Before anyone can understand why academic research has spent half a century pathologizing ADHD, they must first understand the structural trap that governs modern scientific inquiry.
The modern research university and its federal funding conduits do not exist simply to solve complex systemic problems. They are institutions organized around disciplines, departments, laboratories, funding categories, peer-review committees, publication incentives, and careers. Those structures are extraordinarily effective at producing specialized knowledge. They are considerably less effective when a problem requires someone to step outside those organizational lanes and ask how all the pieces fit together.
When an independent mind arrives with a framework that cuts across established silos to explain how a complex system actually works, the institution does not necessarily mobilize to test it. Sometimes the institution simply has nowhere to put it.
No story illustrates that collision more vividly than the arc of Jeff Hawkins.
A Theory-Poor Science
In 1979, Francis Crick published “Thinking About the Brain” in Scientific American. Neuroscience was accumulating an extraordinary amount of information about the brain: cellular anatomy, electrical activity, neurotransmitters, sensory pathways, and increasingly sophisticated observations about localized neural function. Yet the larger theoretical problem remained. How did this enormous biological machine actually produce intelligence?
That question captured Hawkins. He was not interested merely in adding another observation to the growing pile. He wanted neuroscience to develop something closer to what the harder sciences had achieved: a mathematical and computational framework powerful enough to organize observations, explain mechanisms, and generate predictions that could subsequently be tested.
Physics provided the model. Relativity did not emerge because physicists accumulated enough observations and eventually stacked them into a theory. The theoretical framework generated predictions about what investigators should subsequently be able to observe. If the theory was right, certain things should happen, and scientists could go looking for them.
String theory represented the same aspiration at a different stage of development: construct a mathematical framework capable of reconciling observations that otherwise appear disconnected and determine what must follow if that framework is correct.
Hawkins wanted neuroscience to possess that kind of theoretical architecture.
More than two decades later, On Intelligence would make that ambition explicit. Hawkins and Sandra Blakeslee did not simply describe the memory-prediction framework and leave it there. The book concludes with an appendix devoted to testable predictions. If the proposed architecture of the neocortex was correct, there should be things researchers could eventually find in the biology that had not been required to construct the theory in the first place.
Theory had to explain mechanism, mechanism had to generate predictions, and those predictions had to encounter experiment. Whatever survived could strengthen the theory; whatever failed had to force revision. Hawkins wanted neuroscience to work more like that.
First, however, he had to find somewhere that would let him try.
The First Wall: MIT
In 1981, Hawkins approached MIT’s Artificial Intelligence Laboratory believing that understanding biological intelligence should be central to understanding intelligence itself. He encountered a fundamentally different intellectual culture.
Classical symbolic AI dominated the field. Intelligence was approached primarily as computation through formal symbols, rules, and logical operations. Biological brains—with their slow neurons, enormous parallelism, messy chemistry, and apparent redundancy—did not look like an elegant blueprint for intelligent machines.
Hawkins wanted to move in the opposite direction. Instead of beginning with computers and asking how they might be made intelligent, he wanted to begin with the only machine already known to produce intelligence and ask how it worked.
MIT was the first wall, so Hawkins went back into technology.
GRiD, Berkeley, and the Second Wall
Hawkins had something enormously important available to him when academia failed to provide a path forward: he was a very good engineer. After beginning his career at Intel, he joined GRiD Systems, one of the pioneering companies in portable computing.
But neuroscience never disappeared. In 1986, Hawkins stepped away from a promising technology career and enrolled in the doctoral program in biophysics at the University of California, Berkeley. He was trying again.
He completed his graduate coursework and passed his qualifying examinations, but what Hawkins wanted to pursue for his dissertation was not another narrowly bounded experimental problem. He wanted to develop a broad theoretical account of neocortical function, and once again he ran directly into the structure of the institution.
Graduate students did not simply announce that they were going to spend the next several years constructing a grand theory of the neocortex. Doctoral research existed inside laboratories, under principal investigators, within established research programs supported by existing funding structures. Hawkins could work on the questions the machinery knew how to accommodate.
The question he wanted to ask did not fit.
He walked away from Berkeley as ABD—All But Dissertation. That could have ended the story. Instead, it created the route by which Hawkins would eventually become independent enough to ask the question anyway.
