Christof Koch Interview: Conversations with History; Institute of International Studies, UC Berkeley

Consciousness and the Biology of the Brain: conversation with Christof Koch, Lois and Victor Troendle Professor of Cognitive and Behavioral Biology at Caltech, March 24, 2006, by Harry Kreisler

Page 4 of 6

More Thoughts on Consciousness

Do the combination of these experiments over time, pursuing this framework -- do you hope it will lead to a general picture that answers some of these specific questions?

Yes, of course. There's really no alternative.

Consciousness is a difficult problem. We as humanity have thought about it, in the Western [historical] interpretation, at least since 2300 years [ago], and the only way to approach it is in an empirical manner, step by step. You try to discover something about the material basis. Ultimately, of course, you need a theory of consciousness. You need not just the neural correlates but you want to understand what types of systems, and at what conditions, have conscious sensation.

I'd like to know, for example -- in principle I'd like to know -- what about a young baby that doesn't speak? What about a fetus? What about the patient who is in coma or who has a persistent vegetative state, like Terry Schiavo? I'd like to have a measure for theoretical understanding: is she or is she not conscious? What about a monkey, what about a dog, what about a squid, what about a fly? Right now, we don't have a good theory, for example, for insects.

We accord consciousness to the degree that other people are like us. You can go back in history. Even the Greeks, they first thought, "Well, the only true humans are other Greeks." Right? The other ones were barbarous. And we decided slowly, the last couple of thousand years, "Well, we'll extend this privilege of being 'real humans' to other people, and to other races." Now we're beginning to extend this magic circle to other animals. Yes, a monkey or a dog, they don't speak, they're a little more furry, they look a little bit different, but their brain is very similar to your mind. If I give you a little bit of monkey brain or dog brain, and a little bit of human brain, it takes an expert to really tell the difference. The behavior is very similar, appropriately measured. If a dog's in pain, that behavior is not that dissimilar to when you're in pain: it yelps, it has a higher motor activity, it wants to run away and hide, similar things to [what] you do.

Evolutionary continuity, all those reasons, suggest that these animals are also conscious. They're not as conscious as we are, they don't know about Macintosh computers or things like that, but they certainly seem to have all the basic conscious sensations of pain and pleasure, and they can be depressed, they can be happy, they can be excited. Now where does it end? What about a squid, and what about a bee? A bee is a very complicated organism.

So, finally, we need a theory that tells us which organism, including which artificial organism -- at what point can you say, for example, if ever, that the internet is conscious? What about a robot, what about a computer? So, we're finding what we need is a theory that tells us which system, under what condition, will be conscious.

What about this notion that one had in earlier periods of time, of the notion of a "little man"?


Yes, homunculus in your head. Obviously, there is no little man there, but is that a kind of hypothesis that one can work with, and do you still try to use it? You're talking about something that moves beyond the zombie effect where I believe you say you're responding to your present environment, but also it's related to the planning area. This is the point at which we need some entity, some unit, to be making choices.

It's a very compelling illusion. We all have it. There's a "person" actually sitting in my head, I can tell you where -- he sits here, roughly here. If you [have only] one eye, I presume it would shift it over. He's looking out at the world. We all have this homunculus illusion, it's very compelling. Francis Crick argued there has to be some neuronal basis for that to explain this very compelling feeling, and we have an explanation for that.

But of course, as you point out, there is no homunculus, it's just a series of very congregated neural networks, and one network is "looking at another network," and that's being looked at by another network, and ultimately what you get out of this system -- you get this feeling that there is a person sitting inside my head. But there is, of course, no point where you can say, "Well, see? That area, that's Christof." It's going to be, like most other things, a collection of neurons that do this process that ultimately gives rise to this mysterious feeling, there's a person inside my head.

One of the images that struck me in the essay was the notion that there are elements of the brain that are like a movie. Could you help us understand that, that in other words, motion may come out of a series of images that are in your brain, and presumably the images are an output of these coalitions of neurons? Help us get a handle on that.

All right. There's a hypothesis: is consciousness like a river, or is it more like a movie? We all know we have the illusion of seeing life on the screen, but in fact, if you look closely, there are 24 frames a second, or 72 frames a second, and so you have a frame like this, the next frame like this, the next frame like this, but what you see, of course, is smooth motion. Psychologists have asked this question for a long time and there's now more evidence for it, that perception is actually discrete, it's actually organized in discrete snapshots. Each snapshot may be 30 milliseconds, 15 milliseconds, 100 milliseconds, so a small fraction of a second, and each snapshot is a little bit frozen, so it's like a drawing that you make of a runner like this. Clearly this drawing implies that person's moving forward but the drawing itself is stationary, but the stationary drawing comes with the sense of motion.

There is this area that people have just identified called MT, you have it on both sides of your brain, or sometimes called V5. It seems to be involved in motion. If you stimulate it, there's a sense of motion in animals; if you lose it, you lose a sense of motion. The hypothesis is that these neurons in this part of the brain encode motion in a series of these snapshots, and usually, just like in a movie, you're fooled by it. Although these are discrete snapshots, the world looks continuous.

