A Blog by Ed Yong

Why cooperation is hard for people with borderline personality disorder

Blogging on Peer-Reviewed Research
Social lives are delicate things. We’ve all had situations where friendships and relationships have been dented and broken, and we’re reasonably skilled at repairing the damage. This ability to keep our social ties from snapping relies on being able to read other people, and on knowing a thing or two about what’s normal in human society.

Fatalattraction.jpgFor instance, we appreciate that cheating fosters ill-will, while generosity can engender trust. So cheaters might try to win back their companions with giving gestures. These little exchanges are the glue that bind groups of people into happy and cooperative wholes. Now, a new study uses psychological games and brain scans to show what happens when they go amiss.

Brooks King-Casas at Baylor College of Medicine used a simple game to compare the social skills of healthy volunteers with those of people diagnosed with psychiatric condition called borderline personality disorder (BPD). People who suffer from BPD show erratic mood-swings and find it difficult to trust and understand the motives of others. As a result, they suffer from fraught personal relationships with friends, colleagues and partners.

So it was in the games. Each one was played by two players, an investor and a trustee, over the course of 10 rounds. The investor was given a princely sum of $20 and could split as much of it as they liked with the trustee. This investment was tripled and the trustee could then decide how much to return to the investor. Trust and cooperation is essential if both players are to benefit. The investor can make the most money by trusting the trustee with a sizeable share, on the assumption that some of it will find its way back. If the trustee violates that agreement, they are likely to get smaller investments in future rounds.


Mind games

When healthy people played both parts, King-Casas saw that their alliance remained stable throughout the game. Investors tend to start the games by offering just under 50% of their loot, and at the end, they were still as generous. But when people with BPD played the trustee role, cooperation started to break down and by the 10th round, investors were putting in a meagre 35% or so of their money.

In both cases, the odd lapse in trust threatened to sunder the players’ cooperative spirit. The difference was that the healthy trustees knew how to remedy the situation, and King-Casas discovered this by focusing on “low-cooperation” rounds where the investors offered $5 or less. Such a low offer could well be seen as a slight, and trustees could either retaliate and giving back a similarly paltry sum, or try to fix the broken relationship by offering the investor a larger profit-making return.

The healthy volunteers were almost twice as likely to use the latter strategy than those with BPD (45% of the time compared to 25%). And this “coaxing” proved to be the most successful strategy; it proved that the trustee was trustworthy, and led to increasingly large investments in later rounds. So in social interactions, people with BPD are both more likely to cause rifts in cooperation, and less likely to be able to mend them successfully. Why?

It’s not fair

To find out, King-Casas looked for activity in the brain that might reflect these behavioural tics. Specifically, he searched for areas which respond differently in healthy and BPD brains, in response to the size of the investors’ investment. He found one – the anterior insula.

Other psychological studies have suggested that this part of the brain plays a role in assessing fairness, and it has a particular propensity for reacting to injustice. In ultimatum games, where one player offers a share of a pot and the other decides whether to take it, the anterior insula is most active when offers are low and when players reject. When people watch someone else being punished, their anterior insula is most active when the parties are punished after apparent fair play, and least active when an actual cheat gets their comeuppance.

King-Casas found similar patterns in his players. In the brains of healthy trustees, activity in the anterior insula shot up in response to low investments, and diminished as the offerings grew bigger. But in the brains of BPD trustees, there was no such clear-cut pattern; their insulas seemed to be equally active in all cases. To account for other differences between the two groups, the team adjusted for differences in age, sex, IQ, education, income level or any medications they were taking, but to no effect – the results stood their ground.


However, the insula’s activity also changed when players were deciding how much to send back and this itme, both groups, with or without BPD, behaved in the same way. The stronger the activity in this area, the greater the chances of low repayments. So in volunteers with BPD, the insula is insensitive to the size of an offer but not its repayment. Clearly, it’s not the case that the entire region is faulty so what’s actually going on?

Great expectations

King- thinks suggests that people with BPD have naturally poor expectations of other people. Their insulas don’t react when low offers are made because to them, not because they failed to realise they were getting a raw deal, but because they expected to get a raw deal. In healthy people, activity in the insula spikes in response to low offers, not because they are inherently unfair but because they are unexpectedly so and violate the person’s normal expectations of social behaviour.