Back to GRiD
Hawkins returned to GRiD and, during this second period with the company, began working seriously on pen-based computing. The result was the GRiDPad, one of the earliest commercially significant tablet computers.
That experience changed the way Hawkins thought about personal computing. Computers did not have to remain machines sitting on desks; they could become things people carried. But the tablet still wasn't the endpoint. If you could take the computer out of the office and put it into someone's hands, the next logical step was to make it small enough to put into someone's pocket.
That path eventually led him to found Palm Computing in 1992 and ultimately to the PalmPilot.
It would be easy to tell the next part of Hawkins's life as a conventional Silicon Valley story: GRiDPad, Palm, PalmPilot, Handspring, Visor, Treo. One successful technological leap followed another. But that version misses the story that matters here, because Hawkins never stopped wanting to understand the brain.
The Neuroscience Never Stopped
From the outside, Hawkins's biography can look like a twenty-year interruption in which a would-be neuroscientist left Berkeley, became a technology entrepreneur, made his fortune, and eventually returned to his original passion. That isn't what happened. Neuroscience continued alongside the technology career, and so did his family.
Hawkins had made a decision about the kind of life he wanted to live. He wanted to be home in the morning to have breakfast with his children and see them off, and he wanted to be home again for dinner. He did not accept the Silicon Valley mythology that building something important necessarily required surrendering every waking hour to the company.
Donna Dubinsky, his longtime business partner, understood the arrangement. As Hawkins has recounted it, her response was essentially that he had chosen his path and other people had chosen theirs. There was nothing illegitimate about choosing a life that did not revolve around working sixty or eighty hours every week.
Hawkins made a related point in his Stanford talks. He rejected the assumption that the number of hours spent working was a useful proxy for the quality of the work. His engineering instinct was almost the opposite: make good decisions, choose the right problems, and build things intelligently. If those decisions are sound, much of the frantic labor people assume is inevitable becomes unnecessary.
This mattered enormously after Palm cleared its major early product hurdles. Hawkins was Palm's chief technology officer, not its CEO. Once the company reached a stage where he no longer needed to devote every available hour to getting the product out the door, his arrangement allowed him to devote roughly twenty hours a week to neuroscience.
While helping build one of the defining personal-computing companies of the era, Hawkins was spending approximately half of a conventional workweek thinking about the neocortex. At the same time, he was raising his children and maintaining the boundaries he had established around family life. The neuroscience was not something he rediscovered after becoming wealthy. It was already there.
The Work That Kept Interrupting the Work
That balance was never stable. There were moments when the technology business demanded Hawkins's full attention again. The corporate changes surrounding Palm and the creation of Handspring pulled him back into company building. The Visor required another enormous product-development effort, and the work that ultimately produced the Treo pushed handheld computing further toward the convergence of organizer, communications device, and mobile computer.
At those moments, neuroscience had to move aside, but those were interruptions rather than abandonment. Technology was simultaneously the thing that repeatedly took Hawkins away from neuroscience and the thing that eventually gave him the freedom to pursue neuroscience without asking anyone's permission.
The moneymaking part of his life sometimes required him to put the brain aside. It was also financing his escape from the institutional structures that had prevented him from studying the brain the way he wanted. Eventually, those two trajectories converged.
The Twenty-Year Plan
By the time Hawkins emerged from Palm and Handspring, his circumstances were radically different from those of the young engineer who had approached MIT. He had credibility, resources, successful companies behind him, and technologies used by millions of people.
Neuroscience itself had moved as well. The growing understanding of neuroplasticity reinforced a picture of the brain not as fixed circuitry but as an adaptive learning system whose representations and connections changed through experience. The scientific landscape was becoming more compatible with the kind of problem Hawkins had wanted to pursue for decades.
But the central question had barely changed: What computational principles allowed the neocortex to learn a model of the world?
Now Hawkins possessed something he had not possessed at MIT or Berkeley. He no longer needed an academic department to give him permission to ask. In 2002, he founded the Redwood Neuroscience Institute.
The entrepreneur had finally built the research environment the graduate student had been unable to find.
Building the Place That Didn't Exist
Redwood was designed around a problem rather than a conventional academic lane. Hawkins wanted people thinking about the thalamocortical system as an information-processing architecture. He wanted theoretical neuroscientists, experimental neuroscientists, mathematicians, and computational thinkers talking to one another about the same machine.
