HOW WE DECIDE: mind-blowing neuroscience of decision-making

Jonah Lehrer's How We Decide is the latest in a series of popular neuroscience books (Brain Rules, Stumbling on Happiness, Mind Wide Open, The Brain that Changes Itself) to (literally) blow my mind.

Lehrer, author of the celebrated Proust Was a Neuroscientist, lays out the current state of the neuroscientific research into decision-making with a series of gripping anaecdotes followed by reviews of the literature and interviews with the researchers responsible for it.

Lehrer is interested in the historic dichotomy between "emotional" decision-making and "rational" decision-making and what modern neuroscience can tell us about these two modes of thinking. One surprising and compelling conclusion is that people who experience damage to the parts of their brain responsible for emotional reactions are unable to decide, because their rational mind dithers endlessly over the possible rational reasons for each course of action. The Platonic ideal of a rational being making decisions without recourse to the wordless gut-instinct is revealed as a helpless schmuck who can't answer questions as basic as "White or brown toast?"

But overly emotional decisions are also likely to lead us into trouble. There is clearly a sweet-spot between white-hot emotional thinking and ice-cold reason, and Lehrer is trying to find it. By the end of the book, I'm nearly convinced he has.

My copy of How We Decide has literally dozens of dogeared pages that I've marked to return to in this reviews as examples of the kind of thing that made me go Wow! and sometimes even buttonhole nearby friends to read them passages. I'll run a few down for you here:

Lehrer's description of the amazing ability of dopamine to "predict" upcoming events is gripping all the way along, but I was delighted to learn that neuroscientists call signals for missed predictions (that is, the signal released when dopamine is released in anticipation of a reward that doesn't come), emanating from the anterior cingulate cortex the "Oh shit" circuit. The ACC is closely wired to the thalamus, so activation of the "Oh shit" circuit galvanizes the conscious mind, bringing the stimulus right to the front of our attention.

These mistakes are critical to good decision-making, as they are our best tutors. Lehrer describes a famous study from Stanford psych research Carol Dweck, who administered easy tests to 10-year-olds, who did well on it. The control group was praised for "being smart." The experimental group was praised for "trying hard." With only this difference, the two groups were then administered progressively harder tests. Dweck discovered that the "smart" kids did worse: they believed their initial good result was due to some innate virtue beyond their ken or control, and feared that a failure would show that they lacked this intangible. But the "hard-trying" group had been rewarded for taking intellectual risks, and so they continued. Afterwards, the "smart" kids rated the hardest tests as their least favorite; the "tryers" rated it as their most favorite.

Dopamine is the neurochemical star of the book, and its many pathologies make for gripping reading. There's a case study of Ann Klinestiver, a sedate school-teacher who was given strong doses of Requip a dopamine agonist (it imitates dopamine's action in the brain), as treatment for worsening Parkinson's Disease. Like 13 percent of Requip patients, Ann developed a gambling compulsion for slot machines that eventually ruined her life, costing her her husband, her family, and all her assets (she finally went off Requip and opted for severely constrained movement but no gambling).

The pathology here is all about missed predictions. Dopamine helps the brain to find patterns and thus make predictions about the future. But slots are random, and so in a normal brain, slot-play follows a common pattern: first the brain is delighted by the chance to chew on such a meaty problem. It formulates hypotheses about the slots' action, and then new input (mistakes that light up the Oh shit circuit) cause it to start over. But after a short time, a normal brain gives up -- there is no pattern to see, so there's no point in playing on.

But in a brain where the dopamine levels are abnormal, surrender never happens. The brain is in a constant state of reward, because of all the "new input" (random noise) that arrives every time the lever is pulled.

Irrationality doesn't just play a role in pathological gambling; the big casino on Wall Street is also a great confounder of reason. Neuroscientist Read Montague performed an experiment in which subjects were given play money and sat down in front of stock-market simulators that had, unbeknownst to them, been programmed to simulate great crashes (Dow 1929, Nasdaq 1998, Nikkei 1986, S&P 1987). Montague found that the subjects played out exactly the same panics that real-world investors fell prey to.

Subjects set out conservatively, with small bets that rocketed upward in the pre-crash bubble. Their Oh shit circuits lit up at the thought of all the money they hadn't made (the brain overvalues loss, which is why "One day only!" sales work). Subjects progressively increased their bets, putting more and more money into the bubble (which grew and grew). And then the bubble burst and Oh shit fired again, and the same subjects refused to cut their losses and take their money out of the market, because they were fixated on how much they'd lost, and couldn't bear the thought of leaving the game while they were down.

