 When one speaks about the brain, neurons come to mind, and more specifically the complex network of a critical signal that's travelling constantly throughout the brain. And all brain functions do depend on this communication network, and we saw that in the previous talk. However, neurons are not static objects that are fixed in time. Those are deformable cells that can grow, squeeze into type places, push through the cells and potentially die. And this is all dictated by the mechanical properties of the cell themselves. And the protein network right underneath the cell membrane that's filling external forces and potentially changing configuration in response to them. As any type of structure, neurons can be damaged reversibly or irreversibly. And the stronger and the faster the force you apply to them, the more probability is that the functional properties of the neuron will actually be affected. To tackle this challenge, people have been using different means. And one of them is computational modelling. And we went from 1970s really rough head model to very advanced patient-specific MRI and CAT scan-based head model that can be put in different situations that we'll see afterwards. The technique that's being used in general is called the finite element method in which you take your real geometry, you discretize it in small elements, and you associate mechanical properties in each one to each one of these elements, whether it's stiffness, strength, of ductility. Once you have this head model, you can then put it in an environment of your choice and apply external forces to it. And it can be an impact, a punch, or even shock waves. You can then simulate how the whole of this head and all the organ inside it deform. You can measure the pressure and shearing and all these kind of forces. You can then zoom in into one area of interest and then transfer those forces at the cell level and actually calculate how the cell is being deformed. However, neurons are first and foremost communication machines. And knowing how much the deform doesn't tell you much a priori on how healthy they are. Also you need to go down one more scale and you reach the protein scale. And you're using a technique that's called molecular dynamics in which you actually tracked the movement of each one of the atoms individually. And we're looking here at an ion channel. And why do we look at an ion channel? One ion channel is an important protein in neurons. When a signal is reaching this ion channel, a trickle signal, they do open up and organize an exchange of ions in and out of the cell. This potentially change the electrical voltage, the potential, right underneath the membrane locally and that participates to the propagation of the signal. So you can easily understand now how deforming such protein will actually change the electrical property of the neuron itself. And this is one of many examples of the protein that's being damaged when submitted to an external force. Once you know that at the protein level, then you can go back up the scale and basically reach again the whole neuron level. And then if you send a signal on the right-hand side of your neuron, damage it in the middle. You can then measure the alteration and electrical signal on the left-hand side. And that's done at the neuron level. Again, you have different techniques that can be used to scale that up again to the neuron network, actually first neuron interaction level, then small network, then large network, and eventually cortical column. And this is actually done by the blue brain project originally, now the U-band brain project in Europe that is from actually which those pictures have been taken. Once you know that, at large network level, you can finally reach the brain level, that's the main idea, and you can map electrical signal at the organ level to pathologies of traumatic brain injuries such as short-term memory loss, epilepsy, or even headache. So let's sum up a bit the approach that I propose here. We're going to the human scale, and you're applying an external force to the whole head. Then you're going to go down to the organ, the brain, the cell, and then eventually protein using numerical technique at each level. Then you're transferring mechanical property information to functional properties, whether it's electrophysiology or biochemistry. And go back up the scale to the overall brain function level. However, the brain is not just a V-shaped process, right? This is an optimized organ, by nature, that will feel that something is wrong and try to use all these vertical scales and horizontal properties to try to find another optimization state in which you would be healthy again. And we'll see that a bit later. I have a question for you. In my research, I need to go down to this very small scale. But how can we actually link together mechanical properties and functional properties while staying at the macroscopic level without this very expensive numerical way to go down at the micro scale? So how do we link mechanics and function at the macro scale? And that'd be my question for you.