The objective was not simply to collect another mountain of observations. It was to explain the mountain already sitting there.
This distinction runs through Hawkins's entire scientific life. He was not hostile to empirical science. Quite the opposite: a theory that cannot survive contact with empirical evidence is useless. But collecting evidence and constructing an explanatory architecture from that evidence are not the same scientific activity. Someone has to measure the pieces, but someone also has to ask how the pieces could possibly belong to the same machine.
In 2004, Hawkins and Sandra Blakeslee published On Intelligence. The book proposed the memory-prediction framework: the idea that intelligence emerges from a hierarchical cortical system that learns patterns across time and continuously generates predictions about incoming sensory information.
Hawkins then did exactly what he believed a theoretical scientist should do: he made predictions. The appendix to On Intelligence laid out observations that should eventually be found if the proposed framework was correct. The theory was not supposed to become another vocabulary for describing intelligence after the fact. It was supposed to tell scientists where to look next.
Two Paths From Redwood
By 2005, Hawkins had reached another transition point. Redwood Neuroscience Institute did not disappear, nor did it simply become Numenta. Instead, the work moved forward along two paths.
Hawkins gifted the Redwood Neuroscience Institute to the University of California, Berkeley, where it became the Redwood Center for Theoretical Neuroscience. Its theoretical-neuroscience mission continued there: developing mathematical and computational models of brain function and connecting those models to experimental neuroscience.
At essentially the same transition point, Hawkins, Donna Dubinsky, and Dileep George founded Numenta as a separate organization.
The distinction matters. Redwood had demonstrated that theoretical neuroscience could be organized around the larger question Hawkins had been pursuing for decades. Moving it into Berkeley gave that work an academic home, while Numenta provided another path: an independent research organization through which Hawkins and his collaborators could continue developing theories of neocortical computation.
There is an extraordinary symmetry to that history. Berkeley—the institution where Hawkins had been unable to pursue the theoretical neuroscience dissertation he wanted in the 1980s—eventually became the permanent home of the theoretical neuroscience institute he had to become independently wealthy to create.
Numenta gave Hawkins the independent research structure to continue pushing the computational theory forward. Hierarchical Temporal Memory led into sequence memory and eventually toward cortical columns, grid-cell-like mechanisms, and reference frames.
The theory changed, and that matters. Hawkins was not protecting a doctrine. He was trying to construct an explanatory framework capable of surviving evidence. When pieces did not work, the framework had to change. When new biological evidence appeared, the architecture had to accommodate it or eventually be discarded.
By the time Hawkins published A Thousand Brains in 2021, the theory had evolved substantially from the memory-prediction framework described in On Intelligence. But the question had remained almost absurdly consistent across forty years:
How does the neocortex actually work?
The Third Wall: NIH
The full scope of the structural problem became obvious when Hawkins encountered the federal research apparatus itself.
By the early 2000s, he possessed enough commercial success and stature to meet directly with senior leadership at the National Institutes of Health. He argued that neuroscience needed large-scale, multidisciplinary theoretical models capable of making sense of the enormous amount of empirical information the government was already paying researchers to collect.
The response from the institutional gatekeepers was extraordinary in its bureaucratic candor: the work was important and needed, but the existing system could not fund it.
The problem was structural. NIH study sections and peer-review panels necessarily divide enormous scientific territories into manageable areas of expertise. Researchers submit proposals that specialists can evaluate according to established methodologies, existing literature, feasibility, and expected empirical outcomes. That makes administrative sense, but it also creates a profound problem when the scientific question exists between the domains.
The system becomes exceptionally good at funding another measurement, another localized experiment, another dataset, and another paper. Each can be legitimate science, and each can contribute another piece of knowledge.
But where is the study section for explaining how the whole thing works?
There was no obvious bureaucratic home for the kind of theoretical architecture Hawkins wanted to build. The professors inside universities operated inside the same funding ecosystem. Graduate students worked inside their laboratories; laboratories depended on grants; grants were evaluated by specialists; and specialists were trained within disciplines that had developed their own methodologies, vocabularies, journals, and standards of evidence.
Nobody had to conspire to stop grand theoretical work. The structure itself selected against it.
The Institutional Trap
Viewed as a whole, Hawkins's path becomes almost absurd. He wanted to understand intelligence by understanding the brain. MIT did not provide the path, so he went into technology. He tried Berkeley. Berkeley did not provide the path, so he returned to technology.
