Iain Couzin is an Assistant Professor in the Department of Ecology and Evolutionary Biology at Princeton University, where he manages the Couzin Lab. His research focuses on collective behavior and self-organized pattern formation in a variety of biological systems, including fish schools, bird flocks, insect swarms, human crowds, and cellular networks.
Question: As an expert on animal collectives, do you consider yourself part of the crowd?
Iain Couzin: Well, I feel myself very much part of the crowd and part of society and sort of, but I'm also susceptible to, you know, being swayed in a similar way to other people. And so one of the fascinating things when we look at nature, when we look at collective behavior in nature, is how prevalent it is, you know, from, you know, looking at ants forming a swarm to, you know, bees, to, you know, fish schools, bird flocks, you know, all around us, and of course, our own human societies. And so it really, what I'm trying to understand and what I'm fascinated by, is how and why natural selection leads to individuals living in societies or living in groups.
Question: Is your favorite band, the Pixies, non-conformist or mainstream?
Iain Couzin: When you're old like me, no, I mean, in the late '80's and the early '90's, the Pixies were revolutionary. They wouldn't have been bands like Nirvana had it not been for the Pixies and, you know, Kurt was one of the first people to say that. You know, they were incredibly influential. And of course, the nice thing about music as well as science, is that things can build upon other things and Pixies had an enormous role to play in that and had a sort of influential, you know, hit me at an influential time in life. So I love them as a band, in fact, I saw them here in New York City just a few weeks ago and it was fantastic. And I was one of the oldest people there! I think the median or mean age was about 22, 23, I was very surprised and very impressed.
Question: What are some of the most surprising behaviors animal groups display?
Iain Couzin: Well, I started working on ants because I was fascinated by how very small animals, you know, with relatively small brains, can live in these large societies. And one of the things I was studying was traffic organization. And so I decided to look at an ant that's literally blind, so we have a case of the blind leading the blind, they're called army ants, they live in Central and South America, and you literally have a million and a half of these blind individuals living together and forming huge swarm raids across the forest floor and collectively able to subdue much larger prey items than themselves, which they rip apart and they bring back to the nest. And what I found was that these ants, even though they're blind, even though they're just interacting locally, would spontaneously form lanes on the trails to minimize congestion.
So whereas we, you know, I just walked through New York City and there's congestion everywhere, these ants have just developed very simple rules to minimize that so they can maximize the flow of resources back to the nest.
Question: How do some animal groups achieve such extraordinary coordination?
Iain Couzin: Well, the ants are exceptional in some respects because they have a very strange genetic relatedness, it's called haploid diploidy, which means that they're more closely related to each other, their sisters, than they would be to their own offspring. And so the workers actually forego reproduction to benefit the colony. And that's quite a special type of evolution because it means that the ant colony itself behaves as some people have described as a super organism. You know, the queen is the productive entity and the workers then work for the benefit of the whole colony. And so there we have some of the most exquisite collective behaviors, because of course, it's in everyone's interests to do so.
Now if look, you know, at the other extreme and say human crowds, we actually get some similar types of patterns. If you were to look above a crowd, sort of unbeknown to you, you actually also spontaneously form lanes. You fall in the slipstream of other individuals walking in the same direction as you. But of course, you know, you may want to be minimizing your own travel time, but you don't necessarily care whether you do so at the expense of other people. And so we see this when we look at vehicular traffic, on the motorways and highways, people can often be very selfish and this can lead to congestion. And the ants don't have that type of problem because they don't have the sort of genetic conflict.
Question: What are the survival advantages of acting collectively?
Iain Couzin: Well, when we look at highly coordinated behavior, say schooling fish. I mean, you've probably all seen in a television or even, you know, if you've gone diving, this incredibly, highly-coordinated group behaviors. And to be honest, it's still a bit of a mystery as to how and why these groups form. We know that **** pressure causes individuals to come together because if a predator was to attack, it's always better for someone else to be taken rather than you. But in actual fact, the coordination of these behaviors also leads to a very interesting properties. It gives individuals within these groups access to what we call higher order computational powers. And what do we mean by that?
Well, in actual fact, you know, one individual, or a few individuals detecting the predators to move away and this can lead to a whole wave crossing the group. And in actual fact, because these groups are often quite large, they can process information from multiple sources simultaneously. And what we've discovered is that the way that the individual interactions are tuned by natural selection, is such as to make very fast, yet very accurate collective decisions. And decisions that are very difficult for individuals alone. So by being within a group, they're actually better at sensing and responding to the world around them.
Question: What’s an example of the power of collective decision-making?
