It is hard not to come to the view that the very appreciation of the right brain’s sensory phenomenal sphere (that is, everything out there on the left side of space) is housed in, and local to, the right hemisphere. It is bound to the local, physical processors that we know are active in apprehending that part of the visual world. Local processing reigns, very local processing, and it underlies all of brain organization. And much of that local processing goes on outside the realm of conscious awareness. It’s modular, pervasive, and fast.
THINKING ABOUT MODULARITY
Modularity is a product of our big brains. A general principle of brain organization is that the larger an area is, the more neurons it has. The more neurons it has, the more neurons it is connected to. There is a limit, however, to all this connectivity. If each neuron were connected to every other one, our brains would have to be twenty kilometers (twelve miles) in diameter.9 Talk about a big head. The distances that axons would have to travel across the brain would slow the processing speed down to where our body’s movements would be all out of whack and our thinking would be tediously slow and dull-witted. That fat brain would also require so much energy that we would have to eat constantly and then some. So, as the ape brain evolved and got larger and the number of neurons increased, not every neuron connected to every other neuron. This resulted in an actual fall in the percentage of connectedness.
Since the internal structure and connectivity patterns change as the proportional connectivity decreases, a high level of clustering occurs, which gives the overall system greater tolerance for the failure of individual components or connections. The local networks in the brain are made up of neurons that are more highly connected to one another than to elements in other networks. This division of circuits into numerous networks both reduces the interdependence of networks and increases their robustness. What’s more, it facilitates behavioral adaptation,10 because each network can both function and change its function without affecting the rest of the system. These local specialized networks, which can perform unique functions and can adapt or evolve to external demands, are known as modules.
Modules! Modules! Everywhere Mother Nature screams it at us. If something is useful and already there, and modularized, Mother Nature uses it and keeps on going. As pointed out by Andy Clark, the distinguished philosopher from the University of Edinburgh, Hod Lipson and colleagues at Cornell University showed that “the control of apt finger motion is not enabled solely by the nervous system but involves complex and essential contributions from the network of linked tendons.”11 In other words, since we evolved to become more dexterous, why not use the information that is already in the torque relationships established in the tendons of the hand? That way the brain only needs to supply a much simpler set of instructions in order to carry out the complex task of individuated finger movements. It’s a command more like “Grab the cup” than “Okay, thumb-press down with the force of x for y amount of time, and you, middle finger, spread to the side. . . .” This would allow for a “much wider range of directions and magnitudes of fingertip forces” than would otherwise be possible for the brain to orchestrate alone. In this case, and this is the part that Clark loves to remind us of, “part of the controller is embedded in the anatomy, contrary to current thinking that attributes the control of human anatomy exclusively to the nervous system.” It’s not just your brain calling the shots for that Chopin étude; it’s your fingers that are partly in charge.
Mother Nature doesn’t reinvent the wheel after every few revolutions. And likewise, not only does the brain modularize and push tasks and minute instructions out of its central processors, but the whole cognitive system, including the working brain, the body, and the environment, will call upon information embedded in the other to carry out a goal or action.
And yet, there is something wildly unsatisfactory about such an opinion if one holds tightly to the simple linear view of brain function: Module A goes to module B which goes to module C. If this thinking were correct, it would seem to suggest that many, many cables are crisscrossing the brain in an effort to keep all of the modules up to date, which as mentioned above, would require that twenty-kilometer brain. The simple reductionist view would also require a final box where all the toiling of the parts would finally be deposited and coordinated at a single point, and voilà, conscious experience itself is produced. That model was already under fire fifty years ago from the basic split-brain studies. Why, we wondered, is it that cutting the corpus callosum, the largest communication line in the brain, instantly results in two fairly similar conscious entities, side by side enjoying a common body? Suddenly there are two final loci generating conscious experience? How could a simple linear model suddenly produce, with one slice of the surgeon’s knife, two conscious systems, side by side? In short, the simple linear model where A produces B, which produces C was not supported by the split-brain findings. New, or at least different, concepts were needed to grasp the very phenomenon being studied.
A ubiquitous concept in science is the notion of emergence: that more complex systems arise out of relatively simple interactions. Biology arises out of chemistry, which in turn arises out of particle physics. Similarly the mind arises out of neuronal interactions and above it economic principles arise out of psychology. It is a very slippery concept. And it seems tangibly present, especially when working with neuropsychological phenomena. The mental and the physical are always running into each other. Is there an emergent something that is coordinating all of the brain’s modules?
