sponsored links
TED2014

Deborah Gordon: What ants teach us about the brain, cancer and the Internet

March 20, 2014

Ecologist Deborah Gordon studies ants wherever she can find them -- in the desert, in the tropics, in her kitchen ... In this fascinating talk, she explains her obsession with insects most of us would happily swat away without a second thought. She argues that ant life provides a useful model for learning about many other topics, including disease, technology and the human brain.

Deborah Gordon - Ecologist
By studying how ant colonies work without any one leader, Deborah Gordon has identified striking similarities in how ant colonies, brains, cells and computer networks regulate themselves. Full bio

sponsored links
Double-click the English subtitles below to play the video.
I study ants
00:12
in the desert, in the tropical forest
00:14
and in my kitchen,
00:17
and in the hills around Silicon Valley where I live.
00:19
I've recently realized that ants
00:23
are using interactions differently
00:25
in different environments,
00:27
and that got me thinking
that we could learn from this
00:29
about other systems,
00:31
like brains and data networks that we engineer,
00:32
and even cancer.
00:38
So what all these systems have in common
00:41
is that there's no central control.
00:42
An ant colony consists of sterile female workers --
00:45
those are the ants you see walking around —
00:49
and then one or more reproductive females
00:51
who just lay the eggs.
00:54
They don't give any instructions.
00:56
Even though they're called queens,
00:58
they don't tell anybody what to do.
01:00
So in an ant colony, there's no one in charge,
01:02
and all systems like this without central control
01:05
are regulated using very simple interactions.
01:08
Ants interact using smell.
01:12
They smell with their antennae,
01:15
and they interact with their antennae,
01:17
so when one ant touches another with its antennae,
01:20
it can tell, for example, if the other ant
01:23
is a nestmate
01:24
and what task that other ant has been doing.
01:26
So here you see a lot of ants moving around
01:30
and interacting in a lab arena
01:33
that's connected by tubes to two other arenas.
01:35
So when one ant meets another,
01:39
it doesn't matter which ant it meets,
01:41
and they're actually not transmitting
01:43
any kind of complicated signal or message.
01:45
All that matters to the ant is the rate
01:49
at which it meets other ants.
01:51
And all of these interactions, taken together,
01:53
produce a network.
01:56
So this is the network of the ants
01:59
that you just saw moving around in the arena,
02:01
and it's this constantly shifting network
02:04
that produces the behavior of the colony,
02:07
like whether all the ants are hiding inside the nest,
02:09
or how many are going out to forage.
02:12
A brain actually works in the same way,
02:15
but what's great about ants is
02:17
that you can see the whole network as it happens.
02:18
There are more than 12,000 species of ants,
02:23
in every conceivable environment,
02:26
and they're using interactions differently
02:28
to meet different environmental challenges.
02:31
So one important environmental challenge
02:34
that every system has to deal with
02:36
is operating costs, just what it takes
02:38
to run the system.
02:40
And another environmental challenge is resources,
02:42
finding them and collecting them.
02:45
In the desert, operating costs are high
02:47
because water is scarce,
02:50
and the seed-eating ants that I study in the desert
02:52
have to spend water to get water.
02:54
So an ant outside foraging,
02:57
searching for seeds in the hot sun,
02:59
just loses water into the air.
03:01
But the colony gets its water
03:03
by metabolizing the fats out of the seeds
03:05
that they eat.
03:07
So in this environment, interactions are used
03:09
to activate foraging.
03:12
An outgoing forager doesn't go out unless
03:13
it gets enough interactions with returning foragers,
03:16
and what you see are the returning foragers
03:18
going into the tunnel, into the nest,
03:20
and meeting outgoing foragers on their way out.
03:22
This makes sense for the ant colony,
03:25
because the more food there is out there,
03:26
the more quickly the foragers find it,
03:29
the faster they come back,
03:30
and the more foragers they send out.
03:32
The system works to stay stopped,
