Brain Buys (2 page)

We all know that the brain is ill suited for number crunching. But why does a device capable of instantly recognizing faces and handling the calculations necessary to catch a fly ball on the run struggle with long division? As the components of a watch reveal much about the precision with which it can tell time, the building blocks of any computational device disclose much about the types of computations it is well suited to perform. Your brain is a web made of approximately 90 billion neurons linked by 100 trillion synapses—which in terms of elements and connections surpasses the World Wide Web, with its approximately 20 billion Web pages connected by a trillion links.
¹³
As information processing elements go, neurons are extroverts, adept at building connections and simultaneously communicating with thousands of other neurons. They are ideally suited for computational tasks that, like pattern recognition, require making sense of the whole from the relationship of the parts. We will see that, not coincidentally, much of the brain’s computational power derives from its ability to link the internal representation of bits and pieces of information that are somehow related to each other in the external world. By contrast, numerical calculations are best performed by the virtual infallibility and discrete switchlike properties of each of the millions of transistors on a computer chip. Neurons are noisy elements that make lousy switches, and no one designing a device to do arithmetic would build it from neuronlike units, but someone designing a face-recognition system might.

The inherent and irrepressible ability of the brain to build connections and make associations is well illustrated by one of my favorite illusions, the
McGurk effect
.
¹⁴
During a typical demonstration, you see a woman saying something in a video clip. As you look at her face you see her lips moving (but not touching) and hear her repeatedly say, “dada dada.” But, when you close your eyes the sound transforms into “baba baba.” Amazingly, what you hear depends on whether your eyes are open or closed. The illusion is created by splicing an audio track of the speaker saying “baba” over a visual track of her saying “gaga.” So why does this result in hearing “dada” when your eyes are open? One thing the brain is incredibly adept at doing is picking up the correlations, or associations, between different events. Unless you have been watching an extraordinary number of badly dubbed Kung Fu movies, 99 percent of the time when you have heard someone pronounce the syllable “ba,” you’ve seen his lips come together and then separate. Your brain has picked up and stored this information and uses it to decide what you are hearing. The McGurk effect arises out of conflicting auditory and visual information. Although your auditory system hears “ba,” your visual system does not see the lips touch, so your brain simply refuses to believe that someone said “ba.” It settles for something in between “ba” and “ga,” often “da.” (The position of the lips when saying “da” is intermediate between the closed lips of the “ba” and the wide-open “ga.”) Know it or not, we are all lip-readers. This is a helpful feature when we are trying to understand what people are saying in a noisy room, but a potential bug when listening to dubbed movies.

It is difficult to overstate how many of our mental faculties rely on the ability of our neurons to share information with partners near and far, and create links between the sounds, sights, concepts, and feelings we experience. It is what the brain is programmed to do. It is through auditory and visual associations that children learn that the spoken word
belly button
corresponds to that fascinating structure in the middle of their tummy. The ability to learn the strokes that make a letter, the letters that make a word, and the object that a word represents, all derive from the ability of neurons and synapses to capture and create associations.
¹⁵
But the
associative architecture
of the brain also contributes to why we confuse related concepts and why it’s harder to remember the name
Baker
than the profession
baker
.

Memory flaws are far from the only brain bug related to how the brain stores information. As we will see, our opinions and decisions are victims of arbitrary and capricious influences. For example, our judgment of how good wine tastes is unduly influenced by the purported price of the bottle.
¹⁶
The brain’s associative architecture is also closely tied to our susceptibility to advertising, which to a large extent relies on creating associations within our brain between specific products and desirable qualities such as comfort, beauty, or success.

EVOLUTION
IS
A HACKER

Neurons and synapses are impressive products of evolutionary design. But despite the complexity and sophistication of the nervous system, and the awe-inspiring diversity and beauty of the life-forms that currently inhabit planet Earth, as a “designer” the evolutionary process can be spectacularly inelegant. Over billions of years life has been painstakingly sculpted by trial and error, each success at the expense of a vastly superior number of teratological dead ends. Even the successes are riddled with imperfections: aquatic mammals that cannot breathe underwater, human babies whose heads are too big to fit through the birth canal, and a blind spot in each of our retinas. The evolutionary process does not find optimal solutions; it settles for solutions that give one individual any reproductive edge over other individuals.

Consider the problem of making sure a newly hatched goose knows who its mother is—an important piece of information, since it is a good idea to stay close to whoever will be providing the meals, warmth, and flying lessons over the next few weeks. The solution nature devised for this problem was that hatchlings
imprint
on one of the first moving objects they see during their first hours outside the egg. But imprinting can backfire. Goslings can end up following a dog, a toy goose, or the neuroethologist Konrad Lorenz, if one of these is the first object they see. A more sophisticated solution would be to provide goslings a better innate template about what mother goose looks like. Imprinting is an evolutionary hack, a solution that gets the job done and is relatively easy to implement, but may become the weak link in the overall design; it is often the case that a solution devised by evolution is not one an intelligent designer would stoop to.

An aeronautical engineer who sets about the task of developing a new airplane will start by performing theoretical analyses involving thrust, lift, and drag. Next she will build models and run experiments. And most important, as the plane is built, its components will be assembled, adjusted, and tested while the plane is safely on the ground. Evolution has no such luxury. As a species evolves it always has to be done “in flight.” Every sequential modification has to be fully functional and competitive. The neuroscientist David Linden has described the human brain as the progressive accumulation of evolutionary kludges, or quick-and-dirty fixes.
¹⁷
During brain evolution, new structures were placed on top of the older functional structures, leading to redundancy, waste of resources, unnecessary complexity, and sometimes competing solutions to the same problem. Furthermore, as new computational requirements emerged, they had to be implemented with the current hardware. There is no switching from analog to digital along the way.

