Brain Buys (13 page)

The brain’s modularity underlies the symptoms of many neurological syndromes, including the aphasias, loss of motor control, and body neglect that can emerge after strokes. The causes of alien hand syndrome and Capgras syndrome are more mysterious, but they are probably attributable to the loss of specialized subsystems in the brain. Alien hand syndrome might be the consequence of broken communication channels between the “executive” areas of the frontal cortex responsible for deciding what to do and the motor areas responsible for actually getting the job done (that is, translating goals into actual movements of the hand).
³¹
Capgras has been suggested to be a consequence of damage to the areas that link facial recognition with emotional significance. Imagine running into someone who looks identical to a dead family member. Your reaction may be one of bewilderment, but it is unlikely that you will embrace him and have a positive emotional reaction toward this person. You recognize the face but the emotional impact of that face is not uploaded. In Capgras patients, the recognition of a parent’s face, in the absence of any feelings of love or familiarity, might reasonably lead a patient to conclude that the individual is an impostor.
³²

So the modules of the brain do not correspond to tidy well-defined traits like intelligence, spirituality, courage, or creativity. Most personality traits and decisions are complex multidimensional phenomena that require the integrative effort of many different areas, each of which may play an important but elusive role. We should not think of the brain’s modularity as resembling the unique and nontransferable specializations, like the parts of a car, but more like the members of a soccer team; each player’s performance depends to a large extent on the other players’, and if one team member is lost, the others can take over with varying degrees of effectiveness.

The brain’s remarkable ability to learn, adapt, and reorganize has a flipside: in response to trauma, neural plasticity can be responsible for disorders including phantom limbs and tinnitus.
³³
It is not particularly surprising that brain bugs surface in response to trauma, because our neural operating system was probably never tested or “debugged” under these conditions. Cortical plasticity evolved primarily as a powerful mechanism that allowed the brain to adapt to, and shape, the world around it, not as a mechanism to cope with trauma or injury. In a red-in-tooth-and-claw world, any serious injury pretty much guaranteed that an individual would no longer be playing in the gene pool. Thus, relatively little selective pressure would have ever been placed on removing the glitches that arose from the interaction between brain plasticity and serious trauma to the body or brain.

The cockpit of an airplane has indicators and gauges about flap and landing gear positions, engine temperature, fuel level, structural integrity, and so on. Thanks to these sensors the main cockpit computer “knows” the position of the landing gear, but it does not
feel
the landing gear. The human body has sensors distributed throughout, which provide information to the brain regarding limb position, external temperature, fuel levels, structural integrity, and the like. What is exceptional about the brain as a computational device is that evolution has not only ensured that the brain has access to the information from our peripheral devices, but that it endowed us with conscious awareness of these devices. As you lay awake in the dark your brain does not simply verbally report the position of your left arm, it goes all out and generates a sense of ownership by projecting the feeling of your arm into the extracranial world. A glitch in this sophisticated charade is that under some circumstances—as a result of the brain’s own plasticity mechanisms gone awry—the brain can end up projecting the sensation of an arm into points in space where an arm no longer resides. This may simply be the price to be paid for body awareness—one of the most useful and extraordinary illusions the brain bestows upon us.

I decided that blackjack would be the ideal gambling endeavor on my first trip to Las Vegas. Surely, even I could grasp the basics of a game that consists of being dealt one card at a time in the hopes that they will add up to 21. After I received my first two cards, my job was to decide whether I should “stick” (take no more cards) or “hit” (request another) and risk “busting” (exceeding 21). My opponent was the dealer, and I was assured that her strategy was written in stone: the dealer would continue to take cards until the sum was 17 or more, in which case she would stick. In other words, the dealer played like a robot obeying a simple program—no free will required. To avoid having to actually memorize the optimal strategies, I decided to also play as a robot and use the same set of rules as the dealer. Naively it seemed to me that if I adopted the same strategy I should have a 50 percent chance of winning.

This of course was not the case. As everybody knows, the house always has the advantage, but where was it? Fortunately, Las Vegas is a city where people are eager to pass on gambling advice, so I asked around. The taxi driver assured me the dealers had the advantage because they got to see your cards, but you did not get to see theirs. An off-duty dealer informed me that it was because I had to decide whether to take a card before the dealer. But the strategy of sticking at 17 does not require looking at any cards other than your own, so who sees whose cards or who sees them first is irrelevant. Further inquiries led to a number of fascinating, albeit incorrect, answers.

When I asked for a third card and it added up to more than 21 the dealer immediately made it abundantly clear that the hand was over for me by picking up my cards and chips, and proceeding to play with the other players at the table. When no one else wanted another card, the dealer revealed her cards and their sum, at which point I realized that she busted. Since I also busted, I actually tied with the dealer. If we had both ended up with a total of 18, it would indeed be a tie, and I would get my chips back. The casino’s advantage is simply that the patron loses a tie when both have a hand that adds up to more than 21.
¹
But why couldn’t I, or the others I spoke to, readily see this?

