Acute Tryptophan Depletion increases offer rejection in the Ultimatum Game

Here is an interesting article showing the cross-over between neuropharmacology and decision making. Crockett et al. show that if you use acute tryptophan depletion to lower the levels of serotonin in subjects, they are more likely to reject unfair offers in the ultimatum game.

Background

The ultimatum game is an experimental economics paradigm. It works something like this. The proposer in the game gets to divide a certain pre-specified quantity of money between themselves and another player. The other player sees that division, and then gets to decide whether to accept it or reject it. If they accept it, then they and the proposer get the amount of money specified by the proposer. If they reject it, they and the proposer get nothing.

Say the proposer is given 20 dollars to divide. A perfectly fair division is 50-50 -- ten dollars to each person. As the division goes more in their favor, the other player becomes more and more likely to reject the offer in order to penalize the proposer for being stingy.

The fact that people want to penalize the proposer is an interesting finding, and it gets to the importance of the ultimatum game as an experiment in economics. "Rationally" the other player should accept any amount of money the proposer offers. They are better off regardless of whether the division of funds is better. But people don't think that way. They also have a psychological tendency to push for fairness, and they are willing to be penalized -- i.e. get no money -- in order to punish proposers who are unfair.

We know from experiments around the world that different groups of people behave differently in the ultimatum game. For example, some groups of people are particularly vigorous at punishing people who give unfair proposals. This suggests that these groups of people have particularly strong senses of "fair play."

A couple words of caution, however, in interpreting findings of the ultimatum game. First, when you are interpreting the ultimatum game, you are forced to confront our a priori assumptions of what is "rational" behavior. One of these assumptions is that any refusal to accept money given is irrational. But is this really true? For example, if the proposer knows that he or she lives in a society with very strict standards of fairness, this could encourage them to propose a fairer division. The impression on the part of the proposer could result in a greater pay-off overall for the other player. You could argue that it is actually more "rational" for the other player to guarantee to the proposer that he or she will receive nothing if the proposal is unfair. Frequent penalizers might get more in the long run.

Second, if you are looking for realistic assumptions of how trades work in the real world, the single ultimatum game is lacking. Very rarely is there a circumstance in which a dictator decides your fortunes and in which you can tell them to go to hell or not. Very rarely do you have single interactions with people. Human behavior is shaped by repeated interactions over multiple time scales. The effects of reputation -- how fair you are perceived to be -- matter as well. I am certain that there are variations on the ultimatum game that account for these issues -- perhaps repeated games with a single proposer. But you should only read as much into these results as you believe this game reflects real interactions between people.

Crockett et al.

Taking these caveats in mind, Crockett et al. took two groups of people -- one subjected to acute tryptophan depletion (ATD) and had them play the ultimatum game. (One tiny complexity is that the "proposer" in this game was actually a computer program. The subjects were shown pictures of the "proposer" and subjected to many one-shot games.)

Acute tryptophan depletion works by shunting metabolic activity away from the production of serotonin. Tryptophan is an amino acid that is a precursor of the neurotransmitter serotonin. In the ATD protocol, you have the person drink a shake filled with lots of amino acids. If tryptophan is omitted from these amino acids, metabolic activity in neurons is shunted temporarily away from serotonin production to the production of other neurotransmitters. This results in a temporary reduction in serotonin in the brain.

Serotonin has been implicated in a lot of psychiatric disorders, most notably depression. The most common drugs that we use to treat depression are SSRIs or selective serotonin reuptake inhibitors -- like Prozac. The function of these drugs is to increase the lifetime of serotonin in the synapse to compensate for what we believe is a deficit in serotonin function in the brains of depressed people. ATD was originally created as an experimental model for people with depression. People in remission for depression were treated with ATD, and it triggers temporary depression like symptoms. People without depression showed no effect. (The data is actually really mixed and complicated. The authors screened out subjects with a history of depression in this study as a possible confound to the data.)

The subjects in this study were divided into groups subjected to ATD and groups that actually had tryptophan included in their shakes. About 6 hours later -- how long it takes for ATD to take effect -- the subjects were tested on the ultimatum game. To confirm that ATD had actually lowered the subjects serotonin levels, blood tests were administered before and after the shakes were consumed to look at metabolites for serotonin. (Incidentally, the subjects were also tested about whether the ATD altered their moods, and the results showed no change.)

Here is the data (Figure 1C from the paper):

ATD is the acute tryptophan depletion group. PLA is the control. The groups on the bottom represent the percentage of the payout offered in the proposal. The y-axis is the percentage of those offers that were rejected. The groups on the far right show that for offers that were particularly unfair (20% of the payout was proposed), the subjects with ATD were more likely to reject the offer than those without. This discrepancy was not observed for more fair proposals.

First, these results are interesting because I don't remember the last time I read a study where such a direct manipulation of brain chemistry in humans caused a recognizable change in decision making. This makes this study more believable than correlative studies which just look at the association between a brain chemical or hormone and a behavior in the population.

There are several interesting interpretations of this data -- aside of course for the identification as serotonin as a modifier of our sense of fairness. My personal favorite is how this data related to what is called the somatic marker hypothesis. The somatic marker hypothesis is a theory of how we calculate the value of different decisions. The hypothesis argues that when we are comparing the desirability of two or more outcomes we try on those outcomes "emotionally" to establish how much we like them. We visualize those outcomes and calculate whether they make us feel good or bad. The relative feeling of good or bad is called the somatic marker. By comparing somatic markers, we can choose what to do.

Particularly considering the role of serotonin in regulating mood, I find it very interesting that it also is regulating an incentive calculation. I think that this lends credibility to the somatic marker hypothesis. The mood system and the reward system may be linked.

Hat-tip:Science News

Crockett, M.J., Clark, L., Tabibnia, G., Lieberman, M.D., Robbins, T.W. (2008). Serotonin Modulates Behavioral Reactions to Unfairness. Science DOI: 10.1126/science.1155577