Sunday, April 01, 2001

The CEO of the Cerebral Promontory

The Executive Brain: Frontal Lobes and the Civilized Mind

By: Patricia S. Goldman-Rakic, Ph.D.

“Without the great development of the frontal lobes in the human brain (coupled with the development of the language areas), civilization could never have arisen.”

–Oliver Sacks in the forward to The Executive Brain

What thinking man or woman has not pondered the “mind-brain” problem, asking in awe how three pounds of organic matter can conceive War and Peace or enjoy “Yankee Doodle”? Like the study of the cosmos, the mind-brain question inspires wonder and passion in observers. It stirs the deepest questions about who we are and how we got here. And now, as never before, it is a problem that can be studied by incisive methods. 

In the cerebral pecking order, the frontal lobes occupy the exalted position of “house master,” “corporate executive officer,” “court of last appeal,” or—in the parlance of cognitive psychology—“central executive.” No other subject today has more currency for scientists who use imaging techniques to study human brain function; interest is even beginning to invade the sacrosanct domain of synaptic physiology. 

The frontal lobes have been an abiding subject of intellectual inquiry since antiquity, but two events in the 18th century can be seen as the beginning of modern understanding. The first was a report of a remarkable transformation in behavior resulting from accidental massive injury to the frontal lobes of railway foreman Phineas Gage: “He was no longer Gage,” said John Martyn Harlow, M.D. Then, in 1875, just four years after Harlow’s clinical report, Sir John Ferrier published an equally prescient observation of a monkey’s behavior following experimental removal of the frontal lobes in the Philosophical Transactions of the Royal Society

The experiments show that an animal deprived of its frontal lobes retains all of its powers of voluntary motion unimpaired and that it continues to see, hear, smell and taste and to perceive and localize tactile impressions as before...the result, therefore, is almost negative...and yet the facts seem to warrant the conclusion that a decided change is produced in the animal’s character and disposition. 

Frederick Tilney, a Columbia neurologist and leading comparative anatomist of the early 20th century, wrote in his preface to The Intellectual Functions of the Frontal Lobes, the 1936 classic description of a frontal lobe patient by Richard M. Brickner: 

It is possible to lose everything—sight, hearing, taste, smell, locomotion—and even the power of speech, and still to get along in some fashion. But to lose the mind is like passing out into the starless night of unending darkness...The frontal lobe is, by common consent, the most dominant part of the brain. Yet, in spite of its supreme position in the guidance of life, it is the least understood and perhaps also the least studied structure of the body. 

Tilney attributed this to “the great complexity of structure” and “the inhibiting dread of understanding its full potency.” He maintained that the topic of brain functioning and, in particular the frontal lobe, were part of a “rich crop of taboos,” an indelicate subject to bring up in polite society. 

Today we are witnessing a sea change in interest in and understanding of the importance of this “cerebral promontory.” Publication of The Executive Brain by Elkhonon Goldberg may be a kind of capstone in educating the lay public about the brain. It could also serve as an introduction to neuropsychology for the beginning neuroscientist and academics in other fields who want to learn about the biological basis of human creativity. 


Before we delve into Goldberg’s thesis and treatment of the subject, we need some background about the field from which it springs, neuropsychology. Neuropsychologists endeavor to map the functions of parts of the human brain by studying how those functions (mainly behavior) are destroyed or compromised by injury or disease in a particular part of the brain, such as the frontal lobes. This field embraces the study of higher, or mental, function (as distinguished from sensory or motor function) that takes place in the surface layer of the brain known as the neocortex. It uses methods ranging from case studies to functional magnetic resonance imaging. Neuropsychology also encompasses case reports by neurologists that describe the nature of a patient’s injury and consequent changes in behavior. From such descriptions, inferences are formed about the relationship between brain structure and function. Although they still study individual patients with rare brain injuries or abnormalities, neuropsychologists today usually compare groups of patients with common injuries or diseases to other brain-injured and normal groups. 

Goldberg himself was a student of the great Russian neuropsychologist Alexandr R. Luria, and this treatise on the frontal lobe, written in an engaging, personal, and somewhat Socratic style, is dedicated to his teacher, whom he affectionately remembers as “my professor, my mentor, my friend and my tyrant.” 

