Primary Visual Cortex: How brain processes what we see

Key Takeaways

Greetings fellow learners. In this video we are going to talk about the primary visual cortex. So what is it? The primary visual cortex is the part of the brain where initial visual processing happens. And specifically if we were to take a crosssection of the brain the primary visual cortex falls towards the back of this brain. So just to be very clear about the overall visual processing, let's say that there's an object that's like right over here. Light rays are going to pass into the eyes here. They are going to create an image on the retina and there will be nerve fibers that are going to be bundled together to carry this visual information and this will be the optic nerve. Now some of these fibers remain on the same side of the brain. Some others will cross over to the other side of the brain at this point known as the optic kayazm. Now these nerve fibers are then going to culminate in the lateral geniculate nucleus or LGN. From here the information is going to be relayed to the primary visual cortex which will be responsible for some initial visual processing of this information within the brain. Now here's another question. What does the primary visual cortex actually see? Well, if we actually take something called a retinopic map, we can see what information is being projected at different points of the visual pathway that we just described. So, let's say that the eyes looking at this realworld image of person waving in front of us. The image that's going to be on the retina is going to be inverted due to the convex lens structure of the lens and the cornea. And so it's going to invert the image on the retina itself. And as it passes through the visual pathway, eventually it's going to end up in the brain, that part of the brain called the primary visual cortex. And specifically, this is going to be like the left side of the primary visual cortex. So you can see this information here is inverted and a part of this that's specifically the central part that we want to focus on is being blown up and this is due to something known as cortical magnification and this is done to extract certain details about you know the parts that we want to actually focus on. So here's another question. How does the primary visual cortex process visual information? Well, a lot of our understanding of how the primary cortex processes visual information initially started with some experiments in the 1950s and60s conducted by Hubil and Weisel and they conducted experiments on cats and monkeys and some of the highlevel understandings or findings of this included that they proposed a hierarchical model of visual processing. So this suggests that visual processing in our brain occurs in a series of stages where we start with visual processing in the retina followed by the lateral geniculate nucleus and then parts of the primary visual cortex here. And as information is going to be processed, we see that the complexity of the feature that is being processed is going to be increased and it's going to be more abstract features that are processed as we go down this pipeline. Another thing that they also found was that the cells within the primary visual cortex could be classified into two major types. that is simple cells as well as complex cells and these simple and complex cells were best activated by edges and hence they are known as edge detectors. So, how they actually found out all of this was through this kind of experiment where they took a cat, sedated it, presented it with a visual scene, and then tried to shine light onto different parts of this visual scene like you see over here. And they would then record the activity of specific neurons within the primary visual cortex. And broadly when they were recording patterns they noticed first of all that the cells within the primary visual cortex responded best to bars of flight and hence they were best known as edge detectors. Another thing that they noticed was that the cells were also orientation specific. So looking at this figure over here, if you know they connect an electrode to a specific cell, that cell would only fire if the orientation was let's say here in this case if it was vertical. But if the light bar or the edge was in any other orientation, we would not get as much stimulus that's recorded. So that's a highle detail of the experimentation. Let's now get into some details about what simple cells and complex cells are and what are their receptive fields. So simple cells and their receptive fields. So according to this hierarchical model that was proposed by Hubil and Weisel basically a set of lateral geniculate nuclei cells could collectively synapse into a simple cell. Why did they think that this was the case? Well, this is because of the observations that were made in the experiment. So, let's say that this here dotted line is the visual scene itself. And in this visual scene, let's say that we have a bunch of concentric circles here. Right? Now, this set of concentric circles is the receptive field of one of these LGN neurons. So, this is known as a center surround receptive field. What a center surround receptive field entails is that if you shine light like right over here at the center but not on this surround region, this cell is going to be activated. It'll fire an action potential and hence it's going to increase the number of neurotransmitters that are ejected and hence it will increase the likelihood of this simple cell being fired too. Similarly, we have this fourth center surround receptive field that corresponds to this cell over here. So if we shine light here, right at this center, this cell is likely going to fire an action potential which increases the neurotransmitters emitted by the cell into the simple cell and hence this simple cell is also going to be higher or more likely to fire an action potential too. And so what you can imagine here is that if we shine a bar that is in this orientation specifically right