03:34
unless something positive happens.
03:37
So interactions function to activate foragers.
03:39
And we've been studying
the evolution of this system.
03:43
First of all, there's variation.
03:46
It turns out that colonies are different.
03:47
On dry days, some colonies forage less,
03:49
so colonies are different in how
03:52
they manage this trade-off
03:54
between spending water to search for seeds
03:55
and getting water back in the form of seeds.
03:58
And we're trying to understand why
04:01
some colonies forage less than others
04:03
by thinking about ants as neurons,
04:05
using models from neuroscience.
04:08
So just as a neuron adds up its stimulation
04:10
from other neurons to decide whether to fire,
04:13
an ant adds up its stimulation from other ants
04:15
to decide whether to forage.
04:17
And what we're looking for is whether there might be
04:20
small differences among colonies
04:21
in how many interactions each ant needs
04:23
before it's willing to go out and forage,
04:26
because a colony like that would forage less.
04:28
And this raises an analogous question about brains.
04:32
We talk about the brain,
04:35
but of course every brain is slightly different,
04:37
and maybe there are some individuals
04:40
or some conditions
04:41
in which the electrical properties of neurons are such
04:42
that they require more stimulus to fire,
04:46
and that would lead to differences in brain function.
04:49
So in order to ask evolutionary questions,
04:53
we need to know about reproductive success.
04:55
This is a map of the study site
04:58
where I have been tracking this population
05:00
of harvester ant colonies for 28 years,
05:03
which is about as long as a colony lives.
05:06
Each symbol is a colony,
05:09
and the size of the symbol is
how many offspring it had,
05:11
because we were able to use genetic variation
05:14
to match up parent and offspring colonies,
05:16
that is, to figure out which colonies
05:19
were founded by a daughter queen
05:22
produced by which parent colony.
05:24
And this was amazing for me, after all these years,
05:26
to find out, for example, that colony 154,
05:28
whom I've known well for many years,
05:31
is a great-grandmother.
05:33
Here's her daughter colony,
05:35
here's her granddaughter colony,
05:37
and these are her great-granddaughter colonies.
05:39
And by doing this, I was able to learn
05:42
that offspring colonies resemble parent colonies
05:44
in their decisions about which days are so hot
05:47
that they don't forage,
05:49
and the offspring of parent colonies
05:51
live so far from each other that the ants never meet,
05:53
so the ants of the offspring colony
05:55
can't be learning this from the parent colony.
05:58
And so our next step is to look
06:00
for the genetic variation
underlying this resemblance.
06:01
So then I was able to ask, okay, who's doing better?
06:07
Over the time of the study,
06:11
and especially in the past 10 years,
06:12
there's been a very severe and deepening drought
06:14
in the Southwestern U.S.,
06:17
and it turns out that the
colonies that conserve water,
06:19
that stay in when it's really hot outside,
06:22
and thus sacrifice getting as much food as possible,
06:27
are the ones more likely to have offspring colonies.
06:29
So all this time, I thought that colony 154
06:32
was a loser, because on really dry days,
06:34
there'd be just this trickle of foraging,
06:37
while the other colonies were out
06:39
foraging, getting lots of food,
06:41
but in fact, colony 154 is a huge success.
06:43
She's a matriarch.
06:46
She's one of the rare great-grandmothers on the site.
06:47
To my knowledge, this is the first time
06:50
that we've been able to track
06:53
the ongoing evolution of collective behavior
06:55
in a natural population of animals
06:58
and find out what's actually working best.
07:00
Now, the Internet uses an algorithm
07:04
to regulate the flow of data
07:07
that's very similar to the one
07:10
that the harvester ants are using to regulate
07:12
the flow of foragers.
07:14
And guess what we call this analogy?
07:16
The anternet is coming.
07:19
(Applause)
07:21
So data doesn't leave the source computer
07:22
unless it gets a signal that there's enough bandwidth
07:26
for it to travel on.
07:29
In the early days of the Internet,
07:31
when operating costs were really high
07:33
