Amazing rats

by Brian Mossop Ever since the size of our brains outgrew our closest animal relatives, we humans have declared ourselves far smarter than any other creatures in the animal kingdom.  But our big brains, and bigger egos, may underestimate the intelligence of other critters, simply because we’ve been asking the wrong questions. A study published in January in PLoS One shows that if we define intelligence not in terms of communication but in terms of problem-solving, then our animal brethren may be a lot smarter than we’ve given them credit for – starting with the rat.

Ah, but how to measure problem solving skills? A classic way is to place two players in direct competition with each other.  Do the players cooperate or stand alone?  Do they adjust strategies according to how their adversary plays?  It’s a simple and clear way to test cognitive abilities. The epitome of such challenges is The Prisoner’s Dilemma, a battle of wits that uses a tiered reward system to pit two players against each other.  The game goes like this: during each round, both players are asked if they want to cooperate or defect against their opponent.  If both players cooperate, each player gains 3 points.  If one player sells the other out, the winner gets 5 points, while the “sucker” – the player that doesn’t defect – gets zero points.  If both defect, they each get 1 point.

Clearly, the highest payout for each round is to sell out one’s opponent.  Yet if the game is played indefinitely, the key to success is for both players to continually cooperate, picking up 3 points every round.  Past experiments hinted that humans are the only species capable of figuring out how to win the game, the only species that can concoct a strategy and suss out an opponent’s. But that may not be the only word.

The PLoS One study, conducted by Duarte Viana and colleagues at the Instituto Gulbenkian de Ciência, Oeiras, Portugal, showed that rats were able to cooperate and adjust tactics depending on the strategy of their opponent, when put in a Prisoner’s Dilemma scenario. The results shattered the idea that only humans can solve the Prisoner’s Dilemma – and may bode a whole new approach to how we think about intelligence in other species.

The study adapted the Prisoner’s Dilemma for rats by giving food rewards when either both animals cooperated or one rat defected.  When both rats defected, their tails were pinched. The “sucker” rat also had his tail pinched. The experiment used two T-mazes, stacked back-to-back and separated by mesh screens so that the animals could see and smell each other.  The researchers fixed one rat’s strategy (the “stooge”) to either a tit-for-tat or pseudorandom approach, by forcing him to go into either the left or right side of one of the T-mazes on each trial.  The experimental rat could then decide whether to cooperate with the stooge rat, or go for the largest food payout by defecting.
The results showed that the rats quickly figured out their opponent’s strategy.  For example, if the experimental rat defected, the stooge playing a tit-for-tat strategy would defect on the next trial.  Rather than continually going after the high food reward, the experimental rat fell in line and quickly started cooperating again, avoiding a continuous cycle of defection.  In fact, when competing against a tit-for-tat opponent, the rats cooperated about 60% of the time.  When playing against pseudorandom opponent, where there’s no clear advantage to cooperating, the cooperation rate dropped to ~20%.

Studies conducted in other labs previously concluded that rats didn’t grasp how to succeed in the Prisoner’s Dilemma.  The authors of the PLoS study noted that when experimenters observed low cooperation rates, the animals had been food deprived.  Fully satiated rats, on the other hand, freely cooperated and easily solved the Prisoner’s Dilemma.  These results show that the primordial drive for food in a hungry animal simply clouds judgement.

It may not be entirely surprising that rats cooperated in the Prisoner’s Dilemma.  After all, animals often cooperate in nature to altruistically serve the group, whether that means hunting in packs to get more meat, or a surrogate mother animal adopting an abandoned baby to boost the pack’s numbers.  Still, there’s no direct evidence that shows rats grasp the concept of direct reciprocity.  Given that the rats in this study changed their strategy based on the game their opponent was playing, and cooperation rates were only high when the rats played against a tit-for-tat opponent, the authors showed, perhaps for the first time, that rats directly reciprocate. But an even more surprising finding was how well the rats played the game.  They plotted and schemed.  They manipulated their opponents by taking calculated strategic risks for the high payout reward.  In essence, these rodents challenged our perception of animal intelligence and proved that they, too, can master both the game, and the psychological component of competition.

Viana, D., Gordo, I., Sucena, É., & Moita, M. (2010). Cognitive and Motivational Requirements for the Emergence of Cooperation in a Rat Social Game PLoS ONE, 5 (1) DOI: 10.1371/journal.pone.0008483

Brian Mossop is a neuroscientist and a science blogger. He completed a bachelor’s degree in electrical engineering from Lafayette College in 2001, but soon traded his soldering iron for a surgical instruments when he enrolled in a graduate program in biomedical engineering at Duke University. He finished his Ph.D. in 2006, where his thesis investigated gene and drug delivery for cancer treatment. Brian’s postdoctoral work focused on the the brain, and he conducted in vivo neuroscience research at UCSF and a few Bay Area companies. Most recently, he’s been working on an early-stage startup company that he co-founded, which is creating neuroscience-based treatments for addiction.
Tags: , , , , .

Filed under Literature at Large.


  1. Pingback: Tweets that mention Amazing rats « Naturally Selected --

  2. Pingback: PLoS

  3. Bliss says:

    “These results show that the primordial drive for food in a hungry animal simply clouds judgement.”

    Cool study. This isn’t neccesarily true however. You don’t need to study game theory to know that the relative payoffs determine game solutions and equilibriums. In addition, outcomes change dramtically as contnued gaming changes marginal preferences. So the above statement may be true – or it may just be that the relative value of food to pain dramatically changes the game – which of course confirms the thesis. But I don’t know all the specifics of the experiment….Intersting nonetheless.

  4. Pingback: Ali Sohani

  5. Pingback: Blog Pick of the Month – March 2010 « everyONE – the PLoS ONE community blog

  6. Pingback: Shouting Gorilla Book Blog » Blog Archive » Prisoner’s Dilemma and Monty Hall played by rats and pigeons

  7. Pingback: Game Theory Everywhere « Polynomics

  8. Pingback: Rats can learn how to win at the Prisoner’s Dilemma

  9. relepast says:

    that the rats got it right 60% isn’t that definitive. In my experience rats are always uncooperative 🙂

  10. Pingback: Edward Harrison

  11. Pingback: Other Ten Percent » Blog Archive » Other Ten Percent 4/14/10

  12. Pingback: Sandeep Gautam

  13. Pingback: Knowing .NET » Blog Archive » Rats Are Smart Players of Prisoner’s Dilemma

  14. Pingback: fritz blog » Link dump

  15. Pingback: Mark Needham

  16. Pingback: …but they are masters of game theory. « Animals don't think…

  17. Pingback: …but they are masters of game theory. « Animals don't think…

  18. Pingback: Amber Ying

  19. Pingback: Laurie