Iain Couzin: So in one experiment, we were interested in disease within populations and typically grouping is always considered to be risky because it's, you know, increases transmission of disease. But, you know, if we are right, if there is some form of collective intelligence, then perhaps by being in a group, you're better able at discriminating between diseased individuals and non-diseased individuals. And so using model replica fish, one of which had a sort of a little parasitic infection, you know, which we painted on the side of the individual and one that didn't, we can then ask, "Well, what does an individual do?" An individual is pretty bad at making that decision, it's very difficult for it to do so, so it gets it right maybe 60 percent of the time, maybe less. However, when we get to groups of four or eight or sixteen, the way that they're interacting actually allows them to integrate their information. And in actual fact, the accuracy shoots up and they're able to make very accurate decisions and as a group they're able to avoid the diseased individual. And this is just one example of this type of collective decision making.
Question: How much non-conformity do we observe among animal groups?
Iain Couzin: Non-conformity within these groups often tends to be very dangerous. So individuals that stand out within the group are often those that are taken by the predator. So there's a strong selection pressure, not only to look like the other individuals in the group, but also to behave like other individuals in the group.
And also, if one were to consider a predator sort of attacking a fish school or a bird flock, you know, the reason that the concerted motion, these waves work so effectively, is because everyone is essentially following a similar type of rule. If you tried to do something completely different, you could of course cause congestion, you could cause a pile up within these groups, and that would of course be selected out of the population and so individuals that follow such rules tend not to last very long.
Question: Does acting collectively ever harm animals?
Iain Couzin: So, you know, in a wide range of scenarios, there is a risk that animals face. By copying the behavior of each other, they risk a sort of cascade of misinformation. And this is actually akin to what happens sometimes in human societies when people start believing something without actually checking it out themselves. So if a decision depends on that made by a few individuals at the start of the process, then everyone then joins in and decides that they agree with that without actually having assessed the information directly themselves, then this can lead to the sort of autocatalysis and the spread of maladaptive information.
Now what we found in animals is that they're, the way that they interact with each other is tuned to minimize this risk. But, you know, there's a compromise as well because it does sometimes happen that they do make mistakes, of course. Because if you really were to sort of look at all of the information and assess it, that takes time. And time is precious. And so just like, you know, a human when you're looking for say, an insurance policy or whatever, you don't actually read all of the details, all of the fine print of all of the alternatives. You tend to look online and perhaps believe what Google tells you, or perhaps you ask a few friends who you trust. And then you rely on that information.
And so, you know, we also have this type of positive feedback, so our decision making, as well, is interested by these social processes.
Question: What is the truth behind the cliché of suicidal lemmings?
Iain Couzin: It is, of course, a myth. The lemmings don't actually hurl themselves off cliffs. You know, these animals are highly susceptible to population fluctuations. And so the predators tend to sort of increase in number and the rodents can increase in number and so on. And this can actually lead to vast fluctuations, so one year there may be many, many individuals, other years very few. And also, I can't remember the exact example of the lemmings, but there's huge potential in these populations for mass migration, for individuals to migrate en masse. And, you know, even in North America, the squirrel that we see everywhere, the grey squirrel, used to, once upon a time, not that long ago, a few hundred years ago, used to form mass migrations as far as you could see, there are descriptions of squirrels and they, when they come to rivers, they all swim across the rivers and people would go out with, you know, guns and dogs and, you know, and just, it was a phenomenon.
And, you know, also, there's an extinct bird now called the passenger pigeon, which also used to live in the United States. When these flocks were so dense during their migrations that the sky would go black. You know, and these, of course, have been the, the passenger pigeons have been wiped out by man and the squirrels' environment is now so fragmented, the forest is so fragmented, that no longer do they form these mass migrations. But these kind of collective behaviors are actually ubiquitous, you know, that all scales from, you know, bacterial colonies to, you know, very large aggregates like locusts, that we study, which are some of the largest groups on the planet.
Question: Do cellular collectives behave like animal groups in any way?
Iain Couzin: Well, we are just beginning to start working on collective behavior at the cell level. We're sort of particularly inspired by some recent discoveries with soft body tumors and so we're very interested in issues of leadership and consensus decision making based upon our understanding of local rules. So although we started with animal groups and, you know, we work with animal groups extensively, I think these types of understanding actually allow us to get a different type of insight into cellular systems, and so this is a very new exciting sort of direction for my laboratory.
Question: What practical applications might this research have?
Iain Couzin: Well, we're certainly looking at cancer, we're interested in how and why collective migration is so common. So until relatively recently, it wasn't realized that, you know, collective behavior is so prevalent in soft body tumors. And now it's seen as one of the primary modes of spreading. And so that's something we want to understand.
And another very interesting aspect to tumors is the sort of the level at which selection is acting. So we're trying to simulate using new computational simulations, we're trying to simulate very large aggregates, where we can keep track of individual identities, we can look at mutations, we can see how this leads to individual behaviors that then scales up to a collective phenomenon. So as I mentioned, it's a new area for us. You know, we're still heavily working on animal groups, but it's very exciting to think that there really is this cross-understanding because the evolutionary principles and the principles of understanding how interactions scale to collective behavior, tend to be relatively universal.