and it was really important not to lose any data,
07:35
then the system was set up for interactions
07:38
to activate the flow of data.
07:40
It's interesting that the ants are using an algorithm
07:44
that's so similar to the one that we recently invented,
07:46
but this is only one of a handful of ant algorithms
07:50
that we know about,
07:53
and ants have had 130 million years
07:54
to evolve a lot of good ones,
07:57
and I think it's very likely
07:59
that some of the other 12,000 species
08:01
are going to have interesting algorithms
08:03
for data networks
08:06
that we haven't even thought of yet.
08:07
So what happens when operating costs are low?
08:10
Operating costs are low in the tropics,
08:13
because it's very humid, and it's easy for the ants
08:15
to be outside walking around.
08:17
But the ants are so abundant
08:20
and diverse in the tropics
08:21
that there's a lot of competition.
08:23
Whatever resource one species is using,
08:26
another species is likely to be using that
08:28
at the same time.
08:31
So in this environment, interactions are used
08:33
in the opposite way.
08:36
The system keeps going
08:38
unless something negative happens,
08:39
and one species that I study makes circuits
08:41
in the trees of foraging ants
08:43
going from the nest to a food source and back,
08:45
just round and round,
08:48
unless something negative happens,
08:49
like an interaction
08:51
with ants of another species.
08:52
So here's an example of ant security.
08:55
In the middle, there's an ant
08:58
plugging the nest entrance with its head
09:00
in response to interactions with another species.
09:02
Those are the little ones running around
09:05
with their abdomens up in the air.
09:07
But as soon as the threat is passed,
09:10
the entrance is open again,
09:12
and maybe there are situations
09:15
in computer security
09:16
where operating costs are low enough
09:18
that we could just block access temporarily
09:20
in response to an immediate threat,
09:23
and then open it again,
09:25
instead of trying to build
09:27
a permanent firewall or fortress.
09:29
So another environmental challenge
09:33
that all systems have to deal with
09:35
is resources, finding and collecting them.
09:36
And to do this, ants solve the problem
09:42
of collective search,
09:43
and this is a problem that's of great interest
09:45
right now in robotics,
09:46
because we've understood that,
09:48
rather than sending a single,
09:49
sophisticated, expensive robot out
09:51
to explore another planet
09:54
or to search a burning building,
09:56
that instead, it may be more effective
09:58
to get a group of cheaper robots
10:01
exchanging only minimal information,
10:05
and that's the way that ants do it.
10:08
So the invasive Argentine ant
10:11
makes expandable search networks.
10:13
They're good at dealing with the main problem
10:15
of collective search,
10:17
which is the trade-off between
10:18
searching very thoroughly
10:21
and covering a lot of ground.
10:22
And what they do is,
10:24
when there are many ants in a small space,
10:25
then each one can search very thoroughly
10:28
because there will be another ant nearby
10:30
searching over there,
10:32
but when there are a few ants
10:33
in a large space,
10:35
then they need to stretch out their paths
10:37
to cover more ground.
10:39
I think they use interactions to assess density,
10:41
so when they're really crowded,
10:44
they meet more often,
10:45
and they search more thoroughly.
10:46
Different ant species must use different algorithms,
10:49
because they've evolved to deal with
10:52
different resources,
10:54
and it could be really useful to know about this,
10:56
and so we recently asked ants
10:59
to solve the collective search problem
11:00
in the extreme environment
11:03
of microgravity
11:04
in the International Space Station.
11:06
When I first saw this picture, I thought,
11:08
Oh no, they've mounted the habitat vertically,
11:09
but then I realized that, of course, it doesn't matter.
11:12
So the idea here is that the ants
11:15
are working so hard to hang on
11:17
to the wall or the floor or whatever you call it
11:19
that they're less likely to interact,
11:22
and so the relationship between
11:25
how crowded they are and how often they meet
11:27
would be messed up.
11:29
We're still analyzing the data.
11:30
I don't have the results yet.
11:32