over here then this simple cell will have the greatest chance to actually fire. And this is exactly what we observe in the experiment. When we shine a bar of light at a specific orientation at a specific position on the visual scene, we do see that some cells in the primary visual cortex do actually activate. They fire action potentials. And these cells are known as simple cells. And so we can see that these simple cells detect edges of a specific orientation on the visual scene. Now changing the electrode at different positions within the mammal's brain, Hubin Wisel noticed these simple cells, but also another type of cell behavior. And these are cells where if you take like a bar of light and you move it in a specific direction across the visual plane or the visual space, we see that these cells continuously fire action potentials. So what this entails or the thought process here is that this encompasses multiple simple cell receptive fields. And so the idea here is that we have multiple simple cells that will synapse into a cell over here which we call a complex cell. Now the idea of calling it a complex cell is because you would think that if you just shine bars of light anywhere over here or here or here where we have a receptive field of a simple cell, you are going to see this cell complex cell activate. But it's actually not that simple because not only does the receptive fields require it to be like an edge of a specific like angle or orientation, it also has to be moving in a very specific direction. And because of this complex interaction between simple cells and this other type of cell, this other type of cell is given a name complex because of its more complex higher abstract level interaction which also shows that it processes you know higher level information like motion. And so what this experiment kind of says at a much higher level is that as light travels from the retina all the way to the brain along this visual pathway the complexity of the features detected are going to increase. So for example in the retina the features detected are like color then light is normalized and there's also a retinopic mapping that is created. Then in the next layer at the lateral geniculate nucleus, this is going to preserve the retinopic mapping and it also detects color along with contrast and this is possible due to its center surround receptive field and it also reconciles information coming from both eyes. And then next it's going to land into the primary visual cortex where we have a simple cell. This is going to be your edge detector that also detects edges of a specific orientation which these will then synapse into complex cells which will detect edges of a specific orientation and also detect motion. And then we keep going further down into the next layers of the brain that detect curvature, complex shapes and more. And so, as you can see, through this hierarchical model, the complexity of features detected increases. Quiz time. Have you been paying attention? Let's quiz you to find out. Which of the following is true about simple and complex cells? A. Simple cells respond to motion while complex cells respond to only orientation. B. Simple cells have distinct excitatory inhibitory regions in the visual space while complex cells do not. C complex cells typically integrate information over a wider area of the visual space than simple cells. Or D complex cells are found earlier in the visual pathway than simple cells. I'll give you a few seconds to answer this question. The correct options are B and C. But can you tell me why? Please comment your reasoning down below and let's have a discussion. And if you like this video at this point and you think I deserve it, please do consider giving this video a like because it will help me out a lot. Now, that's going to do it for quiz time and this video. But before we go, let's generate a summary. So we started off with the definition of a primary visual cortex at a high level. It's the part of the brain where initial visual processing happens. And then we took a look at the visual pathway all the way from the eyes all the way down to how that information transmits to the primary visual cortex over here. And we also ask the question, what does the primary visual cortex see? in which case it looks at an invertical cortical magnification mapping of the actual visual scene. Now how does the primary visual cortex process visual information? Well, we took a look at Hub and Wisel's experiment on mammals as they proposed a hierarchical model where visual processing happens in a series of stages and the complexity of visual features processed increases through the stages. Here during the experimentation they discovered two main types of cells as being simple cells and complex cells. Both of them were edge detectors. We saw that multiple LGN cells can synapse into a simple cell and saw how that was the case with their observed receptive field behavior. Similarly, this model also proposes that multiple simple cells synapse into a complex cell and also justified this by what was observed by the behavior of their receptive fields too. And overall we kind of just laid out this like table which has layers of the visual pathway along with the kind of features that are being detected along the way just to highlight the point that as we go down the visual pathway the complexity of visual features detected increases along the way. So I hope that this information all makes sense. There are a lot of resources here back, you know, dating from the 1950s and60s experiment that I'm going to link down in the description below along with some cool video links of their experimentation results. So, please do check those out. If you think I deserve it at this point, please do consider giving this video a like. It'll help me out a lot. Thank you all so much and I will see you in the next video where we talk more about how all of this is related to starting the field of computer vision. Thank you and I'll see you soon. Bye-bye.