Question: Could your crowd simulations of smaller organisms facilitate similar studies of humans?
Iain Couzin: Yes, we actually simulate across the spectrum from cells up to schooling fish and to human crowds. And one of the challenges is and has always been the computational power that you require when you're actually simulating these individuals. Individuals have to look at each other and see, you know, am I within range to interact with you, and so on. And as we start increasing the number of individuals, the computers tend to sort of chug to a stop and it's very difficult to work.
But there's a breakthrough. In the last couple of years, there's been programmable video game cards. So these cards that have been developed for, you know, gamers, so they can, you know, live in these virtual environments and so on, actually have hundreds of processing cores on them. And this has been an absolute revolution in terms of scientific computing for us. So we're investing heavily in our efforts to try and program all of our simulations on these video game cards. And to give you sort of a rough impression, we're getting around 300 or more times faster. You know, and if computing gets twice as fast or three times as fast, then that's wonderful. But if you can get 300 or 500 times as fast, if what used to take a month now takes you an afternoon, that changes the way we work.
And also, because we can harness this vast computational power, we can start asking questions about evolution, we can start simulating these groups of reasonable size with the reasonable resolution in how they interact in space over such long time scales that we can now start, you know, having a sort of virtual process of evolution to understand how and why collective behavior has evolved.
Question: What are the similarities and differences between animal and human crowds?
Iain Couzin: Well, in many of the animal groups that we're looking at, you know, I mentioned ants, and they're a sort of special genetic sort of closeness, so we can exclude them for now. But even when we look at selfish organisms like the fish, you know, they're all locusts, you know, they've evolved to live within incredibly large groups and incredibly large societies. Now, we humans haven't. You know, we're thrown into these large cities now, but that's not actually how we evolved. And so we don't have sort of specific rules to optimize our situation here. And furthermore, you know, our mode of communication, verbal communication and written communication, is relatively slow compared to the sort of the mass media and all of the information that we have now. So there's some sort of fascinating challenges from a biological perspective to our societies.
But what we do find is that people are good at learning how to behave in crowds. For example, if you were to take people from, I don't know, the center of America, who have never been to a large city and pop them in the middle of Manhattan, not only would this be a sort of cultural shock, it would actually be difficult for them to behave within these crowds. These are sort of learnt behaviors that you sort of try things out and you eventually kind of optimize how to behave within these environments.
And then we film and we track people within crowds, what sort of astonishes me is how predictable, from a statistical perspective, how predictable crowds are. And so each individual, we can't necessarily predict, you know, whether you're going to go left or right, but when we start getting large groups of individuals, we can predict properties like how many lanes will form or, you know, what are certain times of day who will generally tend to move where. And we can also make predictions about congestion in these environments, and so forth.
Question: Are there species that are more individualistic than humans?
Iain Couzin: Yeah, I mean, there are, I mean, group living isn't, I mean, it's very, very common, but it's not, you know, the only way in which to live. And in fact, we've been studying some fascinating organisms that actually flip between a solitary existence and an incredibly strong collective existence. So one example are locusts. Now, locusts are sort of well known as being sort of biblical plague and so forth. But they're still a real plague. They can actually cover or invade up to one-fifth of the earth's land surface during plague years, which is astonishing. And the Food and Agricultural Organization of the United Nations estimates that the impact to livelihood of 1 in 10 people on this planet to the impact on agriculture and so on. But of course, if you look at the range over which these huge hoards sweep, it's across Sub-Saharan Africa into Asia. And so there's not much financial incentive to do research in this area, but there's, of course, a big humanitarian incentive and it's also a fascinating problem.
And so what locusts actually do is, when times are good, they will avoid each other. They're shy, they're green, they're cryptic grasshoppers, they can't stand being near each other. It's only when times get hard, and the physical sort of contact or the sights and smells of others, causes them to change. And they literally change their color, they become sort of black and yellow with red eyes, and they start marching together, they start swarming together across the landscape. And this is before they get wings, you know, when they're adults, they can then fly together, which disperses them even further.
And so collaboratives of mine at University of Oxford and now at Sydney University, discovered this behavioral switch and so what we were fascinated by was, you know, why are they all swarming together? Why are their kilometers-long locusts all marching in unison, it looks like a cooperative behavior. But what we discovered, is in actual fact, it's driven by cannibalism. When times get hard, they're resorting to eating each other. They run short of protein and salt and water and they start to eat each other. And so the optimal strategy seems to be to try to eat those ahead of you, but try to avoid being eaten by those coming from behind and the outcome is everyone starts to move together collectively. And of course, in that way they can also benefit by saturating predators, they can also benefit by potentially finding new areas of resource. But we discovered that cannibalism is actually critical here. So there's always a surprise when you look at collective behavior. You know, even when something looks really coordinated and really cooperative, you have to look twice and see whether in actual fact, it's driven by aggression.
