Janelia Research Campus Launches Decade-Long Initiative to Decode the Vertebrate Brain Using AI and a Transparent Fish
核心洞察
Janelia Research Campus (搜索) announces a decade-long effort to build the first mechanistic account of how a vertebrate brain generates behavior, linking molecules to action.
The initiative centers on Danionella, a tiny transparent fish that remains clear into adulthood, enabling whole-brain imaging during complex behaviors like mating and social interaction.
A novel "AI-in-the-Loop" methodology will integrate artificial intelligence as a genuine partner in biological discovery, from hypothesis generation to experimental execution.
The Howard Hughes Medical Institute (搜索)'s Janelia Research Campus (搜索) is embarking on what it describes as one of its most ambitious scientific undertakings to date: a decade-long effort to construct the first mechanistic account of how a vertebrate brain generates behavior. The initiative pairs cutting-edge neuroscience with a novel "AI-in-the-Loop" discovery methodology and places a tiny, transparent fish called Danionella at the center of the research.
"This new direction is precisely the kind of long-horizon scientific challenge that Janelia was created to pursue," said HHMI President Erin O'Shea. "We are tackling a problem whose solution will benefit not only science but human health, and in the process, we are reenvisioning how science is done."
A Question Decades in the Making
Neuroscience has long pursued a fundamental question: How does a vertebrate brain produce the rich, flexible behavior of a living animal? Answering it is core to making real progress against brain disorders that impact hundreds of millions of people, yet the question has historically outrun the tools science had available. Two converging forces have now brought the previously impossible within reach — a new generation of tools that permit observation of whole biological systems with unprecedented completeness, and artificial intelligence capable of parsing data at scales no human scientist can manage.
Scientists still cannot explain, in mechanistic detail, how the brain integrates sensory information, internal states, memories, and physiology to produce behaviors. Janelia's new strategy aims to change that by connecting two levels that have rarely been studied together: the neural circuits that generate behavior, and the cellular and molecular machinery that drives it.
AI-in-the-Loop: A New Mode of Discovery
Central to the initiative is the "AI-in-the-Loop" methodology, a research approach in which AI systems do far more than analyze experiments after the fact. They help interpret data, generate hypotheses, design experiments, and — as the methodology matures — increasingly direct experimental execution alongside human scientists. AI can simulate proposed approaches before a single experiment runs, help design the tools and methods the science requires, and, once experiments are underway, detect unexpected patterns, flag candidate circuits, and update predictions as fresh data arrive.
The discovery loop compresses from months to days. AI and laboratory experiments work together as one continuous system.
The effort works in close partnership with AI@HHMI, the Institute-wide initiative bringing AI methods to biomedical research across a wide range of organisms and scientific questions. Discoveries from Janelia's grand challenge flow outward to HHMI Investigators and Freeman Hrabowski Scholars, while findings from their collaborations in turn sharpen Janelia's own work.
Danionella: The Transparent Vertebrate
At the center of Janelia's new vision is Danionella, a fish barely larger than a grain of rice that remains transparent throughout its entire life — an unusual trait among vertebrates. Larval zebrafish, the current standard for whole-brain imaging in a vertebrate, are transparent only as infants, and their behavioral repertoire is correspondingly limited. A young larva can swim, hunt prey, and respond to stimuli, but it cannot court a mate, defend territory, or navigate the complex social world of an adult fish.
Danionella is different. Because it stays clear into adulthood, researchers can, for the first time, watch a vertebrate brain in action while the animal performs the full range of behaviors that make a brain worth studying — including mating, social interaction, learned navigation, and decision-making. Researchers could potentially record activity across the brain while simultaneously tracking behavior, sensory input, and body physiology, all in the same living vertebrate.
"Twenty years ago, Janelia committed to determining the wiring diagram of the fruit fly brain and developing sophisticated genetic tools for studying the roles of individual cell types. Many thought such efforts were infeasible, but we succeeded," said Gerry Rubin, Janelia's founding Executive Director. "The same approach can be applied to Danionella. But having a transparent animal where brain-wide activity patterns can be measured and interpreted using AI provides possibilities we lacked in the fly."
Building the Tools from the Ground Up
Designing tools for a new model organism will take time. In the first several years of the new initiative, Janelia scientists will build the genetic tools, imaging systems, behavioral assays, and AI methods required to study Danionella at this depth. In parallel, the fruit fly and larval zebrafish serve as proving grounds where AI approaches and experimental strategies will be refined. Each system will yield real biological discoveries along the way — findings that matter in their own right, even as they sharpen the questions asked next.
Many of the necessary tools do not yet exist. Scientists still do not know whether AI systems can reliably help generate mechanistic biological understanding at the level the project requires. O'Shea said that uncertainty is part of the point. "The test is simple: Does this research require Janelia to succeed? We are concentrating Janelia's distinctive capacity on big bets that no one else can make."
The Janelia Advantage
The new initiative marks a major evolution for the campus as it approaches its twentieth anniversary. Founded in 2006 to pursue scientific challenges beyond the reach of traditional university or industry labs, Janelia became known for building transformative tools, including fluorescent sensors, imaging technologies, connectomes, and genetic methods now used throughout biology.
"Few places in the world can bring together tool builders, tool users, and AI experts in the deeply integrated way this effort requires," said Nelson Spruston, Janelia's Executive Director. "Janelia was designed for exactly this kind of transformational science."
Rubin, who helped shape Janelia's original scientific strategy, steps into a new role as Head of Biology to lead aspects of the new effort. The strategy leverages Janelia's commitment to openly sharing tools, technologies, and datasets with the broader scientific community.
The answers that emerge could unravel the logic of brain function and, over the long term, reveal what goes wrong in disorders like Alzheimer's disease (搜索), autism (搜索), and depression (搜索).