The study of higher cortical functions is also the province of experimental neuroscience, often referred to as “systems neuroscience” because it tackles the perceptual, memory, and motor command functions of the brain mechanistically. This field seeks causal explanation at many levels—anatomical, physiological, and, increasingly, cellular, molecular, and genetic—using experimental animals that in many cases display the capacity for true cognitive operations. Goldberg is less familiar with this approach.


It is not surprising that the prefrontal cortex is among the last areas of the cerebral cortex to be understood at the level of specific neural processes and cognitive operations. It is this part of the brain’s “supreme position in the guidance of life” that doubtless has inspired so many theories about its origin, function, and plasticity. Nevertheless, this critical role is still understood mainly by academicians. Parents of children with developmental disorders, afflictions of drug abuse, or mental illness may be somewhat aware of the role of the prefrontal cortex, but it is not common public knowledge. Further, in spite of the frontal lobes’ currency in neuroimaging and centrality in mind-brain relationships, some still argue that the relationship between mind and brain in terms of cognitive processes will never be understood. Others, however, claim already to have unraveled these mysteries. 

A balanced statement would be that during the past two decades there have been substantial, even remarkable, advances in understanding the psychological and neurobiological underpinnings of higher cortical functions. The “molecular revolution” has to some extent overshadowed progress in the field now called cognitive neuroscience. The present account by Goldberg, understandably, is not fully informed or entirely critical in outlining the blueprint that lies before neuroscientists attacking the big issues of intentionality and mentation on the level of circuits, cells, and molecules. His account of neural systems is rather more intuitive, but who can speak authoritatively for this complex field? Who can be expected to know all the details of methodology and the conceptual issues in frontal lobe neuroscience well enough to articulate a theory of cognition and make it stick? 


Goldberg gives us a fluid account of his personal theory of higher cortical functions as expressed in human behavior. The Executive Brain is the product of his personal experience as a clinical psychologist and a neuropsychological investigator. As summed up in an appreciative foreword by his friend and colleague Oliver Sacks, “The Executive Brain is a highly engaging and intimate memoir, a sort of intellectual autobiography, no less than a grand piece of scientific reporting and ‘popular’ science.” With much of this, I basically agree. 

The book focuses on the frontal lobes of the brain, in Goldberg’s words “the one part of the brain that defines who you are and defines your identity, that encapsulates your drives, your ambitions, your personality, your essence.” The “orchestra conductor” is his metaphor, although the reader should be aware that the executive role accorded the prefrontal cortex has been a well-entrenched concept for at least 150 years. What Goldberg achieves in this book, more than originality of concept, is a remarkable capacity to relate the subject to ordinary experience and thereby to bring it into the domain of human interest. 

Developing a theory of executive function, consciousness, and free will is a daring and daunting task, requiring intimate knowledge of methodology, experimental design, and a broad span of clinical and basic research. This is an exciting and propitious time; many scientists, including myself, believe the questions addressed by Goldberg are no longer inaccessible to neuroscience and neurobiology. I am convinced that knowledge of the cellular basis of operations in the brain and characterization of interactions between neurons can illuminate our understanding of the human psyche and of behavioral disturbances resulting from injury and disease. 

This brings me into conflict with some aspects of Goldberg’s thesis, or at least the terms in which he presents it. For example, he contrasts the classic theory that the brain consists of separate modules dedicated to particular functions with what he calls “gradient thinking.” In the latter, there is a gradual transition from one cognitive function to another, corresponding to a continuous gradual trajectory along the cortical surface, and the many parts of the brain operate in parallel. Sadly, he also treats the prefrontal cortex as a homogeneous command center, rather than a structure with separate, distinct functions, as modern neuroanatomy and neurophysiology show it to be. 

Goldberg has woven this tapestry out of his own personal experience and research by others. After introducing the metaphor of the prefrontal cortex as “orchestra conductor/corporate executive officer,” he supports it with a mistaken account of the anatomical connections that presumably position the prefrontal cortex as the “command post” of the brain. A part of the thalamus in the middle of the brain (known as the mediodorsal thalamic nucleus) is presented as the point where subcortical sensory information converges and is integrated on its way to the cerebral cortex. “True to its executive functions, the prefrontal cortex is probably the best-connected part of the brain,” writes Goldberg. In fact, however, the mediodorsal nucleus thalamus receives its input from a different part of the brain, the motor-related structures and the amygdala, not from the sensory centers, as Goldberg indicates. Moreover, the prefrontal areas of the brain are no more well connected than dozens of other cortical areas. 