This sense of what’s normal in social interactions is either badly askew or missing entirely in people with BPD. King-Casas supported this theory by giving his volunteers a questionnaire to assess the outcomes they would expect of typical social situations. Their responses showed that the BPD group showed far less trust than their peers did.

This interpretation helps us to view the exchange game through new eyes, and explains why cooperation broke down between healthy investors and BPD players. The two were badly mismatched because each had a fundamental problem with understanding the other’s intentions.

The investor might offer a small sum to say, “I don’t trust you and I expect a sign of solidarity.” But the BPD trustee wouldn’t understand the investor’s ploy – to their minds, they received what they expected and they had no incentive to try and foster the investor’s good will. So they offer a small return. To the investor, that looks like a slight and a confirmation that cooperation had been abandoned, and they too have no incentive to fix the ruptured relationship.  

Reference: Science doi:10.1126/science.1156902

Images: courtesy of Science

13 thoughts on “Why cooperation is hard for people with borderline personality disorder

  1. Thanks for this. I find it interesting because I have an adult step-daughter who may or may not have BPD, but certainly has some social contextual dysfunctions.

  2. So let me see if I understand this…
    The normal person doesn’t trust and therefore doesn’t offer a fair amount, going low and expecting the other party to do something to ‘prove’ their trust. In other words the normal person is trying to build a psychological heirarchy where they are at the top driving the deal, in effect a priori demonstrating they are not playing fair.
    The abnormal person seeing this decides he’s not going to play.
    Just exactly what is the difference between the two?
    Neither trusts the other.
    Neither is willing to begin with a fair position of equality.
    Both parties don’t get what they want
    And the ‘abnormal’ person is at fault because they were smart enough, tuned in enough, not self-dellusional enough to actually give the other party the better bargaining position.
    This entire article seems specious at best.

  3. An interesting interpretation, but I don’t think it fits what’s happening.
    The important thing about the events as described are that they *follow* a situation where trust has been broken. The initial move by the investor, where they put a small amount in, is not the bit that breaks trust – it is a *reaction* to an earlier move by the trustee, who didn’t return a sizeable enough pot.
    One detail which I should have mentioned is that at the start of these games, the healthy investor almost always offered a reasonable sum (if you look at the error bars in the first diagram, you’ll see they tend to offer about 43-50% of their pot).

  4. This situation is known as the ‘Iterated Prisonders Dilema’ and is well know in game theory.
    The provably optimal strategy for both players to maximize their return fall into a category known as ‘Tit for Tat’.
    The gist is that as long as the person who came before your turn played fair, you play fair. If they defect, you defect. Note that there are several different types of Tit for Tat and if the two players have a ‘short memory’ with regard to who started the defection it can devolve into a ‘defect forever’ strategy which can help minimize your losses (which is a different goal than Tit for Tat).
    So from a game theoretic perspective the supposed abnormal persons strategy of defecting when faced with a previous defection is actually rational and ignoring the defection, which the article proposes as ‘normal’, and continuing to play is not.

  5. So you’re trying to say the abnormal person’s strategy that results in them earning less money is the more rational one?
    I don’t see how they could effectively be “minimizing losses” and still earning less money than the “healthy” group. Logic fails there.

  6. James, you’re effectively arguing that it’s most logical and practical to basically never trust the partner, even though that’s demonstrably false. The gameplay that results in the most profit for everyone is 100% return, here: investor invests 100%, then trustee returns 100%. It triples every round, thus, you start with $20 and end with $20*(3^10), thus you end up with $1,180,980. Split that in the last round and you get $590,490 apiece.
    Are you seriously arguing that hanging on like lockjaw to the piddling little $10 you got in the beginning is a better gameplan than full co-operation that nets everybody over half a million dollars? Because, you know, you’re quantifiably wrong.

