365bet体育在线世界杯

features

A Pioneering Mind

Alumnus Terrence Sejnowski is a groundbreaking neuroscientist and AI researcher who has spent his career connecting ideas—and people

Whatever the topic—gravity waves or brain waves, nebulae or neurons—, (CIT '68) PhD, dives deep.

Sejnowski began his scientific career pondering exotic deep-space phenomena such as black holes, gravitational waves and the nebulae that give birth to stars.

In graduate school, however, his interest shifted to neuroscience. And for the past 50 years—much of it spent at the California-based , one of the world's premier biomedical research institutions—Sejnowski has sought to unravel the deepest mysteries of the human brain: how it comprehends the world, stores memories and learns new skills.

In the process, the Cleveland native has revolutionized two fields—neuroscience and artificial intelligence (AI)—and made lasting contributions to many more, including discoveries that could lead to new treatments for a wide range of neurological disorders. He also helped forge national science policy as an advisor to the White House BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative and become a prolific science communicator, explaining advances in neuroscience and AI through books, media appearances and a free online course ("Learning How to Learn," offered through Coursera) that has reached nearly 5 million people around the world.

In 2024, Sejnowski won the Brain Prize—a major international neuroscience research award, presented annually by the king of Denmark—for pioneering theoretical and computational neuroscience and laying the groundwork for brain-inspired AI.

"Whenever I want to know something about the brain, I just call up Terry," said computer scientist , PhD, who won a 2024 Nobel Prize in Physics—in part for work he did with Sejnowski in the 1980s, during what he regards as the most exciting collaboration he's ever had. "He puts his whole life into brain science."

Sejnowski's brilliance and drive are matched by his warmth and humanity. He has a genius for bridging different fields and connecting diverse groups of people—qualities that have endeared him to his colleagues and enabled him to nurture a cadre of younger scientists. Indeed, when he first made the leap to neuroscience, AI was in its infancy, and researchers were just beginning to study the brain using computational techniques. Sejnowski made huge strides in both areas while mentoring a whole new generation of computational neuroscientists who have built upon his work.

"Terry and the people that he trained brought the field to life," said Peter Thomas, PhD, a professor of mathematics at Case Western Reserve, who worked in Sejnowski's lab from 2000 to 2004.

THINKING MACHINES

Sejnowski's scientific curiosity surfaced early. "Even in grade school, I was fascinated by astrophysics, the solar system, galaxies—and science fairs," he said, adding that he once triggered an elementary-school evacuation with a realistic model volcano that filled his classroom with smoke. ("It worked beautifully," he recalled. "But I didn't take into account that the windows were all closed.")

As an undergraduate at what's now CWRU, Sejnowski studied general relativity with former Manhattan Project physicist Joseph Weinberg, PhD. He then went to Princeton University for graduate work on black holes and gravitational waves with legendary theoretician John Wheeler, PhD. The existence of such phenomena had not yet been proved, however; Sejnowski predicted it would take 40 years to develop instruments capable of detecting them. (He was right.)

Fortunately, Sejnowski discovered an emerging field where the problems were just as interesting, but data was more readily available: neuroscience.

"Here was a topic that was as mysterious and fascinating as the universe: the brain," he said. "And the difference was you could study it with your own hands in your own lab."

365bet体育在线世界杯 Princeton, Sejnowski acquired a new advisor: physicist John Hopfield, PhD, who shared the 2024 Nobel with Hinton for creating an artificial neural network that borrowed concepts from statistical physics to store memories of visual images. (Neural networks use simplified neurons that simulate the behavior of real neurons in the brain and play a central role in modern AI.)

In 1979, Sejnowski applied to present at a workshop on neural networks organized by Hinton, claiming to have discovered the "machine code" of the brain. "I got this application from somebody, and I couldn't decide whether they were crazy or brilliant," Hinton said.

The two became fast friends and, in the early 1980s, when Sejnowski was on faculty at Johns Hopkins University, they further mined ideas from statistical physics to design a neural network that could not only remember images but learn to generate new ones. "We were convinced that we had figured out how the brain works," Sejnowski said.

Their creation—dubbed a "Boltzmann machine" for Ludwig Boltzmann, founder of statistical physics—paved the way for today's deep-learning algorithms and generative AI models and was cited in Hinton's Nobel award.

Soon after, Sejnowski developed a neural network called NETtalk that learned to correctly pronounce English words by analyzing phonetic transcriptions of real human speech—a precursor to modern neural network-based AI programs such as ChatGPT that learn to manipulate language by analyzing written text. "Everybody was amazed," Sejnowski recalled. "I was on the 鈥楾oday' show."