Goldberg treats the prefrontal cortex as a homogenous entity rather than the compartmentalized modular structure that anatomical, physiological, and human imaging studies have repeatedly shown it to be. I do agree with Goldberg that the divisions among prefrontal areas are graded rather than sharp, but the functional and connectional differences are not merely quantitative. The prefrontal cortex is demarcated into regions based on distinct inputs and outputs, and these regions are correlated with different functional specializations. This is not to say that there are no unifying structural or functional principles that explain their coordinated activities, but simply to emphasize that the frontal lobe is heterogeneous. 

The essential difference between Goldberg’s thesis and what I take to be the state of the field is not over the prefrontal cortex’s pivotal role in executive processing; it is about the brain organization envisioned to carry out these functions. It is his neurology, not his psychology, with which I take issue. Getting the brain’s functional architecture straight is the objective of cognitive neuroscience in all its diverse specializations: physiology, anatomy, and neuropsychology. And getting it straight has the potential to demystify the human psyche and help us understand and treat its various maladies: aggression, depression, mania, memory loss, and the rest. It should be clear that I emphatically share Goldberg’s interest in, respect for, and reverence toward the remarkable social and biomedical prowess of the frontal lobe.


To return to the key issue raised above, Goldberg considers the study of separate modules in the brain’s cortex a fallacious approach that offers little information: 

In reality, the modular theory explains very little, since by lacking the ability to reduce multitudes of specific facts to simplifying general principles, it fails the basic requirement of any scientific theory. Like the belief systems of antiquity, it merely re-labels its domain by inventing a new deity for each thing. Nonetheless, like every simplistic notion, it has the seductiveness and illusory appeal of instant explainability—by introducing a new module for every new observation! 

According to Goldberg, modularity is “a high-tech revival of 18th century phrenology.” In his view, a modular system consists of autonomous units that are connected but have limited influence over each other’s complex functions; worst of all, a modular system predicates preordained functions for each unit. This denunciation is based in part on Goldberg’s dismissal of neuropsychological research describing remarkably isolated deficits in function, with related or similar functions being spared: for example a patient who was unable to name a familiar object, “peach,” but had no difficulty naming less familiar objects such as “abacus” or “sphinx.” 

This is one major theme in the book that does not have resounding strength. Rare experiments of nature have informed and inspired much of neuropsychological investigation, but unique cases are proof of principle, not final arbiters of truth. The credibility of cases that show encapsulated deficits is enhanced when it is discovered, as it was in the 1970s, that the brain’s prefrontal areas, no less than its sensory areas, are organized by their connections into anatomically defined and developmentally constructed radial units called modules, columns, or stripes. 

I also take issue with Goldberg’s position that modular units in the brain are autonomous, with little influence on one another—the implication being that a unit committed to a particular function is impervious to change. Goldberg is correct that at the finest level of modular analysis, the neuron, there is evidence of high specificity. We could call this “preordained,” if by that we mean that a neuron is controlled by its developmentally determined inputs and outputs. We have seen in our own research that neurons encode specific content and become engaged whenever an animal is required to hold in mind a specific item of information. It is not generally feasible to record such neuronal activity in the human brain, but neuroscientists have recorded single neurons with similar selective responsivity in the hippocampal region of epileptic patients being prepared for surgery. 

It is also important to note that, while patterns of brain activation are predictable and discrete, a neuron’s responsiveness can be drastically influenced by distracting or rewarding stimuli, by drugs, stress, fatigue, and disease. It has been estimated that a single neuron may receive from 6,000 to 10,000 inputs that influence its performance, and hence the performance of the circuits of which it is a part. The pattern of a single neuron’s connections establish its specificity: that to which it is and is not capable of reacting. It is through the strengthening and weakening of these hard-wired connections that a neuron’s defining response can be changed. 