A CASE STUDY FROM THE NEUROSURGICAL SUITE
The neurosurgeon Mark Rayport, who had practiced at the Medical College of Ohio at Toledo, made a stunning observation many years ago. During the course of craniotomies where the patient maintained consciousness, he would, unbeknownst to the patient, apply a small current of electrical stimulation to the olfactory bulb, the part of the brain that is heavily involved in managing the sense of smell. As Rayport tells the story, he would engage the patient in conversation with a positive tone, say, about the upcoming spring weekend. While chitchatting away, he would apply a pulse of electricity to the brain structure. The patient would suddenly interrupt the conversational flow and say something like, “Who brought the roses into the room?” Moments later, after Rayport had switched the conversation to negatively colored topics, he applied the same electrical impulse to the same exact place in the brain with the same exact intensity. The patient would again interrupt, but this time say, “Who brought the rotten eggs into the room?”12
Here was an example of a mental process constraining a brain process even though all of it was going on in the brain. It was as if a “top-down” mental process were informing a “bottom-up” physical biological process: the mind informing and influencing the brain. In short, while a mental state was generated by the physical brain, it too had a presence and could in turn influence the very physical state that produced it.
FLIRTING WITH EMERGENCE AND ITS IMPLICATIONS
Here is how emergence can be thought of. It occurs when a micro-level complex system organizes into a new structure, with new properties that previously did not exist, to form a new level of organization on the macro level.13 For example, the behavior and properties of atoms are described by quantum mechanics. When those microscopic atoms come together to form a macroscopic baseball, however, a new set of behaviors and properties emerge that are then governed by Newton’s Laws. Neither one predicts the other. Philip Anderson, a leading physicist at Princeton, wrote a famous article in the 1970s titled “More is different.” In it he wrote that “the reductionist hypothesis does not by any means imply a ‘constructionist’ one: The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe. In fact, the more the elementary particle physicists tell us about the nature of the fundamental laws, the less relevance they seem to have to the very real problems of the rest of science, much less to those of society.”14 That nai
led it for me.
Nonetheless, the idea of emergence had a tough time being accepted, especially by neuroscientists. Why was it so tough? Hard reductionists have difficulty accepting that there is more than one level of organization—that different layers can contribute to the causal chain in understanding why things happen the way they do. Even if they do accept that, they can’t accept the notion that the radical novelty that accompanies the emergence of a higher level cannot be predicted by lower-level events. Yet, multiple layers of organization are bread and butter to physicists, who faced these issues when quantum mechanics came on the scene. While some hard reductionists still lurk among the physicists, most believe that elements in nature are inherently unpredictable and, therefore, occur only with probabilities.
As I said, this is all very slippery and difficult to keep straight, and the arguments were energized fifty years earlier by Sperry. At a meeting at the Vatican he defiantly observed:
This is not to say that in the practice of behavioral science we have to regard the brain as just a pawn of the physical and chemical forces that play in and around it. Far from it. Recall that a molecule in many respects is the master of its inner atoms and electrons. The latter are hauled and forced about in chemical interactions by the overall configurational properties of the whole molecule. At the same time, if our given molecule is itself part of a single-celled organism like paramecium, it in turn is obliged, with all its parts and its partners, to follow along a trail of events in time and space determined largely by the extrinsic overall dynamics of Paramecium caudatum. And similarly, when it comes to brains, remember always that the simpler electric, atomic, molecular, and cellular forces and laws, though still present and operating, have all been superseded in brain dynamics by the configurational forces of higher level mechanisms. At the top, in the human brain, these include the powers of perception, cognition, memory, reason, judgment, and the like, the operational, causal effects of forces of which are equally or more potent in brain dynamics than are the outclassed inner chemical forces.15
DO YOU MEAN SUPERVENE OR SUPERSEDE?
Neuroscientists are so heavily reductionistic, as are most scientists, that Sperry’s ideas really didn’t take hold. In fact, Joe Bogen recounts in his lively autobiography how Sperry’s colleagues at Caltech wanted him to get off the topic. At the same time, however, the ideas were being widely considered in the philosophical community and causing lots of thought and reaction. People were arguing about Sperry’s use of supersede versus supervene. As the philosopher Sara Bernal points out, those of a “physicalist” stripe prefer the idea of “supervene” to “supersede.”16 Many years after Sperry’s talk, the idea of supervenience was ably explained to me by Donald Davidson, the distinguished philosopher from the University of California, Berkeley. Davidson once attended a small meeting I held at the Hotel Bel-Air with George Miller, Leon, and others. His way of capturing the idea was to say, “supervenience might be taken to mean that there cannot be two events alike in all physical respects but differing in some mental respects, or that an object cannot alter in some mental respects without altering in some physical respects.”17 Others, including the philosopher David Lewis, have given the example of a dot-matrix picture: “A dot-matrix picture has global properties—it is symmetrical, it is cluttered, and whatnot—and yet all there is to the picture is dots and non-dots at each point of the matrix. The global properties are nothing but patterns in the dots. They supervene: no two pictures could differ in their global properties without differing, somewhere, in whether there is or there isn’t a dot.”18
So, the supervenience argument goes, no global, upper-level difference without local, lower-level differences. A supervenience physicalist/materialist holds that psychological, social, and biological levels supervene on physical and chemical levels. When Sperry says “supersede” and “outclassed,” he vaguely suggests something other than supervenience—a picture where level n floats freer than level n-1. The unrepentant reductionists see a sleight of hand here and claim the determinist Sperry is suddenly talking about something else rather than neuronal cell firings.