Human beings, of course, are not the only animals to end up with brain bugs as a result of evolution’s kludgy design process. You may have observed a moth making its farewell flight into a lamp or the flame of a candle. Moths use the light of the unreachable moon for guidance, but the attainable light of a lamp can fatally throw off their internal navigation system.
¹⁸
Skunks, when faced with a rapidly approaching motor vehicle, have been known to hold their ground, perform a 180-degree maneuver, lift their tails, and spray the oncoming automobile. These bugs, like many human brain bugs, are a consequence of the fact that some animals are currently living in a world that evolution did not prepare them for.

Other brain bugs in the animal kingdom are more enigmatic. Perhaps on some occasion you have had the opportunity to observe a mouse running vigorously on an exercise wheel. Anybody who has had a pet mouse knows it will run in place for hours on end, and likely wondered why it devotes so much time and energy running in the wheel. A somewhat anthropocentric answer would seem to be:
well the poor guy is bored, what else is he going to do?
But a mouse’s devotion to the running on the wheel resembles more an obsession than an outlet for boredom. Decades ago it was demonstrated that when rats are given access to food one hour a day, during which they can eat as much as they want, they can go on to live relatively healthy lab rat lives. However, if a running wheel is placed in their quarters, they often die within a few days. Each day they tend to run more and more, and soon succumb to hypothermia and starvation. Although rats with an exercise wheel in their cage are much more active, they will actually consume less food during the one-hour feeding period than the rats that have no wheel to run on.
¹⁹
Clearly the running does not reflect a healthy interest in aerobic activity. Rats and mice are highly successful species. Along with humans and cockroaches, few animals have managed to survive and thrive in so many different corners of the globe. They are exquisitely adaptable and resilient animals; how can they be so foolish at to be lured to their death by a running wheel? Clearly, the running wheels tap into some neural circuitry that was never properly beta-tested given that they have no precedent in rodent evolutionary history.

The bugs in the brains of the moths and skunks may eventually be corrected as a result of the obvious fact that moths that fly into flames, and skunks that get run over, reproduce less than those that don’t. But, as a designer, the evolutionary process is handicapped as a result of its notorious slowness. Evolution’s original strategy for creating creatures who avoid eating some poisonous yellow sea slug is to let any who do so grow sick or die, and thereby produce fewer offspring. This process could take tens of thousands of generations to be implemented, and, if the sea slug ever changed colors, the process would have to start all over again. Evolution’s clever solution to its own sluggishness was learning: many animals learn to avoid poisonous prey after first nibbling on one, or, better yet, learn which foods are safe by observing what their mothers eat. Learning allows animals to adapt to their environment within an individual’s lifespan—but only to a degree. Like the moths that continue to fly into a candle flame, or skunks that insist on spraying oncoming cars, many behaviors are fairly inflexible because they are hardwired into the brain’s circuits. We will see, for instance, that humans have an innate tendency to fear those things that once represented a significant threat to our lives and well-being: predators, snakes, closed spaces, and strangers. Things that in a modern world of car accidents and heart attacks should be the least of our concerns. In effect, because of evolution’s slow pace, many animals, including human beings, are currently running what we could think of as an incredibly archaic neural operating system.

To understand what I mean by a neural operating system the analogy with digital computers is again useful—albeit potentially misleading. The tasks that a digital computer performs are a function of its hardware and software; the hardware refers to the physical components, such as chips and hard drives, and the software refers to the programs or instructions that are stored in the hardware. The operating system of a computer can be thought of as the most important piece of software: the master program that provides a minimal set of computer bodily functions, and the ability to run a virtually infinite number of additional programs. When it comes to the nervous system, the distinction between hardware and software is fuzzy at best. It is tempting to think of neurons and synapses as the hardware since they are the tangible components of the brain. But each neuron and synapse has an individual personality determined by nurture as well as nature. Neurons and synapses change as we learn, and their properties, in turn, govern who we are and how we behave—the programs the brain runs. So neurons and synapses are also the software of the brain.

A more fruitful analogy between digital computers and the brain may be to compare a computer’s hardware
and
operating system to the genetically encoded program that contains the instructions for how to build a brain. The hardware and operating system are fairly permanent entities of your computer, and were not designed to be regularly or easily altered. Similarly, the genetic blueprint that guides the development and operation of the nervous system is pretty much written in stone. This neural operating system establishes everything from the approximate size of our frontal cortex to the rules that govern how experience will shape the personalities of billions of neurons and trillions of synapses. The genetic instructions coded in our DNA are also responsible for the much less-tangible features of the human mind, such as the fact that we enjoy sex and dislike scratching our fingernails on blackboards. Our neural operating system ensures we all have the same repertoire of basic drives and emotions. Evolution had to provide a cognitive recipe that tunes these drives and emotions: to balance fear and curiosity, establish a trade-off between rational and irrational decisions, to weigh greed and altruism, and set some elusive and capricious heuristic that blends love, jealousy, friendship, and trust. What is the optimal balance between fear and curiosity? Throughout evolution curiosity has driven the desire to explore and the ability to adapt to new horizons, while fear protects animals from a harsh world replete with things that would be best left unexplored. Evolution faced the daunting task of balancing opposing drives and behaviors to cope with a myriad of future scenarios in an unpredictable and fluid world. The result was not a fixed balance, but a set of rules that allowed nurture to modulate our nature. Since we
Homo sapiens
—as opposed to our extinct Neanderthal cousins—currently rule the planet, it seems likely that evolution endowed us with an operating system that was well tuned for survival and reproductive success.