How readily theorists accept, even venerate, the idea that there are specific cortical modules connected with the various senses, but still find the concept objectionable for cognitive structure. I find modularity to be a powerful model for the cognitive workings of the brain. Marcel Kinsbourne, one cognitive neuropsychologist who recognizes this problem and is a “modularist,” takes the view that: 

If the human brain were homogeneous, with all parts equally involved in all intelligent activities, then its manner of operation would be inscrutable and not very interesting... If it were instead composed of discrete components (or modules) like a machine, then in principle its composition would not be hard to determine, though how such a conglomerate would have evolved would be perplexing. In reality, the brain is more interesting than either of these blueprints suggest. It presents a paradox: It is a completed connected nerve net, and yet it is highly differentiated in its parts. 

Goldberg’s dismissal of modularity opens the way for an “a-modular” architecture he calls “gradiental.” According to this thesis, there is “an alternative blueprint”: “a massively parallel interconnected brain.” He writes: 

According to the gradiental principle, massive continuous interactions take place in the brain, while relatively little is preordained about the function of its parts. Instead it is assumed that the functional role of various cortical regions emerge according to certain basic gradients. 

Illustrating his point, Goldberg contrasts the anatomical organization of the thalamus with that of the cortex. He suggests that the thalamus is a collection of nuclei—a modular structure—with little integration among the parts; in contrast, the neocortex is presented as a more recent evolutionary acquisition with “rich pathways interconnecting most areas with most others.” It is true that there are many interconnections among cortical areas, but it is not true, as both Goldberg and Kinsbourne maintain, that most areas in it are connected with most other areas. For example, the dorsal pathway is anatomically quite segregated from the ventral “stream.” 

Goldberg forgets that individual thalamic nuclei project to targets in the cortex and imprint specificity on those targeted areas; this makes each of the cortical targets as functionally dedicated as the thalamic nuclei themselves. Similarly, boundaries among the separate nuclei of the thalamus are no less “gradiental” than are the functional subdivisions of the neocortex. 

Many neuroscientists, myself included, will agree with Goldberg that the circuitry of the neocortex is strongly, if not massively, parallel. Nevertheless, it is more accurate to describe a multitude of cortical circuits, each having specific and limited connections, and with lines of communication between them. Were most areas connected to most other areas, as Goldberg incorrectly assumes, it would be hard to implement parallelism, since parallelism implies a large degree of— if not necessarily absolute—autonomy of separate processing systems.

Goldberg also suggests that the units in a massively parallel architecture are smaller, more numerous, and less functionally complex than are modules. The units of which he speaks are not defined, but they may well be neurons, which are the smallest functional unit that can be envisioned, are numerous, and have functional properties that are reasonably tractable. In the prefrontal cortex, for example, neurons have very specific, defined responses that are constrained by their location and existing connections. Like members of ethnic groups (and Goldberg makes this analogy at length), neurons with given specifications aggregate and have stronger relationships with each other than with unlike neighbors. Yet these modules are the nodes in larger cortical networks that involve components of the limbic, sensory, and motor domains. These domain-dominated networks are the components of the parallel setup. So where is the contradiction between modularity and parallelism? 


Instead, the difficulty is in reconciling a massively parallel architecture with an orchestra conductor. Goldberg makes a valiant attempt: 

At the latest stage of cortical evolution, the frontal lobe developed to meet the “need”... The type of control provided by the frontal lobes is weak, superimposed on a high degree of autonomy of other brain structures. At the same time, the frontal lobe control is “global,” coordinating and constraining the activities of a vast array of neural structures at any given time and over time. The frontal lobes do not have the specific knowledge or expertise for all the necessary challenges facing the organism. What they have, however, is the ability to “find” the areas of the brain in possession of this knowledge and expertise for any specific challenge, and to string them together in complex configurations according to the need. 

Here Goldberg appears to be saying that massive parallelism and a control structure (the frontal lobes) coexist and, if so, the controller is a module, and a grand one at that. Goldberg’s quandary is that the two constructs that he advances— parallel architecture and the orchestra conductor—are not entirely compatible. In my view, the hard neurobiological evidence indicates that “leadership” or “control” can be incorporated into a theory of parallelism; the key is to show that there are multiple large-scale networks, each composed of limbic, sensory, motor, and cognitive nodes that are organized by informational domain. Now we can modify our analogy. Instead of one all-knowing, all-feeling, nondenominational master conductor, there are multiple dedicated instrumental sections of the orchestra, each reading from its distinct portion of the same score. That the same score is read by all sections of the orchestra guarantees that the behavioral symphony produces a more or less coordinated result. It is from multiple interactive specialized domains that knowledge, insight, creativity, and human communication arise. Not only does such a model explain the unusual isolated deficits we observe in cases of brain injuries, it represents a convergence of the two extreme models that Goldberg sees as being in opposition. Perhaps it also explains the rarity of perfection in daily human cognitive performance.


