Collective deci­sions are made with­out any cen­tral con­trol, and there are many exam­ples in nature shown.

There’s a T cell, part of the immune sys­tem; the brain; human orga­ni­za­tions like mar­kets; and of course my favorite exam­ple is ants. And on all of these sys­tems, the decision-making of the col­lec­tive is based on local interactions.

So, the slide does­n’t show ants, but it does show neu­rons that are inter­act­ing, and cells that are inter­act­ing chem­i­cal­ly and through tac­tile con­nec­tions. And in ants, the main inter­ac­tions that reg­u­late the behav­ior of the col­lec­tive or of the group are smell. So, most ants can’t see. They smell with their anten­nae, and when one ant touch­es anoth­er it can decide whether the oth­er ant, based on its smell, is a nest­mate. And the pat­tern of these inter­ac­tions, just the rate at which ants meet each oth­er, no oth­er mes­sage, is what reg­u­lates and cre­ates the behav­ior of the collective. 

So, as an ant is walk­ing around, for exam­ple in this lab are­na, and it’s touch­ing oth­er ants with its anten­nae, smelling them, it’s just adding up those con­tacts to decide what to do. And in an anal­o­gous way, a neu­ron is also adding up its stim­u­la­tion from oth­er neu­rons to decide whether to fire.

So, I’d like to tell you about one prob­lem that ant colonies solved using these net­works of sim­ple local inter­ac­tions. These are ants in the desert in the south­west US and in Mexico, and they for­age for seeds. So, they leave the nest (that’s a nest shown there), they trav­el along, they search for seeds, and they bring them back and store them inside the nest. 

And the main prob­lem that they have to solve is con­serv­ing water. It’s hot in the desert. An ant los­es water just being out­side for­ag­ing, and they get their water out of the seeds that they eat. Shown there is a seed cham­ber. So basi­cal­ly the colony has to spend water to get water, and it con­stant­ly has to man­age this trade-off. Is it worth it right now to go out and look for food?

And the way that they use inter­ac­tions to man­age this trade-off is that an out­go­ing for­ager does­n’t leave the nest until it meets enough return­ing for­agers with food. So what you see there is a film made inside the nest with a video­scope. And the return­ing forg­er comes in and then the out­go­ing for­agers add up their con­tacts with return­ing forg­ers to decide whether to leave.

A line drawing illustrating an ant's path leaving the nest, wandering around, and then returning.

And this makes sense if you think about the for­ag­ing trip of an ant that leaves the nest, trav­els along the trail, then mean­ders around to search for food, and then turns around and brings it back to the nest. So, the more food there is out there, the faster they find food, the faster they come back, the more ants go out. So it’s sim­ple pos­i­tive feedback.

Shown there are ants now in a nest where the top has been exca­vat­ed. The ants with seeds are return­ing forg­ers com­ing in with food, and inside are ants, out­go­ing for­agers, that are meet­ing the return­ing forg­ers. The ants inside are meet­ing the return­ing forg­ers and decid­ing whether to leave. No ant knows how much food there is out there, but that’s what reg­u­lates the foraging.

Now, we find the colonies are dif­fer­ent. Some colonies work like this where the ants are mov­ing very fast no mat­ter how hot it is. But oth­er colonies move much more slow­ly, for­age much less when it’s hot and dry. It’s real­ly the dif­fer­ence in poor con­di­tions that sep­a­rates them. And because I’ve been track­ing these col­leagues for a long time, I’ve been able to see the sig­nal of nat­ur­al selec­tion in this pop­u­la­tion. It turns out, sur­pris­ing­ly, that the colonies that con­serve water, the ones that sac­ri­fice get­ting as much food as pos­si­ble and restrict for­ag­ing when it’s hot and dry, are the ones that are hav­ing more offspring. 

So we like to ask what is dif­fer­ent about the indi­vid­u­als in the colonies that are more restrained in their for­ag­ing behav­ior, and it seems that the indi­vid­u­als are less sen­si­tive to inter­ac­tion. So we find out about that by trac­ing the path of all the indi­vid­u­als in a video like the one I just showed you. 

Graph showing a complex, color-coded web of lines, with dots markin where ants interacted.

Each of those dots is an inter­ac­tion, and we can see that what’s hap­pen­ing is that the indi­vid­u­als in colonies that for­age less it seems have less dopamine in their brain. So they are less like­ly to respond to inter­ac­tions, and the whole colony for­ages less.

So, these dif­fer­ences among colonies in the deci­sions of indi­vid­u­als lead to dif­fer­ences among colonies in the ecologically-important reg­u­la­tion of for­ag­ing. And so this anal­o­gy between how ants make deci­sions and how neu­rons make deci­sions is help­ing us under­stand the col­lec­tive decision-making of the colony and how it is cur­rent­ly evolving. 

So, we see also col­lec­tive decision-making in human orga­ni­za­tions and human social behav­ior. For exam­ple, the way that a soc­cer game, the stock exchange, oth­er mar­kets, even the Internet. And so the ques­tion that I’d like to leave you with is, how can we extend what we are learn­ing about how sim­ple local inter­ac­tions in ant colonies or in brains, in the aggre­gate, pro­duce the col­lec­tive behav­ior of the group and the way that it responds to chang­ing con­di­tions? How can we extend what we’re learn­ing about col­lec­tive behav­ior in oth­er sys­tems to begin think­ing about col­lec­tive behav­ior in human social organizations?

Thank you.

Further Reference

The Gordon Lab at Stanford University.