He built the GRiDPad, founded Palm, and helped create the PalmPilot. While serving as CTO, he deliberately carved out roughly twenty hours every week to continue the neuroscience work that academia had not given him a place to pursue. He helped raise his children, ate breakfast with them, came home for dinner, and returned to neuroscience whenever the commercial work allowed it.
When Palm or Handspring required him, the neuroscience went back on the shelf temporarily. Then he returned to it, building the Visor, helping push handheld computing toward the Treo, and accumulating technological credibility, intellectual independence, and financial capital. Eventually, he became wealthy enough to build the neuroscience institution he had been looking for all along.
That is an extraordinary story of persistence. It is also an extraordinary indictment of the route required to exercise it, because most researchers cannot do what Jeff Hawkins did.
A doctoral student who encounters the same structural wall cannot ordinarily leave the university, invent an entire category of computing, build successful technology companies, accumulate enough money to finance independent theoretical neuroscience, and then return to the scientific question the institution would not let them pursue.
They have to stay inside the system. That means finding an adviser, joining a laboratory, obtaining grants, publishing papers, passing peer review, securing a postdoctoral position, winning another grant, and obtaining tenure. Every step rewards questions that the existing machinery knows how to evaluate.
The result is not a conspiracy. It is something far more durable: an incentive structure.
ADHD Did Have Its Jeff Hawkins
This is where the Hawkins story collides directly with ADHD, because it would be wrong to say that nobody inside adult ADHD ever looked at the existing framework and recognized that enormous pieces of the human experience were missing.
ADHD did have its Jeff Hawkins. His name is Dr. William Dodson.
Dodson came at the problem from a very different direction. Hawkins was an engineer looking at neuroscience and asking how the entire machine worked. Dodson was a psychiatrist looking at decades of patients and asking why the existing description of ADHD did such a poor job of explaining the people sitting in front of him.
One of the most revealing questions emerged from his work with physicians. Over the course of his career, Dodson treated 159 practicing physicians with ADHD. According to the figures he has shared, only one had begun ADHD medication before entering medical practice.
That left 158 people with ADHD who had somehow made it through undergraduate education, medical school, residency, licensing, and into the practice of medicine without the treatment that the conventional impairment model would suggest should have been crucial to navigating those environments.
Dodson's question was wonderfully simple:
How the hell did the other 158 get there?
That is a Hawkins question. It does not deny impairment, nor does it pretend that ADHD disappears because some people succeed. It asks what a deficit-centered model is missing if it can describe why someone should struggle but cannot explain why another person with the same diagnosis can navigate one of the most demanding educational pipelines in society.
Dodson began looking at the people the pathology literature had largely ignored: successful ADHD adults. What conditions had allowed them to develop? What happened when the environment fit the nervous system rather than fighting it? Who had helped them through the periods when it didn't?
Again and again, Dodson encountered something difficult to put into the conventional clinical framework. Successful ADHD adults frequently described someone in their lives who had believed in them, protected their sense of possibility, helped them recover from failure, and continued seeing competence when institutions were seeing deficits. Dodson initially described that person as a kind of cheerleader, although advocate gets closer to what the role actually was.
The observation pointed toward a much larger question. Perhaps the outcome of an ADHD life could not be understood entirely by measuring the individual nervous system. Perhaps the relationship between that nervous system and the people and environments surrounding it mattered far more than the prevailing clinical model had allowed.
That question leads directly toward empathy, emotional regulation, rejection sensitivity, and the social environment surrounding ADHD. It also leads directly into another institutional wall.
The Things That Don't Sit Still
Dodson became one of the clinicians pushing adult ADHD toward phenomena that the established framework had struggled to accommodate, particularly emotional dysregulation and what he calls rejection sensitive dysphoria.
The difficulty is not merely theoretical. It is methodological. Much of modern behavioral research depends on phenomena that can be reliably produced, isolated, operationalized, and measured. Researchers need something that can happen often enough, predictably enough, and under sufficiently controlled conditions to generate analyzable data.
An intense rejection-sensitive episode does not politely arrive at 2:15 Tuesday afternoon because a research assistant has reserved the laboratory. It happens when it happens.
