Okay so I just gave you a few highlights. A few examples of the notion of the brain as a computing device, computation. And that's by the way why the field computation in neuroscience is so powerful. We discuss it later on because basically we believe that the brain, the main mission of the brain. Any brain is to compute, as I discussed before. Let me show you examples that single cells. When you record from single cells in the behaving brain, in the behaving animal. The single cell already shows a reminiscence or aspects of computation. So I'm going to discuss computation at the level of single neurons and the most important probably, direct example. First direct big example is by the two Nobel Laureate The Hubel and Weisel experiment for which they received a Nobel Prize in 1981. So I'm going to show you a little bit of a movie showing the original experiments of Hubel and Weisel. Recording from the cat visual cortex, particular region of the visual cortex of the cat, v1. The visual cortex at cortex area one of the cat. Implanting an electrodein the living seeing cat. And trying to find out what are the parameters being computed, represented by a single cell in the visual cortex of the cat. So here is [CROSSTALK]. >> What that looks like. >> [CROSSTALK] So this is David Hugel. >> Primary visual cortex, this is the part that's tucked underneath. [CROSSTALK] This is the visual cortex of the cat. >> Now you're going to see that a layer structure, some places the cells are packed tightly, other places they're looser. and underneath every square millimeter of cortex, which is about that much, you have about 100,000 cells. >> The researchers actually listened in to individual nerve cells firing in the anasthetized cat. >> [CROSSTALK] So they record from single cells [INAUDIBLE] says, in the cat. >> When we started working, Toreston and I, in the late 50s. >> We set up our first experiment. [CROSSTALK] >> This is David Hubel. >> and they didn't go well. [CROSSTALK] This is Torsten Wiesel. >> At the beginning, we couldn't make the cells fire at all. [CROSSTALK] And nothing seemed to work. >> [CROSSTALK] So the cat is looking at the screen, and they record from this single cell. >> by accident. One day we were shining small spots. Either white spots or black spots onto the screen, and we found that the black dot seemed to be working in a way that at first we couldn't understand. >> [CROSSTALK] and suddenly they hear the firing of the cell. >> [CROSSTALK] the process of slipping the piece of glass into the projector, which swept a line, a very faint, precise narrow line across the retina. [CROSSTALK] >> Can you hear the tac tac tac. >> And everytime we did that we'd get a response. >> Now okay, so the cell was firing, a particular cell recorded indicate visual cortex. The cell was firing in a particular, in a particular occasion and what they found. And this was the Nobel prize, later they developed the model for it. What they found is whenever there is a line, a line crossing the screen in a particular angle. Then this particular cell started to fire, meaning that the cell responds best when there is a line. In the visual world, when there is a line moving, let's say in this direction or in this direction. So let's understand it for a second. Let's try to understand waht is it that they show. They saw Hubel and Wiesel recorded from v1 which is part of the visual area. Of the cat, of mouse, of, of, of humans, so the visual cortex is larger than V1. V1 is the primary visual cortex. They record it from here, from a single cell showing while the cat was looking at the screen, showing different Angles of moving lines. And then they found that when they record from this particular cell, one cell by chance. One cell. They found out that this cell, when there is a line crossing the screen, suddenly this cell starts to fire. And when the recent line crossing the screen in a lit, in a different angle, this cell does not fire. So there are spikes responding, coding for this particular direction of line, in this particular cell. So we can call this one cell an orientation direction selective cell. In this case when this orientation of the line is moving in this direction. And just to summarize what Hubel and Wiesel found in this experiment. Is that when you have an oriented line, 180 degrees or other orientations on the visual sys, on the visual world. And you record from cells in V1 of the cat. In some angle, for this particular cell, the cell fires more rigorously. More, more strongly. In other, in other angles, the cell does not fire. So, for this angle, there are no spikes. For these angles, there are no spikes. For this particular angle, a little bit of spikes. And for this angle, in this, in this case, 90 degree angle. There is a lot of spikes. There are a lot of spikes. Less spikes, no spikes, no spikes. There is a tuning for this cell. This cell is tuned to respond to this angle, not to this angle, and a little bit to this visual angle in the word. This was a breakthrough. This was breakthrough in retina, you don't have cells that responds to line oriented lines. In the deeper region, the thalamus there are no cells that respond to this oriented lines. But in the next level deeper into the brain suddenly there are these computation. You compute. The cell computes. The cell responds, computes, to a particular angle, not to others. Okay. So this is a orientation-selective. And in this case also moving the direction of a cell that is responding to particular parameters in the word computes the parameters. So some of your cells now respond to my nose because some of your cell in your visual cortex are sensitive to 90 degrees. Some of other, other cells respond to my eyes, which is 180 degrees, and so on. So that's how you decompose the world, apparently. We and the cats and the mouse and monkeys, decompose the world. Early on in the visual system, in the v, v1, early visual system, you decompose the world into lines. And also to other parameters. movement, to color and so on. To edges and to other aspects you compute the aspects of the visual scenery scene. And you decompose it for its features, oriented lines, length Ending of lines, moving of lines, movement of lines and so on. So this is at the single cell level, you can see which is not surprising because somewhere a computation has to be done if you recognize a line. Some system must respond to this line. So some neurons must respond to this line. But Hubel and Wiesel showed where this system is.