Many readers of this wonderfully accessible account may be less interested in the brain’s functional architecture and more interested in the practical, medical, and political implications that Goldberg draws from consideration of the frontal lobes. There is much to cover here, but I will confine myself to the cognitive health issues, which are powerfully presented. 

Alzheimer’s disease and other dementias have been society’s wake-up call. Here, in the most affluent of countries in the most affluent of times, human minds were succumbing to decay before human bodies, a sharp challenge to the tacit popular belief that “body is frail but soul is forever.” 

In the section “Jogging the Brain,” Goldberg takes up the theme of mental exercise for preventing mental decline with age or disease, arguing that “we will need our brain more than ever in the brainy new age.” Particularly apt for those of us who are university professors, he warns that: 

In the past, a university graduate could spend the bulk of a career complacently reaping the fruits of his or her accomplishments. Today, large amounts of knowledge must be acquired throughout life in order to stay professionally afloat....The body of essential information is growing exponentially, but the human mind has remained biologically unchanged, or changed very little. 

Goldberg reviews some of the disappointments in cognitive retraining efforts with brain-injured patients in Russia, attributing the lack of success to the limited capacity of healthy as well as brain-damaged individuals to generalize from one situation to another. He offers his own program of cognitive exercise that emphasizes “open ended effects” that can strengthen the brain in a manner akin to strengthening the body. The analogy is compelling, as we know, and Goldberg points out that education is “by far the most powerful predictor of cognitive vigor in old age.” 

I would caution the reader, however, that the value of education and an active, stimulating mental life needs little justification from biology. The issue of how experience influences brain function cuts across areas of neuroscience from systems neurobiology and brain development to cell biology and psychology. It is far too large a subject to cover in this review. 


The Executive Brain: Frontal Lobes and the Civilized Mind is crafted with passion and thoughtfulness. As its author tells us, this book is a personal odyssey begun when he was a 25-year-old graduate student under the tutelage of the remarkable Russian neuropsychologist, Alexandr Romanovich Luria, an icon in neuropsychology. Nothing has transpired in the nearly 30 years since, nor is likely to transpire in the future, to alter the cardinal role played by the frontal lobes in “the most advanced and complex functions of the brain,” named by Goldberg as leadership, motivation and vision, self-awareness, talent and success, creativity, and society. 

Although scientists will disagree with the pre-eminence given to one lobe of the brain and its accessibility to scientific analysis, or even its cognitive functions, the frontal lobe, its information-processing capacity, and its clinical manifestations are unmistakable. Goldberg has brought some of these to light from his clinical experience and broad reading. The very complexity of the frontal lobe may have captured the attention of many major figures in neuroscience. Many younger neuroscientists are dedicating their scientific lives to the study of the cognitive operations carried out by prefrontal circuitry. 

Many, like myself, will fault Goldberg’s knowledge of the new cellular and anatomical discoveries, his understanding of prefrontal function, and his sometimes uncritical acceptance of conclusions in published reports that have methodological flaws. Nonetheless, we are in Goldberg’s debt for articulating in clear and beautiful English the biomedical and societal importance of the subject we endeavor to unravel in mechanistic terms. 


From The Executive Brain: Frontal Lobes and the Civilized Mind by Elkhonon Goldberg. ©2001 by Elkhonon Goldberg. Reprinted with permission of Oxford University Press. 

Since the early 1980s, the functional organization of the brain has been the focus of intense scientific debate. Two radically different blueprints were being considered. The first blueprint is based on the concept of modularity. As we discussed earlier, a modular system consists of autonomous units, each invested with a relatively complex function and relatively insulated from the others. Separate modules provide inputs to, and receive inputs from, one another, but they exert little or no influence over each other’s inner workings. The interaction between modules is limited and gated through a relatively small number of information channels.