Indeed, investors follow this trend more generally, selling stocks that do well, and holding onto stocks that do poorly (because they can't part with them while they're still "behind"). Eventually, the investor's portfolio is filled with nothing but declining bad bets.

However, this loss-aversion can be short circuited with simple gimmicks, especially credit-cards. The brain just doesn't register the same loss when you swipe your card as it does when money leaves your pocket. Carnegie Mellon neuroeconomist George Loewenstein says, "credit-cards...anaesthetize your brain against the pain of payment." MIT business professors demonstrate this by showing that students bidding for tickets to a Celtics game on average bid twice as much when the betting is done by credit-card than by cash.

The answer to this is meta-cognition: think about what you're thinking. Think about what you're feeling. Think about your circumstances and what happened the last time you were here.

But don't think too much. There are classes of problems -- ones in which there are more variables than the conscious mind can juggle -- where thinking overwhelms your brain's ability to synthesize all these variables into a good conclusion. Timothy Wilson, a U Virginia psychologist, asked two groups of female college students to choose and keep their favorite art print from a selection containing a Monet, a van Gogh, and some inspirational kitten posters. A control group was asked to rate each poster from 1 to 9 and keep their top one. The experimental group was asked to fill in questionnaires about what they liked about each poster.

The controls overwhelmingly picked the fine art. Follow-up questions established that they were still happy with their decisions weeks later.

But the experimental group -- the group that had to explain what they liked about each poster -- chose the kittens. And when they were followed up, they were disappointed with their decision.

Wilson explains that the failure arises because the good things about fine art are difficult to describe: they are intangible aesthetic elements. We like them, but most of us can't explain why. On the other hand, the virtues of a kitten-picture are easy to enumerate. When asked to explain, rationally, which one is best, kittens win every time. But it is this very superficiality that causes us to quickly tire of the kittens and wish for a Monet.

Of course, it's not just kittens. Ap Dijksterhuis at the Dutch Radbout University has shown that the same failure plagues house-buyers. When given the choice of a modest house in the city near work and amenities and a huge McMansion in the suburbs, introspection favors the McMansion. It has easy-to-enumerate virtues: we can have big dinners there, the family can come to stay, and so on. But we only have a few big dinner parties and houseguests a year, and the rest of the year we're stuck with long commutes and no night-life.

Introspection is also critical to the placebo effect. Being told that you are about to experience a pharmacological effect primes you to feel that pharmacological effect. And vice-versa: students who are administered an energy beverage after being told that it is expensive experience 30 percent higher alertness than those who are told that it is a discount alternative. Likewise, people tasting wine they are told is cheap have measurably different brain activity -- and preferences -- from subjects who are told the same wine is expensive.

All this introspection takes place in the prefrontal cortex, which has lots of other work that it has to keep on top of, so when it is distracted, our ability to make good decisions decline. In one experiment, control subjects are asked to remember two numbers and are then walked down a hall to another room where they will be asked to recall them. On the way, they pass a refreshment table with chocolate cake and fresh fruit. The experimenters measure their ability to pick the "right" snack -- that is, the one that, in the light of cold reason they would opt for.

The experimental group goes through the same test -- only they're asked to remember seven numbers, which is somewhere near the upper range of what the average person can remember.

The experimental group eats cake. The control group eats fruit. When we're distracted, we stop introspecting and listen to our emotional minds. This fact is not lost on retail psychologists who design stores to maximise this effect.

Having too much information is a plague in many fields. In an experiment with MIT business students, one group is given extremely detailed reports on companies and asked to buy and sell their stocks based on what they learn. Another group is just given the stock-prices. The latter group -- betting blind -- bets better than the "overinformed" group, who have so much information that they can't decide what is and isn't important. The same thing happens to guidance counsellors who are given detailed dossiers on students and asked to predict their academic performance -- they do worse at predicting performance than counsellors who are just given student transcripts.

By the end of the book, Lehrer is ready to draw some conclusions from all this fascinating material. What he comes up with, basically, is Cognitive Behavioral Therapy (a technique that has worked for me during a bout of depression). CBT consists, basically, of introspectively interrogating your emotional response to events, to see where and how emotion is influencing reason and vice-versa. CBT requires that you write things down (at first, anyway) so that your brain can't pull a fast one by selectively recalling your track record. It's the Goldilocks of introspection: not too much, not too little, just enough.

