Sleep is when the brain's electrical activity becomes almost language-like
For generations, the brain has resisted the kind of quiet, continuous observation that the heart has long permitted — leaving neurologists to work from fleeting snapshots rather than the full arc of a life. Beacon Biosignals, born from that disparity, has built a wearable EEG headband that gathers clinical-grade brain data while people sleep in their own homes, feeding machine-learning systems capable of detecting the earliest whispers of neurological disease. Founded by MIT-trained neuroscientist Jake Donoghue and software engineer Jarrett Revels, the company now supports over forty clinical trials and, through a recent acquisition, holds longitudinal brain health records for more than 100,000 patients annually. The ambition is not merely diagnostic but prophetic — to find disease in the silence before it speaks.
- Neurologists have long been forced to diagnose from brief, artificial snapshots while cardiologists have monitored patients continuously at home for decades — a disparity that left brain disease detection perpetually behind.
- Sleep architecture carries disease signals years before symptoms surface, meaning every night of unrecorded home sleep has historically been a missed opportunity for early intervention in conditions like Alzheimer's and Parkinson's.
- Beacon's FDA-cleared EEG headband now runs across 40+ pharmaceutical clinical trials, replacing expensive lab-based recordings with scalable, natural-environment data that captures how the brain actually behaves over time.
- The acquisition of a sleep apnea testing platform serving 100,000+ patients annually has shifted Beacon from trial partner to keeper of longitudinal brain health records — a dataset large enough to reveal disease subtypes no single study could identify.
- With $97 million raised and a foundation model of the brain taking shape, the company is repositioning neurological medicine from reactive treatment toward detection before the disease becomes undeniable.
Jake Donoghue was midway through neurology and psychiatry training when the inequality became impossible to ignore. Cardiologists could watch their patients' hearts over weeks from home. Neurologists had nothing comparable — only brief, expensive lab recordings taken while patients tried to sleep in unfamiliar rooms, wired to machines. That frustration became Beacon Biosignals.
Donoghue, an MIT neuroscience PhD who trained under Earl K. Miller and completed clinical work at Massachusetts General Hospital and Boston Children's Hospital, co-founded the company in 2019 with Jarrett Revels, a research software engineer from MIT's Julia Lab. Their device is a lightweight, FDA-cleared EEG headband worn at home during natural sleep. Machine-learning algorithms parse the resulting signals — sleep stage durations, micro-awakenings, subtle shifts in brain architecture — extracting patterns that clinical labs rarely had the time or scale to find. More than forty clinical trials now use the platform to study treatments for conditions ranging from major depressive disorder to Alzheimer's and Parkinson's disease.
The core insight is temporal. Sleep is when the brain's electrical activity becomes most structured and revealing, and changes in sleep patterns can precede the first symptoms of neurodegeneration by years. A routine sleep apnea screening today might, years later, serve as a baseline against which early Parkinson's progression becomes visible — turning diagnostic data into something prognostic.
Last year, Beacon acquired an at-home sleep apnea testing company serving more than 100,000 patients annually, transforming its reach from clinical trial partner to custodian of longitudinal brain health records at scale. The company raised $97 million in November to accelerate that expansion. Donoghue's goal is a foundation model of the brain — a dataset broad and deep enough to reveal disease subtypes that static tests like genetic sequencing or imaging cannot capture, and to find the markers that precede symptoms before the disease takes hold.
For Donoghue, the work is not abstract. His grandfather lived with Parkinson's disease. The company's focus on early changes in REM and slow-wave sleep in Parkinson's patients is, in part, the attempt to give future patients what his grandfather never had: the chance to know, and to act, while there is still time.
Jake Donoghue was training in neurology and psychiatry when something struck him as fundamentally unfair. Cardiologists could monitor their patients' hearts from home, watching cardiac function unfold over weeks and months through simple, scalable devices. Neurologists had no equivalent. They relied on snapshots—a patient in a lab, wired to machines, trying to sleep in an unfamiliar place while their brain activity was recorded in bursts. It was invasive, expensive, and it captured almost nothing of how the brain actually behaved in the real world.