Question: Will this new knowledge help solve the locust problem?
Iain Couzin: We now understand that this, you know, the limiting resources seem to be protein and salt in many circumstances. And so one of the big challenges with, I mean, there's no way we can just go and like pile loads of protein in these areas because frankly, if you can find the swarms, you may as well apply some sort of pesticides.
The problem is finding them. I mean, they occur across vast areas, no one knows where the egg laying fields are, no one knows when and where these swarms are going to form. So if we can link together our understanding of the biology with our understanding of weather patterns, geography, using what's called hyper spectral imaging from satellites to give us some information about the protein content of vegetation, what we're hoping is it might be possible to actually start building predictive models and using the type of simulation techniques that we've developed to scale up from, you know, in detailed individual interactions up to the behavior of millions of individuals, we're hoping that we can make some breakthroughs in control.
Question: Why do some human societies seem more conformist than others?
Iain Couzin: Well, I'm not an expert on human society and I would defer to, you know, anthropologists and social scientists on this. But, you know, it is certainly clear from some classic work in psychology that, you know, that we have a great capacity of influencing each other and not perhaps being fully aware, you know, the fact that we are influenced.
I'll give you a sort of a toy example, there was a classic study by a guy called Stanley Milgram here in New York City in the '60's, and he was interested in how people sort of copy each other within crowds. And, for example, they'll copy the gaze of other individuals, so he had some students looking up at a window and it was just a silhouette of a couple in the window, nothing much to look at. And what he was interested in was, you know, as he walked past, will you look up also. And of course, the more people are looking up, the more people will stop and look up. And so here we have what we call positive feedback. You know, the fact that you will tend to copy the behavior of other individuals, because, you know, you want to check something out. You use that social information within a variety of contexts.
But there are other examples, you know, such as the Mona Lisa, which is an incredibly famous painting, why is it so famous? Well, it's probably not because it's the best painting in the world, it's probably because of a series of historical events that, you know, that led to its fame and then the more famous it is, the more famous it becomes and so on. And people like Duncan Watts and Matt **** at Princeton have described this as the Justin Timberlake effect--no offense, Justin, but there can be, you know, even sort of an average song can actually believed to be better than it is and that can lead to sort of this positive feedback and people are buying lots of albums. And so there is this very interesting feedback effect that tends to occur. And what they discovered, which is quite fascinating, is that, you know, when you're choosing music, the very worst songs, and the very best songs, it doesn't really matter about this feedback. You know, the very best songs, you would've chosen them anyway, the very worst songs, well, even with feedback, you know it's wrong. But for this large spectrum in the middle, people are very strongly influenced by what they believe others think about that music and will actually change their own preferences based upon what they think others think, even if that information is incorrect.
Question: Does blind conformity threaten human survival?
Iain Couzin: Well, just like in animal groups, blind conformity is almost never a good thing. You know, animals have actually tuned their behavior to have adaptive conformity in a way that they will tend to conform when it pays for them to conform. And in actual fact, having a diversity of responses turns out to be very important. So I was just reading a paper recently about innovation, about how animals can solve problems and even in small birds, like high sparrows, having a larger group allows them to have a sort of a better sweet of different types of personalities and behaviors within the group, which enables the group as a whole to sort of solve problems more readily. And of course, in human society, this is, you know, incredibly important and, you know, the types of specialization and generalization of skills are really interesting within societies. And of course, we can start building structure, you know, high **** or multi-level structure within our societies as well, so in addition to having this kind of self-organized behavior that's kind of distributed intelligence, we can also, you know, harness other aspects. And so we're not as prone to these kind of errors as we could be, but of course, there are multiple examples in history of societies really going quite heavily down a distorted path. And so, perhaps, you know, the animals who have evolved to behave within these large societies are less susceptible to such behaviors.
Question: What insights does your research offer into financial bubbles and panics?
Iain Couzin: Yeah, I think that in general, the understanding of sort of complex adaptive systems is becoming increasingly influential in terms of understanding markets and in understanding banking and in understanding how people within these organizations behave, because again, there is a social, a type of social feedback and to sort of emphasize that, you know, we haven't evolved to deal with information coming at these very high rates. And typically, you know, when you're having to make very fast decisions, those decisions tend to be less accurate, on average, and this can lead to sort of a mass sort of deciding factor in the wrong direction sometimes. And so perhaps it’s no surprise that in this sort of modern society when everything is fast, fast, fast, that sometimes you get these cascades of error and very large mistakes being made.
And so I think in general, understanding how communications work and how feedbacks work and how sort of network structure and topologies work in understanding how our own society is a complex system, can indeed shed great light on how these things function.
Recorded on December 15, 2009
Interviewed by Austin Allen