The alternative blueprint is a massively parallel, interconnected brain. Here, the units are smaller, invested with far simpler functions, but far more numerous. They are closely interconnected and interact continuously through multiple channels. 

The notion of modularity is a high-tech revival of eighteenth-century phrenology. Not only does it presuppose distinct borders between discrete units, but it also suggests their functional prededication. According to this view, a very specific function is rigidly preordained for each such unit. 

By contrast, the notion of a massively interconnected brain owes its ascendancy, in a somewhat circular way, to the formal neural networks, or neural nets...With “experience,” the formal neural nets acquire a rich array of properties, which were not explicitly programmed into them from the outset—the “emergent” properties. The patterns of their connection strengths changes, so that various parts of the net form the “representations” of various incoming types of information. 

Brain modeling with neural nets is among the most powerful tools of cognitive neuroscience today. The studies of the emergent properties, together with the clinical data about the effects of brain lesions and the methods of functional neuroimaging looking at regional interactions, offer a glimpse into an alternative, amodular principle of brain organization. Earlier in the book, I referred to this principle as gradiental. According to the gradiental principle, massive continuous interactions take place in the brain, while relatively little is preordained about the function of its parts. Instead, it is assumed that the functional roles of various cortical regions emerge according to certain basic gradients. 

Both the modular and the gradiental concepts have advocates and detractors. Both of them capture important properties of the brain. Modularity is best applicable to an old structure from an evolutionary standpoint, the thalamus, an assemblage of neuronal collections (nuclei). The interactive principle is best applied to a relatively recent evolutionary innovation in the brain, the neocortex. In particular, the interactive principle of organization captures the properties of the most recently evolved part of the neocortex, the so-called heteromodal association cortex, which is critical for the most advanced mental processes. Reptiles and birds are thalamic creatures with little cortical development. This was probably also true for the dinosaurs. Mammals, on the other hand, have a developed cortex, which is superimposed on, and overrides, the thalamus. 

The thalamus and the neocortex are closely interconnected. The thalamus is often viewed as the precursor of the cortex, containing in a rudimentary way most of its functions. While functionally close, the thalamus and the neocortex differ radically in neuroanatomical structure. The thalamus consists of distinct nuclei, interconnected with a limited number of pathways as the only routes of communication. By contrast, the neocortex is a sheet without distinct internal borders, with rich pathways interconnecting most areas with others. 

If the thalamus is a close prototype of the cortex, then what were the evolutionary pressures for the emergence of the neocortex?...The probable answer to our teleological question is that different principles of neural organization are optimal for different levels of complexity. Up to a point, modular organization is optimal. But once a certain level of complexity is required, the transition toward a heavily interconnected net consisting of a large number of simpler (but diverse in kind) interactive elements becomes necessary to ensure adaptive success. Throughout evolution, the emphasis has shifted from the brain invested with rigid, fixed functions (the thalamus) to the brain capable of flexible adaptation (the cortex). This was reflected in the explosive neocortical evolution in the mammals... 

The transition from the thalamic to the cortical principle of brain organization marks a drastic increase of all possible interaction patterns among brain structures, neuronal groupings and individual neurons. With this development, the ability to select the most effective pattern in a specific situation becomes particularly important. But the growing degree of freedom available to the brain in principle had to be balanced with an effective mechanism of constraining it at any given time; otherwise there would be the neural equivalent of chaos. 

At the latest stage of cortical evolution the frontal lobes developed to meet this “need.”

About Cerebrum

Bill Glovin, editor
Carolyn Asbury, Ph.D., consultant

Scientific Advisory Board
Joseph T. Coyle, M.D., Harvard Medical School
Kay Redfield Jamison, Ph.D., The Johns Hopkins University School of Medicine
Pierre J. Magistretti, M.D., Ph.D., University of Lausanne Medical School and Hospital
Helen Mayberg, M.D., Icahn School of Medicine at Mount Sinai 
Bruce S. McEwen, Ph.D., The Rockefeller University
Donald Price, M.D., The Johns Hopkins University School of Medicine
Charles Zorumski, M.D., Washington University School of Medicine

Do you have a comment or question about something you've read in CerebrumContact Cerebrum Now.