That creates something I have described elsewhere using the analogy of the Heisenberg Uncertainty Principle—not because Dodson uses that phrase, and certainly not because quantum mechanics somehow explains RSD, but because the measurement problem is revealing. The phenomenon researchers most want to observe is spontaneous, context dependent, emotionally intense, and difficult to reproduce ethically on command.
The things that are easiest to measure therefore become the things most likely to become science, while the things that are hardest to capture become easier to leave outside the model. That does not make the difficult phenomenon unreal. It makes it difficult to study using the machinery the institution has built.
Hawkins ran into that problem from the theoretical side.
Dodson ran into it from the clinical side.
The Training Pipeline
The consequences do not remain inside research journals. They travel downstream.
During the years Dodson's RSD and Empathy Research Group was operating, the group examined the extraordinary training gap surrounding adult ADHD in psychiatry. Dodson later shared with me the result: roughly 93 percent of adult psychiatrists surveyed reported receiving no meaningful ADHD education during their four-year psychiatric residency, and ADHD contributed zero questions to the relevant board examination.
Consider the institutional loop that creates. A research literature concentrates overwhelmingly on certain measurable aspects of ADHD. Medical education draws from that literature. Residency programs allocate their limited training time according to what the profession has established as central, while board examinations reinforce those priorities. Another generation of psychiatrists then enters practice having received little meaningful preparation for adult ADHD.
Then the patients arrive.
The clinician encounters emotional volatility, rapid speech, rejection sensitivity, anxiety, trauma, shutdown, interpersonal conflict, or other phenomena that were never integrated adequately into their ADHD training. The clinician reaches for the diagnostic frameworks they were trained to recognize.
The research blind spot becomes a training blind spot, and the training blind spot becomes a clinical blind spot. Eventually, the clinical blind spot helps produce the next generation of data.
That is how an institutional half-truth can perpetuate itself without anyone inside the institution acting dishonestly.
The Difference Was Capital
The parallel between Hawkins and Dodson only goes so far, but where it breaks may be the most revealing part of the comparison.
Both looked at fields filled with observations and believed something larger was missing. Both were willing to ask questions that crossed established boundaries. Both encountered problems that did not fit comfortably inside the existing institutional machinery.
But Hawkins had something Dodson did not.
Palm.
Hawkins could walk away from the conventional research pipeline because he had accumulated enough money to build another one. When theoretical neuroscience had no institutional home, he could create Redwood Neuroscience Institute. He could bring researchers together, give them time to think, and finance a scientific environment organized around the question rather than around the existing funding category.
Dodson could not simply write the check.
That difference matters enormously. We like to tell the history of science as though transformative ideas win because evidence eventually triumphs over resistance. Hawkins's biography shows something considerably messier. Sometimes an unconventional scientific question survives because the person asking it has enough independence to keep asking after the institution says no.
Hawkins eventually acquired that independence. Dodson had to keep working inside and around the structure.
One had Silicon Valley capital behind him; the other had to find funding.
That may tell us as much about what gets studied as any argument about scientific curiosity ever could.
When Measurement Becomes the Science
This is where Hawkins and Dodson become the lenses through which we can look at fifty years of ADHD research.
Academic ADHD research did not spend those decades doing nothing. Quite the opposite. It produced mountains of data measuring reaction times, working memory, inhibition, response variability, delay aversion, cortical thickness, neurotransmitter activity, school performance, occupational outcomes, medication response, comorbidity, family conflict, accidents, incarceration, substance use, and virtually every other measurable impairment associated with the diagnosis.
Much of that work is valuable, and much of it is true. That is precisely where the problem begins, because a collection of truths about pieces of a system is not necessarily a theory of the system.
If researchers repeatedly observe that an ADHD child cannot initiate an assignment, that observation is real. If researchers measure poor performance on a response-inhibition task, that measurement is real. If a student loses track of three hours while engaged in something intensely interesting, that observation is real. If the same student cannot sustain ten minutes of attention to something that produces no engagement, that observation is real too.
Now we have a problem, because all of those observations have to exist inside the same nervous system. Naming each observation does not answer that problem.
Description is not mechanism.
Without a sufficiently encompassing mechanistic framework, behavioral science can accumulate an enormous number of individually accurate observations while still constructing an incomplete picture of the human being producing them. That is the danger of the half-truth: it isn't false; it is incomplete.