The reason is that the casino’s advantage was carefully hidden in a place (or, rather, a time) we did not think to look: in the future. Note that the dealer took my cards away immediately after I busted. At this point my brain said
game over.
Indeed, I could have left the table at this point without ever bothering to find out if the dealer had also busted. One of the golden rules etched into our brains is that cause comes before effect. So my brain didn’t bother looking for the cause of my loss (the house’s advantage) in the events that happened after I had stopped playing. By cleverly tapping into a mental blind spot about cause and effect, casinos play down their advantage and perpetuate the illusion of fairness.

DELAY BLINDNESS

One does not need to learn that cause precedes effect; it is hardwired into the brain. If a rat fortuitously presses a lever and some food falls from the heavens, it naturally comes to repeat the movements it made before the miraculous event occurred, not those that came after. Two of the most basic and ubiquitous forms of learning,
classical
and
operant conditioning
, allow animals to capture the gist of cause and effect. The Russian physiologist Ivan Pavlov was the first to carefully study classical conditioning in his famous experiments. He demonstrated that dogs begin to salivate in response to a bell (the
conditioned stimulus
) if, in the past, the bell’s ringing consistently precedes the presentation of meat powder (the
unconditioned stimulus
). From the perspective of the dog, classical conditioning can be considered a quick-and-dirty cause-and-effect detector—although in practice whether the bell actually
causes
the appearance of the meat is irrelevant as far as the dog is concerned, what matters is that it predicts snack time.

Dogs are far from the only animals that learn to salivate in response to a conditioned stimulus. I learned this the hard way. Every day for a few weeks, my officemate at the University of California in San Francisco shared one of her chewable vitamin C tablets with me (which are well suited to induce salivation due to their sourness)—out of kindness, or in the name of science? The bottle made a distinctive rattling sound every time she retrieved it from her drawer. After a few weeks, I noticed that sometimes, out of the blue, I found my mouth overflowing with saliva. Before presenting this newly found condition to a doctor, I realized that my officemate was sometimes getting a dose for herself and not giving me one. Totally unconsciously my brain processed the conditioned stimulus (the rattling) and produced the
unconditioned response
(salivation.

On the flipside, when Pavlov and subsequent investigators provided presentations of the sound of the bell shortly
after
giving the dogs the meat, the dogs did not salivate in response to the bell.
²
Why should they? If anything, in that case, the meat was “causing” the bell; there is little reason to salivate in response to a bell, particularly if you are in the process of wolfing down a meal.

As in most cases of classical conditioning, the interval between the conditioned and unconditioned stimulus is short—a few seconds or less. The neural circuits responsible for classical conditioning not only make “assumptions” about the order of the stimuli but also about the appropriate delay between them. In nature, when one event causes (or is correlated with) another, the time between them is generally short, so evolution has programmed the nervous system in a manner that classical conditioning requires close temporal proximity between the conditioned and unconditioned stimuli. If Pavlov had rung the bell one hour before presenting the meat, there is no chance the dog would ever have learned to associate the bell with the meat, even though the ability of the bell to predict the occurrence of the meat would have been exactly the same.

The importance of the delay between stimuli has been carefully studied using another example of classical conditioning, called
eyeblink conditioning
.
³
In humans, this form of associative learning typically involves watching a silent movie while wearing some specially adapted “glasses” that can blow a puff of air into the eye, reflexively causing people to blink (this method is a significant improvement over the old one in which blinking was elicited by slapping volunteers in the face with a wooden paddle). If an auditory tone is presented before each air puff, people unconsciously start blinking to the tone before the onset of the air puff.
⁴
If the onset of the tone precedes the air puff by a half second, robust learning takes place; however, if the delay between the “cause” and “effect” is more than a few seconds, little or no classical conditioning occurs. The maximal intervals between the conditioned and unconditioned stimuli that can still result in learning are highly dependent on the animal and the stimuli involved, but if the delays are long enough, learning never occurs.

The difficulty that animals have in detecting the relationship between events that are separated by longer periods of time is also evident in
operant conditioning
, in which animals learn to perform an action to receive a reward. In a typical operant conditioning experiment rats learn to press a bar to receive a pellet of food. Again the delay between the action (cause) and the reward (effect) is critical. If the food is delivered immediately after a rat presses the lever, the rat readily learns; however, if the delay is 5 minutes, the rat does not learn the cause and effect relationship.
⁵
In both cases the facts remain the same; pressing the lever results in the delivery of food. But because of the delay, animals cannot figure out the relationship.