But it would be interesting to know
11:34
how other species solve this problem
11:36
in different environments on Earth,
11:38
and so we're setting up a program
11:41
to encourage kids around the world
11:42
to try this experiment with different species.
11:44
It's very simple.
11:46
It can be done with cheap materials.
11:48
And that way, we could make a global map
11:50
of ant collective search algorithms.
11:53
And I think it's pretty likely that the invasive species,
11:57
the ones that come into our buildings,
11:59
are going to be really good at this,
12:01
because they're in your kitchen
12:03
because they're really good
at finding food and water.
12:05
So the most familiar resource for ants
12:09
is a picnic,
12:12
and this is a clustered resource.
12:13
When there's one piece of fruit,
12:15
there's likely to be another piece of fruit nearby,
12:16
and the ants that specialize on clustered resources
12:19
use interactions for recruitment.
12:22
So when one ant meets another,
12:24
or when it meets a chemical deposited
12:26
on the ground by another,
12:27
then it changes direction to follow
12:29
in the direction of the interaction,
12:31
and that's how you get the trail of ants
12:32
sharing your picnic.
12:34
Now this is a place where I think we might be able
12:36
to learn something from ants about cancer.
12:37
I mean, first, it's obvious that we could do a lot
12:41
to prevent cancer
12:43
by not allowing people to spread around
12:45
or sell the toxins that promote
12:47
the evolution of cancer in our bodies,
12:49
but I don't think the ants can help us much with this
12:52
because ants never poison their own colonies.
12:54
But we might be able to learn something from ants
12:58
about treating cancer.
12:59
There are many different kinds of cancer.
13:01
Each one originates in a particular part of the body,
13:03
and then some kinds of cancer will spread
13:06
or metastasize to particular other tissues
13:09
where they must be getting
resources that they need.
13:12
So if you think from the perspective
13:15
of early metastatic cancer cells
13:17
as they're out searching around
13:18
for the resources that they need,
13:20
if those resources are clustered,
13:22
they're likely to use interactions for recruitment,
13:24
and if we can figure out how
cancer cells are recruiting,
13:27
then maybe we could set traps
13:30
to catch them before they become established.
13:33
So ants are using interactions in different ways
13:37
in a huge variety of environments,
13:40
and we could learn from this
13:43
about other systems that operate
13:45
without central control.
13:46
Using only simple interactions,
13:49
ant colonies have been performing
13:51
amazing feats for more than 130 million years.
13:52
We have a lot to learn from them.
13:56
Thank you.
13:58
(Applause)
14:01

sponsored links

Deborah Gordon - Ecologist
By studying how ant colonies work without any one leader, Deborah Gordon has identified striking similarities in how ant colonies, brains, cells and computer networks regulate themselves.

Why you should listen

Ecologist Deborah M. Gordon has learned that ant colonies can work without central control by using simple interactions like how often the insects touch antennae. Contrary to the notion that colonies are organized by efficient ants, she has instead discovered that evolution has produced “noisy” systems that tolerate accident and respond flexibly to the environment. When conditions are tough, natural selection favors colonies that conserve resources.

Her studies of ant colonies have led her and her Stanford colleagues to the discovery of the “Anternet,” which regulates foraging in ants in the same way the internet regulates data traffic. But as she said to Wired in 2013, "Insect behavior mimicking human networks ... is actually not what’s most interesting about ant networks. What’s far more interesting are the parallels in the other direction: What have the ants worked out that we humans haven’t thought of yet?" Her latest exploration: How do ants behave in space?

sponsored links

If you need translations, you can install "Google Translate" extension into your Chrome Browser.
Furthermore, you can change playback rate by installing "Video Speed Controller" extension.

Data provided by TED.

This website is owned and operated by Tokyo English Network.
The developer's blog is here.