  7. James, I disagree with your assessment. This isn’t the prisoners dilemma, in the prisoners dilemma both players are on equal footing, that’s not the case in this game. The investor has more power, the more they ‘defect’ the less the trustee earns regardless of whether they defect or not. If the investor were to ‘defect’ to the highest degree by investing nothing (which I assume is forbidden in this game)then the trustee loses completely, they don’t have the chance to dissent and ‘even the scores’.
    When the investor gives the trustee less money they’re not trying to develop a ‘psychological heirarchy’, they’re simply trying to earn the most money. They’ve decided that the trustee is a bad investment so they minimise their risk by investing less.
    The normal participants realised that trust had been lost so they gave back a higher portion of the money to prompt future investments. The BPD participants didn’t, so they ended up getting less.
    It’s not a psychological power play, it’s all about earning the most money.

  8. John, I’m not sure that you’re right about that. I think that each round is a separate game and the pot starts fresh each round.
    I’ve got a few questions about the game, do the trustees know how much money the investor has? Do they know how much the initial investment was and that it is automatically tripled?

  9. The neuroeconomics is fascinating, and maybe important.
    Commenters are confused about the Game Theory. There is a profound difference between a game played just once, and the same game played again and again. In the latter, iterative, case what matters is very much how each player models the other player(s), and learns from experience.
    Researchers have proved, at the Santa Fe Institute, that even games as simple as Rock-Paper-Scissors have mathematical Chaos in their trajectories when played iteratively, with learning.
    The main point of the paper cited in this blog thread is that people diagnosed with psychiatric condition called borderline personality disorder (BPD) CANNOT properly model other players in a simple iterated game, and thus speculatively are doomed to lower payoff in a real life in which such games are embedded in social and economic transactions..
    Standard group psychology, even when dumbed down to be taught in (say) Classroom Management graduate courses in colleges of education, now maintains a cycle of people who feel that life is a losing game behave in ways that lead to further failure, this reinforcing their lowered self-esteem.
    MANY of the urban public school students that I encountered, after I’d left my college and university teaching and plunged into High School and Middle School teaching to students on the cusp of dropping out or being expelled, were defeated, angry, withdrawn, revengeful, and disruptive. I think that new science suggests why this is so, although it does not suggest a cure.
    Not Exactly Rocket Science: by the way, before I was an Adjunct Professor of Astronomy (Cypress College) and Adjunct Professor of Mathematics (Woodbury University), I WAS in fact a Rocket Scientist, at Boeing, Rockwell, NASA/JPL, and other venues. Bottom line: Brain Science is HARDER than Rocket Science.

  10. Good stuff folks. Chemniste – neither the paper nor the supplementary info tell us whether the trustees knew about the tripling, but at the end of each round, both players are shown what each one has earned.

  11. Hardly any of the commentators are getting it. Too much in a hurry to have a contrary opinion, and not considering the matter carefully enough. Too self-important, I’d guess.
    It’s hard to understand this situation until you can understand exactly how the game procedure works, and how it feels to play it. It’s best to play several rounds — you’ll probably have to do it with play money. The skeptical commentators will likely want to revise their comments after they do this.
    This is very important research, with revolutionary implications. Borderline personality disorder is very common, and devastating. The bad cases are more crippling than severe mental disorders such as schizophrenia. Many cases end in suicide, and suicide attempts are commonly associated with it, not to mention substance abuse and other self-destructive behavior. The disorder is notoriously resistant to treatment.
    It seems to be more common in women. Many of these people marry, because they are superficially normal, often charming and often hungry for love and sex. They often have children, too. Then the whole family slowly goes up in flames from inexplicable conflict, temper tantrums, suicide threats, etc. My idea of hell is to have a borderline wife or mother. Many people suffer from one disorder.
    This is the first psychiatric disorder related to this brain region. There are probably others, which we don’t yet know about. This is the first psychiatric disorder attributable to defects in fairness perception. There are probably others.

  12. i have borderline personality disorder and i find it very hard to trust anybody girlfriends bosses anybody like that i dont even trust my own family but i have just woundered is that not just being very paranoid it would be good if this research is taken further to see what it suggests ive got to say i nevier agree or disagree

Leave a Reply

Your email address will not be published. Required fields are marked *