This “delay blindness” is not limited to simple forms of associative learning, such as classical and operant conditioning. It is a general property of the nervous system that applies to many forms of learning. If a light goes on and off every time we press a button, we have no trouble establishing the causal relationship between our action and the effect. If, however, the delay is a mere five seconds—perhaps it is a slow fluorescent light—the relationship is a bit harder to detect, particularly if in our impatience we press the button multiple times.

In a hotel in Italy I found myself wondering what a little cord in the shower was for. After pulling on it a couple of times produced no observable effects I assumed it no longer had a function or was broken. Thirty seconds later, the phone rang; only then did I realize the mysterious cord was an emergency call in case you fell in the shower. But if the delay between pulling the cord and receiving a call had been five minutes there is little doubt I would not even have remembered tinkering with the cord, much less figured out the relationship between pulling the cord and the phone ringing.

The delays between cause and effect that the brain picks up on are not absolute, but tuned to the nature of the problem at hand. We expect short intervals between seeing something fall and hearing it crash, and longer intervals between taking aspirin and our headache improving. But across the board it is vastly more difficult to detect relationships between events separated by hours, days, or years. If I take a drug for the first time, and 15 minutes later I have a seizure, I’ll have no problem suspecting that the drug was its cause. If, on the other hand, the seizure occurs one month later, there is a much lower chance I’ll establish the connection. Consider that the delay between smoking cigarettes and lung cancer can be decades. If cigarettes caused lung cancer within one week of one’s first cigarette, the tobacco industry would never have managed to develop into a mammoth multi-trillion-dollar global business.

Why is it so much harder to detect the relationship between events separated by days or months? Of course, as a general rule the more time between two events the more complicated and less direct the nature of the relationship. Additionally, however, our neural hardware was simply not designed to capture the relationship between two events if there is a long delay between them. The primordial forms of associative learning (classical and operant conditioning) are generally useless over time scales of hours,
⁶
much less days, months, or years. Learning the relationship between planting seeds and growing a supply of corn, or between having sex and becoming pregnant, are things that require connecting dots that are many months apart. These forms of learning require cognitive abilities that far exceed those of all animals, except humans. But even for us, understanding the relationship between events separated in time is a challenge. Consequently we often fail to properly balance the short and long-term consequences of our actions.

DISCOUNTING TIME

If you were given two options—$100 now, or $120 one month from now—which would you choose?

An additional $20 for waiting a month is a very good yield; thus, the average economist would argue that the rational decision is to take $120 one month from now. Yet most people would choose the immediate $100.
⁷
This bias toward immediate gratification is termed
temporal discounting
: the perceived value of a potential reward decreases with time. As a consequence, decisions that require comparing immediate and future scenarios are, indeed, often irrational. As artificial as the above example may be, we are continuously making real-life decisions that require balancing short- and long-term trade-offs. Should I purchase the new TV today and pay interest over the next six months, or wait until I have the cash on hand? Should I buy the cheaper gas-fueled car or the more expensive hybrid, which, in the long run, will be better for the environment and allow me to save money on fuel?

For our ancestors, life was a shorter and a much less predictable journey. The immediate challenges of obtaining food and survival took precedence over thoughts about what was to come in the months or years ahead. If you have any reason to believe you may not be alive in a month, or that the person making the above offer is not trustworthy, the rational decision is to take the quick cash. Similarly, if you are broke and your child is hungry today, it would be stupid to wait a month for the extra $20. My willingness to accept a larger reward in the future is contingent not only on my belief that I will still be alive, but that whoever is making the offer can somehow guarantee that I will actually receive the greater sum at the future date. These are conditions that were unlikely to have been met during most of human evolution.

Since our neural hardware is largely inherited from our mammalian ancestors, it is worth asking how other animals behave when faced with options that span different points in time. The answer is not very wisely. Marmoset and tamarin monkeys have been trained to make a choice between receiving two food pellets immediately or six food pellets at some future point in time. How long are the monkeys willing to wait for the four extra pellets? The monkeys make an impulsive five-year-old trying to wait a few minutes to get an extra marshmallow look like a Buddhist monk. The tamarins waited only eight seconds on average. In other words, if the delay was 10 seconds, they usually went for the quick but small snack, but if the delay was 6 seconds they would generally bear the wait. The marmoset monkeys were a bit more patient, waiting on average 14 seconds for the extra food.
⁸
The mere notion of studying short- and long-term decisions in animals may not even make much sense. There is little evidence that other species can conceptualize time or think about the future. Of course, some animals store food for future consumption, but these behaviors seem to be innate, inflexible, and devoid of understanding. In the words of the psychologist Daniel Gilbert, “The squirrel that stashes a nut in my yard ‘knows’ about the future in approximately the same way that a falling rock ‘knows’ about the law of gravity.”
⁹