Still, I am in the camp of Somerset Maugham, who famously observed that he had to be told stuff at least twice in order for it to stick. Someone once observed that a fanatic is someone who doesn’t change his mind and doesn’t change the subject. I am not a fanatic, but I remain unsettled on how all of those modules are organized and coordinated to give rise to unitary psychological experience. Is it enough to trot in the idea of emergence and claim victory? Getting a handle on the idea of emergence, and what it might or might not be, found me at the door of John Doyle, a mathematician at Caltech.
Doyle couldn’t be more different than me. His constant thinking can only occasionally be interrupted by a martini, if force fed. He is also a jock. In the mid 1990s, he set, lost, set, and lost the rowing world record in the 40–45 age group, won a world championship in human-powered vehicles, won two golds (rowing), a fourth place (cycling) and a sixth place (triathlon) in the 1995 World Masters Games in Brisbane, Australia. Even though he is a highly sophisticated mathematician, he speaks plain English, a prerequisite for a conversation with me. To my great surprise, when I asked him one day how come he is so clear in his expositions, he matter-of-factly said, “Oh, I used to be an actor.”
IT’S NEVER TOO LATE TO LEARN
Doyle is a professor of control and dynamical systems, a highly mathematical field full of difficult and challenging engineering problems that run from understanding turbulence to understanding the Internet. With his engineering background, Doyle thinks deeply about the architecture of systems. Any system. How are they organized to do what they do? Is there a universal architecture common to all information processing systems, such as brains, bacteria, cells, and corporate structures? Human-made stuff, of course, has a design and architecture to it. Maybe in the biologic world the forces of natural selection wound up producing entities that have a similar logic to their organization. Maybe if it is interacting parts that make for a global function, then all of these systems do have a similar architecture. At the heart of his search was a disbelief in the idea of emergence, which he finds spooky and undefined. From his engineering perspective Doyle was trying to understand the levels of explanation from the concrete perspective of actually designing and building something. When a something is actually built and carries out its function, it frequently seems as if it had emergent properties, but it doesn’t. It should be understood in terms of its interacting parts.
Borrowing from the field of computer science, Doyle asks: What can we learn from the amazing systems we humans have built to process information, and how can we apply that knowledge to the problem of how the brain performs its tricks? It is commonplace in computer science to speak of a “layered architecture” of systems that build on one another, with one layer of function serving as the platform for the next layer of function. In the computer world they think in terms of seven layers. The top layer is the application or program being used, such as Facebook, while the bottom layer is the actual hardware such as an iPhone. Each layer, while inhabiting the other layers, is remarkably independent of them. Understanding this formulation is the trick. Can an engineering viewpoint help us think about a neurobiologist’s problem? I think it can.
MIND/BRAIN NETWORKS, LAYERING, AND THE BRAIN
Layered architectures are one particular kind of modular architecture. Each layer can be thought of as a module. And, as I have said, there is lots of evidence that modular architecture is selected for in evolution and development because it allows one module to adapt through some sort of change without screwing up all the other modules. Layering, though, is a particular kind of modular architecture in that the layers (modules) are organized in a line. Layer 1 goes to Layer 2 goes to Layer 3, goes to Layer 4. It is not known if this is what the brain actually uses. Instead, it may use a hierarchical modularity consisting of many modules at each of the different scales (for example, neurons, circuits,
lobes). Layering suggests a one-directional arrow (up or down through the layers) whereas hierarchical modularity enables a complex set of interactions between modules within a single scale or between disparate scales.
WHY LAYERING IS A HELPFUL CONCEPT
If you remove the cover and peek inside a mechanical clock, you will see a bunch of interconnected wheels, gears, and springs. There it is, churning away to produce a timekeeper. It doesn’t know it’s doing that, and the parts don’t know anything about its function. Similarly in the brain, the individual neurons that churn away to produce our personal conscious experience know not what they do. In order to understand the various parts’ mechanisms behind a simple clock, it quickly becomes clear that one needs to think in terms other than “this wheel connects to that spring and then to that wheel.” The old “A connects to B which connects to C” story will get you nowhere.
Tales from Both Sides of the Brain : A Life in Neuroscience (9780062228819) Page 34