In the late '80s, Sejnowski joined the faculties of the University of California San Diego and the Salk Institute, establishing a lab at Salk that would become one of the principal engine rooms of computational neuroscience. Sejnowski recruited students and postdocs with backgrounds as varied as his own鈥攅ngineers and electrophysiologists, mathematicians and computer scientists—to build computational models for testing ideas about how the brain processes and stores information.

In the ensuing decades, Sejnowski and his colleagues pursued projects such as analyzing the behavior of individual neurons, modeling the brain's visual system and developing a machine-learning algorithm to analyze brain signals from medical tests including functional MRIs. By constantly updating their models with fresh experimental data and using their theoretical insights to guide further experiments, they helped illuminate not only how the brain works at the level of cells and molecules, but also how entire brain regions such as the cortex and hippocampus interact to enable perception, memory and learning.

Understanding how a healthy brain forms memories could lead to new ways to address memory problems caused by diseases such as Alzheimer's and Parkinson's, while understanding the properties of healthy nerve cells could lead to better treatments for neurological disorders such as multiple sclerosis.

Some of Sejnowski's discoveries have traveled far beyond neuroscience. The algorithm for interpreting brain data, for example, also turned out to be useful for improving voice quality on mobile phones. And the startup that one of Sejnowski's postdocs launched to commercialize the technology was ultimately acquired by the telecom giant Qualcomm.

That propensity for doing fundamental research with wide-ranging practical applications—and for drawing on disciplines as seemingly disparate as biophysics and information theory—resulted in Sejnowski's election to the National Academy of Sciences, the National Academy of Medicine, the National Academy of Engineering and the National Academy of Inventors.

"Terry is a pioneer of neuroscience, and he's a pioneer of AI," said , PhD, a CWRU Distinguished University Professor and vice-chair of the Department of Physics, where Sejnowski is a member of the strategic advisory group. "It's not for nothing that he is one of the single handful of people who are members of all four national academies."

SOCIAL SCIENCE

The range and volume of the lab's output are a reflection not only of Sejnowski's own broad interests and boundless energy (he is famous for firing off emails at 3 a.m.), but also of his gift for creating a fertile work environment.

While most labs skew toward either experimentation or theory, Sejnowski valued both. "There were people doing experiments, and then next door there were carrels full of theorists typing away on computers," Thomas recalled. Every student or postdoc who left the lab received a Swiss Army knife from Sejnowski鈥攁 metaphor for the wide range of skills required to do neuroscience, and of the multifaceted nature of the brain itself.

Sejnowski engineered the environment to encourage serendipitous encounters. His lab features a common area with couches and a kitchen, and the hallways leading to it are lined with whiteboards for scribbling ideas. Afternoon tea is held on Mondays and Fridays: A bell is rung, and everyone gathers to snack and schmooze. Even members of other labs drop in to enjoy the convivial atmosphere and exchange ideas.

"He really understood the social prerequisites for good science," said Michael Lewicki, PhD, a CWRU professor in computer and data sciences who worked in the lab from 1996 to 1998.

Though he continues to pursue a varied research agenda, Sejnowski has also taken the plunge into science policy and communication.

In 2013, Sejnowski joined the advisory committee for the BRAIN Initiative, a multibillion-dollar project launched by the Obama administration that has spurred the development of new technologies to study the brain. Some of those technologies have fed directly into Sejnowski's own research, such as his ongoing work on the connections among sleep, memory and learning: Thanks to advances in brain recording techniques, he has shown that waves of electrical energy flowing between different brain regions during slumber help cement new experiences into long-term memory, facilitating lifelong learning.

Most recently, Sejnowski has been exploring the possibility that contemporary neural networks such as the one that powers ChatGPT might provide insights into how the human brain works, just as Boltzmann machines once promised to do.

He has also thrown his weight into communicating the nature and benefits of the latest advances in neuroscience and AI to the general public. "They deserve to know what [their tax dollars are] being used for, and the impact it is going to have on them," he said.

Since 2018, Sejnowski has published two books to help students and teachers improve their performance by exploiting what neuroscientists have discovered about memory and learning, and two more explaining the latest advances in AI. He is now working on a book about the revolution in neuroscientific research being driven by the BRAIN Initiative—a revolution that promises to further penetrate the mysteries of the human brain, thereby boosting our ability to repair it.

"When you were born, did you get an instruction manual for your brain?" Sejnowski asked. "We don't have it. We still don't understand enough about it. But we're going to figure it out."

And when we do, much of that manual likely will have been written by Sejnowski himself.

— ALEXANDER GELFAND