That frustration became the seed for Beacon Biosignals. Donoghue, an MIT PhD in neuroscience who trained under Earl K. Miller and completed clinical work at Massachusetts General Hospital and Boston Children's Hospital, partnered with Jarrett Revels, a research software engineer from MIT's Julia Lab, to build something different. In 2019, they launched a company built on a simple premise: if you could measure brain activity while people slept in their own beds, you could gather the kind of longitudinal, high-quality data that had never been possible before.
The device they created is a lightweight EEG headband—clinical-grade brain monitoring stripped of the laboratory. People wear it at home, night after night, collecting data while they sleep naturally. Machine-learning algorithms process the signals, extracting patterns that reveal how much time a patient spends in different sleep stages, how many micro-awakenings interrupt their rest, and crucially, what subtle shifts in sleep architecture might signal the early stages of disease. The FDA cleared the technology as a medical device, and pharmaceutical companies began using it immediately. Over forty clinical trials now rely on Beacon's platform to study treatments for major depressive disorder, schizophrenia, narcolepsy, Alzheimer's disease, and Parkinson's disease.
What makes this approach powerful is timing. Sleep, Donoghue explains, is when the brain's electrical activity becomes almost language-like—structured, high-amplitude, revealing. Changes in sleep patterns can emerge years before a patient experiences any symptoms of neurodegeneration. A person might come in for a routine sleep apnea screening, their brain activity recorded and stored. Years later, if they develop Parkinson's disease, that earlier baseline becomes a window into the disease's invisible progression. The data that was once merely diagnostic becomes prognostic—a way to detect disease before it announces itself.
Last year, Beacon acquired an at-home sleep apnea testing company serving more than 100,000 patients annually across the United States. That acquisition transformed the company's reach. Suddenly, Beacon wasn't just running clinical trials for pharmaceutical partners. It was building longitudinal brain health records for tens of thousands of people, creating what Donoghue calls a foundation model of the brain—a dataset comprehensive enough to reveal patterns no single study could uncover. In November, the company raised $97 million to accelerate this expansion.
Donoghue's vision is explicit: to move neurological disease from late recognition to early detection, from reactive treatment to preventive intervention. The brain, he notes, is an electrical organ that changes through synaptic plasticity. Track that electrical activity across many diseases, across many patients, over many years, and you begin to see the heterogeneity that static tests—genetic sequencing, brain imaging—can never fully capture. You discover disease subtypes that don't fit existing categories. You find the markers that precede symptoms by years. You create the possibility of catching disease before it takes hold.
For Donoghue, this work carries personal weight. His grandfather was affected by Parkinson's disease. The company's focus on detecting early changes in rapid-eye-movement and slow-wave sleep patterns in Parkinson's patients is not abstract. It is the attempt to give future patients something his grandfather did not have: the chance to know, and to act, before the disease becomes undeniable.
Citas Notables
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
What struck me most was the inability to measure brain function in the ways that cardiologists can longitudinally monitor cardiac function in patients from home.— Jake Donoghue
La Conversación del Hearth Otra perspectiva de la historia
Why sleep? Why not monitor the brain during waking hours, when people are actually doing things?
Sleep is when the brain's electrical signature becomes most legible. The activity is higher, more structured, almost like a language. During the day, the brain is noisy—responding to a thousand stimuli. At night, you get signal without that interference.
But if someone has Parkinson's disease, wouldn't you want to see how their brain behaves when they're moving, when they're struggling?
You would, and eventually you might. But the point is earlier detection. Changes in sleep architecture can show up years before motor symptoms appear. If you're trying to catch disease before it becomes visible, sleep is where you find it.
So you're building a kind of brain archive—collecting years of data from thousands of people, then looking backward when they get sick?
Exactly. A person comes in for a sleep apnea test. Their brain activity is recorded. Five years later, they develop Parkinson's. Now that old data becomes a map of the disease's invisible progression. You can see when it actually started.
That's only useful if you can do something about it once you detect it early.
True. That's the harder problem. But you can't intervene in a disease you don't know exists. The detection has to come first.