Track brain function across diseases at scale, and you see what imaging could never reveal.
For generations, the brain has been medicine's most elusive organ to monitor continuously — neurologists forced to rely on brief snapshots while cardiologists long enjoyed streams of data from home. Beacon Biosignals, founded in 2019 by neuroscientist Jake Donoghue and engineer Jarrett Revels, is working to close that gap by deploying FDA-cleared EEG headbands that capture lab-grade brain activity during sleep, night after night, in patients' own homes. The wager is that sleep — with its rich, structured neural language — holds early warnings of Alzheimer's, Parkinson's, and other neurological diseases that conventional medicine only recognizes after significant damage has already been done. With $97 million raised and over 40 clinical trials underway, Beacon is attempting to do for the brain what genomics did for cancer: make its mysteries legible, and legible early.
- Neurology has long lagged behind other medical fields, relying on reactive, symptom-driven care while diseases like Alzheimer's and Parkinson's silently progress for years before diagnosis.
- Traditional EEG monitoring, confined to sleep labs and single-night sessions, could never generate the longitudinal, high-volume brain data needed to detect subtle early changes in neural architecture.
- Beacon's at-home EEG headband removes those constraints, allowing continuous data collection across tens of thousands of patients and feeding machine-learning models trained to recognize the electrical signatures of disease before symptoms appear.
- The acquisition of a sleep apnea testing platform serving 100,000 patients annually dramatically expands Beacon's reach, turning routine screenings into potential windows on long-term neurological health.
- With $97 million in fresh funding, the company is now racing to build something that does not yet exist: a longitudinal record of human brain function at scale, a foundation model for the brain that could redefine early intervention across neurological medicine.
The brain has long resisted the kind of continuous monitoring that cardiologists take for granted. Neurologists have worked from snapshots — office visits, imaging scans, a patient's own account of what they remember — while diseases like Alzheimer's and Parkinson's quietly advance. That disparity is what Jake Donoghue, a neuroscientist trained at MIT and later in clinical neurology and psychiatry, set out to address when he co-founded Beacon Biosignals in 2019 alongside research engineer Jarrett Revels.
Donoghue's formative insight came from watching oncology transform through genomic sequencing into something data-driven and precise, while neurology remained largely reactive, waiting for symptoms to declare themselves. His answer was to look to sleep. Neural activity during sleep is not only far more abundant than during waking hours but also more structured — almost a language. An FDA-cleared EEG headband worn at home night after night could capture data no sleep lab ever could, unconstrained by a single unfamiliar night in a clinical setting. Machine-learning algorithms could then read that data for changes in sleep architecture — shifts in REM and slow-wave sleep that may precede cognitive decline by years.
Beacon has now deployed this approach across more than 40 global clinical trials, studying conditions from major depressive disorder and schizophrenia to narcolepsy, Alzheimer's, and Parkinson's. For Donoghue, whose own grandfather was affected by Parkinson's, the stakes are personal: the goal is to move these diseases from late recognition, when damage is already substantial, to early, data-driven detection — perhaps even during a routine sleep apnea screening a patient undergoes without any neurological concern.
To accelerate that vision, Beacon acquired an at-home sleep apnea testing company serving more than 100,000 patients annually across the United States, and in November raised $97 million to expand further. The long-term ambition is a longitudinal record of brain function that does not yet exist — a growing foundation model built from the electrical signatures of many patients over time, capable of revealing novel disease subtypes and early warning signs that static tools like imaging or genetic sequencing could never surface. What began as a wearable headband is becoming an attempt to give neurology the same data-driven precision that has transformed the rest of medicine.
The human brain has long resisted the kind of systematic monitoring that cardiologists take for granted. A doctor can track a patient's heart function from home with a simple device; neurologists, by contrast, have relied on snapshots—office visits, imaging scans, the patient's own account of what they remember. This gap in precision medicine is what Jake Donoghue noticed during his training in neurology and psychiatry, and it became the seed of an idea that has now grown into Beacon Biosignals, a company mapping brain activity during sleep to catch neurological disease before symptoms ever appear.
Donoghue, who completed his PhD in neuroscience at MIT under Earl K. Miller, spent his clinical training rotating through oncology, neurology, and psychiatry. In oncology, he watched genomic sequencing transform cancer care into something data-driven and precise. In neurology and psychiatry, he saw the opposite: treatment remained largely iterative, reactive, dependent on waiting for symptoms to declare themselves. The contrast struck him as a problem worth solving. He began developing the conviction that if you could process large amounts of brain data and correlate it with actual brain function, you could transform how neurological diseases were identified and treated. By the end of his training, he had recruited Jarrett Revels, a research software engineer he'd met at MIT's Julia Lab, and together they founded Beacon in 2019.