There may be a rare condition, neurologist Oliver Sachs described one of these (he himself has had this condition under migraines), where this mechanism breaks down. He describes this very vividly himself, and also the experiences of some of his patients. He says the world appears shattered, atomized. "I see the nurse comes towards me and she's in these discrete stills. I don't see her smooth anymore. It's like a movie run too slow." This well may be a manifestation of something that's more general, that maybe all of our perception is actually illusory, this smoothing, changing river of consciousness may actually be illusory in the sense that it is actually a discontinuous set of snapshots but we don't experience this as such. It would be another example of this perfect con job that is perception. There're so many things that are not really the way they seem but our brain suggests them, and our brain is very good at that.

Is it the case that the brain will often substitute information where there is no information? Will it fill in frames?

Yes. For instance, the blind spot. We well know, at least since the fifteenth century, there is a spot called the blind spot, where there's no information. You can find it roughly here. There are no photoreceptors there, because in that location the Ys leave the eye. At that location there's no visual information. It's like you take your camera, your video camera or your consumer camera, and some of the pixels are black, they would always be black. You would immediately send it back. Well, why don't we complain? Well, it turns out there's a mechanism in the brain that compensates for it. So, I don't see a hole. I sort of have to test it, but I think, well, okay, somewhere around here my finger disappears. Again, it's a con job. It fills in the information. There's no information there, but it cleverly fills it in.

There are many, many cases like that where the brain fills it in, usually using appropriate stimuli, because it learned over the course of evolution, it makes sense, on average it makes sense. But if you test it carefully, you can see there is a lot of this filling constantly going on. Even at the high level, the cognitive level, I constantly make implicit assumptions about people, about their behavior, and I do it all the time unconsciously. It's just another example of the brain filling in and predicting things that are missing or predicting things into the future. That's what our brain does.

Now in addition to designing experiments, I guess pathologies and disease are a fruitful place to look. This comes out in Oliver Sachs' essays in the New York Review. Talk a little about that, because they offer the possibility of the breakdown of the systems, and then a testing of the map of the mind at that time.

Probably one of the most richest sources, the most fertile sources, of information about consciousness is humans, because we can talk about it. You have to go to the clinic where you see a lot of sometimes bizarre conditions but that tell you a lot about consciousness. So, for example, for the prosopagnosia, you're unable to see a face (and there are different forms of it). The patient says, "I can see your mouth, I can see your eyes and I can see your ears, so I know there's got to be a face because I know intellectually they go together, but I actually don't see it." It turns out that very often a particular part of the brain is destroyed. That tells us a really important point. That tells us consciousness, at least for faces, is localized. It's not global, distributed, holistic, gestalt-like, as a lot of people think. In this case, there's actually discrete foci. There are a bunch of neurons that code for faces, that code for the conscious perception of the faces. If you lose them, you're unable to see faces. You can still think about them abstractly but you can't see a face anymore.

Another very important discovery was found at Caltech by Roger Sperry, based again on the brains that don't work very well, namely epileptic seizures. Sometimes to prevent seizures from going from a small attack, from a so-called focal attack, into a grand mal attack where people thrash around, you cut 200 million fibers connecting the left brain and the right brain. What they then realized was that you actually have two consciousnesses, two minds in one skull. You have the left hemisphere, usually by and large that talks, that's usually the linguistic competent one. The right one can sing, the right one can grunt, the right one can answer simple questions but it's not very sophisticated, but it seems to be conscious. So, now you really have the dilemma -- you have two people, in a sense, two conscious entities inside this one skull. It's a really interesting discovery.

We work with patients, we work with a neurosurgeon, Dr. [Itzhak] Fried at UCLA, again with epileptic patients, where we can record for many days from individual nerve cells, something you can never do otherwise. You can discover very interesting nerve cells in the brains of these patients. So, again, the clinic has been one of the most fecund sources of information about the brain and about consciousness.

Ultimately, the insights that you draw on perception potentially could lead to insights that are relevant for the other areas of consciousness -- smell, and so forth. Is that the hope here?

Yes, just like in molecular biology, Harry, where people argued early on, like Max Delbert, for example, who said, "Let's study phages," -- phages are viruses that prey upon bacteria -- "And let's study the genetics there." Some people argued, "That's so remote from human genetics, it's not going to tell us anything." Well, it turned out it's the same mechanisms, RNA, DNA and proteins, the same mechanism. So, here the hypothesis is -- it may be wrong -- that consciousness is something biological, it evolved we don't know when, 100 to 200 million years ago, who knows, but it's something that probably first came with consciousness for simple things like pain and pleasure, then evolution discovered it's a great thing. We think the function is to plan. It probably used the same mechanism for visual consciousness, for olfactory consciousness, and later on in more complicated animals like great apes, and like dolphins, and ultimately like us, also for self-consciousness, knowing who I am and projecting myself in the future of going through different scenarios: if we have an earthquake here, which door did I come in, which door should I try to egress. All these different forms of consciousness have something common. They have this subjective consciousness in common, but they also will have or may have something common at the neurobiological base. They're all caused by a similar set of neuro-mechanisms. That's the hypothesis.

Next page: Cutting-Edge Science

© Copyright 2006, Regents of the University of California