It's great advice, and a great book, too.

How We Decide



  1. Wow, that’s, er, quite a review you have there Cory. But this book does sound EXCELLENT, and I’ll certainly give it a look!

  2. I’m wondering how failure and experience affect introspection. Divorced parents, freed of the shackles of joyful suburban conformity, tend to purchase the smaller house in the city as they are neither here nor there and are in both need of and have reason to avoid long commutes and experience nightlife, even infrequently. I am wondering if practical stressors are in fact helpful as they get us out of our mind games.

  3. As Dr Frasier Crane would point out: Unless the book pages somehow flapped wind onto your cerebellum, the book did not “(literally) blow” your mind. It figuratively blew your mind.


  4. That last paragraph reminds me of a quote from Calvin Coolidge:

    “Nothing in the world can take the place of persistence. Talent will not; nothing is more common than unsuccessful men with talent. Genius will not; unrewarded genius is almost a proverb. Education will not; the world is full of educated derelicts. Persistence and determination alone are omnipotent. The slogan Press On! has solved and always will solve the problems of the human race.”

  5. Oh no! Not the “Proust Was a Neuroscientist” guy! Really? Do people like COry Doctorow take him seriously? That’s disheartening. I’d never heard of him before *that* book and hoped it was a one-off. Now I see that he’s entwined with all kind of people I respect! Yet *that* book couldn’t have been more flawed and specious and historically and scientifically wrong! It is even a standing in-joke between my friends and me. Sigh.

  6. Why do ya think they call it ‘dopamine,’ dope?

    How do these ideas relate to politics? I’ve long suspected that “rational” generally ain’t got a hell of a lot to do with it. I don’t recall which of Heinlein’s books – “Starship Troopers”, maybe? – suggested that one’s politics are an extension of one’s emotional makeup, for which the mind supplies handsome and admirable rationalizations. That idea has really been with me, this last several years.

    Can’t wait to read it!

  7. Helps explain why those who strive for the “Platonic ideal” of purely rational thought often come off as whimpy flip-floppers since one can often find ways to rationalize decisions any number of ways. Decision-making must have a time limit, and emotional gut response must be the tie breaker to avoid a perpetual state of analysis-paralysis. The accuracy of that gut response improves with practice… use the force!

  8. I may be misremembering (which would be fitting), but I don’t recall much (if any) neuroscience in Stumbling on Happiness. I recall it being pretty much all straight behavioral, experimental psychology.

  9. Okay, I hadn’t read the whole post when I commented that Stumbling on Happiness isn’t (by my recollection) about neuroscience. Based on the descriptions of the experiments in this post, though, it sounds like How We Decide isn’t either, at least not completely.

    Or it might be better to say that, to the extent that it’s about neuroscience, it is only trivially so, by virtue of the fact that we use our brains when we behave. Hence, observations of behavior can be used to infer neurological underpinnings of behavior.

    Going much further than that with behavioral results seems problematic to me, though. ‘Homuncularizing’ the brain (e.g., saying things like ‘the brain overvalues loss’) doesn’t explain anything, it just moves the problem down (or in) a level. People (may) overvalue loss; it seems more appropriate to me to discuss constructs like ‘overvaluation’ in terms of organisms (i.e., the units exhibiting the behavior we’re observing).

    To the extent that neuroscientific models offer something other than homuncularization, they are, or at least can be, useful. To the extent that they provide information that is redundant with behavior, I’m not so sure.

    By the way, Jerry Fodor has a nice essay on why the brain is irrelevant to cognitive science:link.

  10. not really fair to say the “*Platonic* ideal of a rational being making decisions without recourse to the wordless gut-instinct.” Plato famously believed the soul was divided into reason, will/spirit, and emotions/”gut-instinct” and that a person only operated at his best when all three were balanced.

    socrates was the rationality=knowledge=virtue guy. plato abandoned that formula, even without the help of brain scans

  11. I get that the book cover graphic is meant to represent “choices,” but my first thought was still “yum, Neapolitan!”

  12. This is a fascinating subject, to be sure, but I remember reading at least 4 reviews of “Proust Was a Neuroscientist,” and each of them was rather negative, so I don’t know whether I’d call it “celebrated.”

Comments are closed.