The company's insight was deceptively simple: sleep is the ideal window into brain function. Neural activity during sleep is not only more abundant than during waking hours—it can be an order of magnitude higher—but also more structured, almost like a language. A lightweight EEG headband, worn at home night after night, could capture lab-grade data that no sleep lab ever could, because patients wouldn't be constrained by a single night in an unfamiliar place. Machine-learning algorithms could then extract patterns from that data: how much time a patient spent in different sleep stages, how many small awakenings interrupted the night, what subtle changes in sleep architecture might precede cognitive decline. The FDA cleared the device, and Beacon began partnering with pharmaceutical companies to deploy it across clinical trials. To date, the company has been involved in over 40 trials globally, studying conditions ranging from major depressive disorder and schizophrenia to narcolepsy, Alzheimer's disease, and Parkinson's disease.
What makes this approach powerful is its potential for early detection. Neurodegenerative diseases like Parkinson's and Alzheimer's often announce themselves through changes in sleep architecture years before a patient experiences any clinical symptoms. By analyzing patterns in rapid-eye-movement sleep and slow-wave sleep, Beacon's algorithms can identify these early markers—the subtle electrical signatures of a brain beginning to change. Donoghue, whose grandfather was affected by Parkinson's, sees this as an opportunity to shift these diseases from late recognition, when damage is already substantial, to much earlier, data-driven detection. A patient might come in for a routine sleep apnea screening and never know that the data collected that night could become, years later, a window into disease progression before symptoms emerge.
Last year, Beacon accelerated this vision by acquiring an at-home sleep apnea testing company serving over 100,000 patients annually across the United States. In November, the company raised $97 million to expand further. The long-term ambition is to build longitudinal records of brain function over time—a dataset that doesn't yet exist but that Beacon believes will transform brain health. Each patient becomes part of a growing foundation model of the brain, a repository of electrical signatures that researchers can mine for patterns, for novel disease subtypes, for the early warning signs that might allow intervention before irreversible damage occurs. The brain, Donoghue notes, is an electric organ that changes through synaptic plasticity. Track that function across many diseases at scale, and you begin to see things that static modalities like genetic sequencing or imaging could never reveal. What was once a constrained, facility-based test becomes a scalable source of high-quality data for diagnostics, drug development, and the kind of precision medicine that has transformed other fields but has eluded neurology for too long.
Notable Quotes
There's a step-change in what becomes possible when you remove the sleep lab and bring clinical-grade EEG into the home. It turns sleep from a constrained, facility-based test into a scalable source of high-quality data for diagnostics, drug development, and longitudinal brain health.— Jake Donoghue, CEO of Beacon Biosignals
We're analyzing features of rapid-eye-movement and slow-wave sleep to detect early changes that precede clinical symptoms. It's an opportunity to move these diseases from late recognition to much earlier, data-driven detection.— Jake Donoghue
The Hearth Conversation Another angle on the story
Why sleep, specifically? Why not monitor the brain during the day, when people are actually living their lives?
Because the brain during sleep is doing something more intense and more legible. Neural activity is an order of magnitude higher, and it's organized—almost like a language. You can read it. During waking hours, the brain is too noisy, too scattered across a thousand different tasks.
But doesn't that mean you're missing something? The disease happens during the day, in real life.
The disease leaves traces in sleep. Changes in sleep architecture can emerge years before someone notices symptoms. It's like reading the brain's diary—what it's processing, what's breaking down, before the person even knows something is wrong.
And the machine learning—what's it actually doing with all that data?
It's finding patterns humans can't see. How much time in each sleep stage, how many micro-awakenings, what the electrical signatures look like. Then it links those patterns to outcomes. That's never been done with this precision before.
So you're building a kind of brain fingerprint.
More than that. We're building a longitudinal record. A patient comes in for a sleep apnea test today, and if they develop Parkinson's in ten years, that earlier data becomes a window into the disease before it announced itself. Routine testing becomes the foundation for detecting disease earlier, maybe before symptoms ever begin.
That's a different kind of medicine than we have now.
It is. It's the kind oncology moved toward with genomics. We're trying to bring that precision to the brain.