When institutions organize themselves primarily around studying impairment, impairment becomes extraordinarily visible while competence under different conditions can disappear from the frame entirely.
The Profession of Half-Truths
This is how an academic literature can become simultaneously enormous and incomplete.
ADHD research became extraordinarily sophisticated at measuring failure against externally imposed conditions: sustain attention here, inhibit this response, begin this task now, estimate this interval, organize these materials, sit still for this period, perform consistently regardless of whether the nervous system finds the activity interesting.
When the person fails, the failure becomes data.
But the same nervous system can perform very differently under a different set of conditions. Interest can change the result. Novelty can change it. Challenge can change it. Urgency can change it. A person who cannot sustain ten minutes of engagement under one condition can remain locked onto another problem for six, eight, or twelve hours. During that state, even the experience of time itself can change.
Those observations belong to the same machine, and a theory of ADHD eventually has to explain all of them.
That is the lesson Hawkins carries into this chapter, and it is the question Dodson's clinical work keeps forcing back onto the table. Neither requires us to discard the mountains of empirical ADHD research that already exist. The problem is not that the pieces are worthless.
The pieces are not the machine.
Hawkins spent decades trying to construct an architecture capable of explaining how the pieces of the neocortex fit together. Dodson spent decades listening to ADHD patients whose lived experience contained phenomena that the established clinical architecture did not adequately explain. Both encountered institutions extraordinarily capable of studying what fit inside their existing methods and considerably less capable of reorganizing themselves around what did not.
Hawkins eventually became wealthy enough to build the place where he could ask his questions anyway. Dodson did not have a PalmPilot, and that difference matters.
Chapter One began with a deceptively simple proposition: ask what the machine is doing, look for the repeated architecture, build a model, generate predictions, and try to break it. In principle, it isn't that hard.
Then you try doing it inside the institutions responsible for deciding what gets researched, what gets funded, what gets taught, what gets tested, and ultimately what clinicians are trained to see.
And suddenly—
It’s actually that hard.
References
Crick, F. H. C. (1979). Thinking about the brain. Scientific American, 241(3), 219–232.
Dodson, W. W. Clinical lectures, writings, and correspondence concerning adult ADHD, emotional dysregulation, rejection sensitive dysphoria, empathy, and adult psychiatric training.
Hawkins, J. (2002). Genesis of Palm Computing. Stanford Technology Ventures Program / Stanford Entrepreneurial Thought Leaders.
Hawkins, J. (2002). Serial Entrepreneurship: Redwood Neuroscience Institute. Stanford Technology Ventures Program / Stanford Entrepreneurial Thought Leaders.
Hawkins, J., & Blakeslee, S. (2004). On Intelligence: How a New Understanding of the Brain Will Lead to the Creation of Truly Intelligent Machines. Times Books / Henry Holt and Company.
Hawkins, J. (2005). Entrepreneurship Viewed as a Tool and When to Use It in Industry & Science. Stanford Entrepreneurial Thought Leaders Seminar, Stanford Technology Ventures Program.
Hawkins, J. (2009). Inside the Mind of a Reluctant Entrepreneur. Stanford Entrepreneurial Thought Leaders Seminar, Stanford Technology Ventures Program.
Hawkins, J. (2009). The Twenty-Year Plan. Stanford Technology Ventures Program / Stanford Entrepreneurial Thought Leaders.
Hawkins, J. (2012). Intelligence and the Brain: Recent Advances in Understanding How the Brain Works. Charles M. and Martha Hitchcock Lectures, University of California, Berkeley.
Hawkins, J. (2012). Intelligence and Machines: Creating Intelligent Machines by Modeling the Brain. Charles M. and Martha Hitchcock Lectures, University of California, Berkeley.
Hawkins, J., & Ahmad, S. (2016). Why neurons have thousands of synapses, a theory of sequence memory in neocortex. Frontiers in Neural Circuits, 10, 23.
Hawkins, J., Lewis, M., Klukas, M., Purdy, S., & Ahmad, S. (2019). A framework for intelligence and cortical function based on grid cells in the neocortex. Frontiers in Neural Circuits, 12, 121.
Hawkins, J. (2021). A Thousand Brains: A New Theory of Intelligence. Basic Books.
University of California, Berkeley. (2005). Jeff Hawkins, computing pioneer, endows new center to develop model of brain.