cells because you know, for the full radius of half of the cells of glare cells. And the most of the glare cells have the energy consumption basically is accounting for the housing keeping mechanism also counting for the resting potential maintenance, maintenance, because the most, the most, the almost all the glare cells have the resting potential around minus 70 minus 80 millivolts. Except for for those energy consumption, the glare cells also have the cash in response are, you know, along with the nearest activities, you know, you know, a lot of you have the neural entities, you know, response for that for the stimulus glare cells also have the cash in response activities related to these neural activities together to support the neural activities. Those are cash in activity, I think it's also with some other potassium transmission in other potassium transformation, maybe play some rules to supporting the glutamate reception and also for the ionic channels are always, you know, recapture during the actual natural generation process. Okay, so anyway, it's a glare cell is how this is a case in activities is cost energy. So we need to come for that in order to understand the whole brings energy consumption. So we calculate for all of them. And then we use those energy consumption week to calculate the budget for the whole of momentum because for the whole of momentum, it's about 30, you know, 14.3 billions of neurons, most of them are exactly neurons. 20% of them are Egyptian neural. And also for the, you know, cortex, we have the 26 billions of the glare cells, and most of them are astrophlea and oligodendron dendro size. And based on the equations, you know, we use to capture the individual neurons energy consumption process, we could, you know, calculate the whole energy budget for the whole of momentum. And in the same, almost the same way, we could calculate the energy budget for the white man, white matter, you know, most of the cells are glare cells, you know, small, a very small amount of the cells, we could call them a neurons, but still we are not quite sure they are real neurons or thick neurons. But basically, if we call them as neurons, still we have the energy, you know, empty boundary for this white man body. Also, we need to calculate the energy for the most of the neuron neurons come from the green man. It's about 5 billion of the exons come from green man communities between different cortex. So, so after we calculate for that, we could get the budget for the, the, the, the agreement and the white man, the whole green. So, basically, the for the green man in the race is the not signaling process. We are the second in the process is mainly is sick actual potential, actual potential generation and propagation plus the synaptic transmission process. So, not signaling process is basically racing potential maintenance and housekeeping maintenance. So, this is for the green man for the white man is seen. It's the signaling support for the extra potential generation of obligation and standard transmission process for the white man that you can see that it's not the processes are mostly costful components is about 80%, 80% of the whole and for white man, the budget for them. So, basically, we add them together. We could say that for the whole as a cerebral include the white man and the green man in the resting state, the signaling process is occupied is a, you know, consume about 50% of the energy while the not signal process also consume 50% of the energy. So, this is based on our calculation. So, you know, once you calculate the audience, you want to know whether you are correct or wrong. So, in order to confirm our calculations, so we, you know, have the data from the from the year. FMI center, we cover if we have measured a lot of energy consumption process for the brain for the human brain or for the right brain. So, we use a different in the technology like to be or to grab or to read, read you graph and also the MIS and also the different technology to measure the energy consumption for the, you know, cortex and also for the white man. And for the animals in different for the animals also for the human brain in different behavior state like, you know, and it's the ties the state also the week, a week rest or week stimulation. And the brain energy consumption is, you know, for different states have different scale quantity. So basically, if we have more, we have more activities in a week state, we have more action energy consumption there. So, our calculated, our calculated energy consumption is almost the linearly correlated with the measure. We have a question from the audience if we can, please. You mean. Yeah. The application question from Josh Iran. You are allowed to speak. Hello. Yeah, could you help me. Yes, we can hear you. Yeah, okay. So, sorry, what's the questions from from the people. Sorry. We cannot see the question. I didn't see the question. So question from, from QA board or from chat, we didn't see that. Okay. There are no questions in the Q&A yet. You can go ahead. You can answer questions. You can go ahead. Yes. Okay. Okay. Okay. Basically, it's basically it's all calculated the energy consumption for the for the human brain and the rent as compared with the measure that you measure the metabolism are almost almost the linearly correlated with it together in order to do this actually for the for the calculation we use different far read of the, of the for brain, you know, activity and use this far read, you know, compared with the measured activities from the rent from a human. Almost we also have the very nice correlation from B and C you can see that. And also from B, we can say that since all the for the rent in the resting state, the far read or for bring every the far read around the whole heart. The calculation is almost like for her for the human brain is calculated calculated for the resume or bring in the, you know, is around the one heart. Also calculate the measure that activity in the resting state in the wiki state and also around one heart. Okay. So, I will go to next, next, the next the next the next one is, you know, after the validation from the energy body. We, I think we're running short of time. It's already been about half an hour. So can we proceed to conclusions. Yeah, seven minutes to finish. Okay, next one is construct the digital 3D human brain and