Beacon Biosignals launches sleep brain-mapping AI platform

Beacon Biosignals unveiled a bedside sleep brain-mapping platform this week that records overnight electrical brain activity and other biosignals, then applies machine learning to extract candidate biomarkers for disease. The company, founded by Jake Donoghue PhD ’19 and former MIT researcher Jarrett Revels, says the system is designed to support diagnosis and treatment planning for sleep and neurological conditions.

The product packages sensors, signal processing, and AI models into a workflow intended for clinical and research settings. Beacon positions the platform as an alternative to single-night lab polysomnography and to short ambulatory snapshots, aiming to deliver higher-resolution maps of brain activity across full nights of sleep. Early demonstrations show the system capturing continuous electrophysiology and auxiliary signals that the company feeds into proprietary analytics.

How it works

The system combines bedside sensor hardware with automated pipelines for signal cleaning and feature extraction. Sensors record cortical electrical activity overnight alongside movement and other physiological measures. Raw recordings are preprocessed to remove common artifacts and to segment sleep stages. Machine learning models then analyze temporal patterns and correlations across channels to generate candidate biomarkers and visualizations for clinicians and researchers.

Beacon’s architecture separates data acquisition, automated analytics, and a user interface. Acquisition runs locally at the bedside before encrypted uploads to a secure cloud environment where models run at scale. Output products described by the company include nightly summary reports, time-aligned brain-mapping visualizations, and cohort-level analytics for research teams. The company says its models prioritize explainability by flagging salient features and epochs rather than returning black-box scores.

Clinical testing and next steps

Beacon was founded by Donoghue and Revels out of academic work at MIT and intends to move the platform through clinical validation. The company plans to run pilot studies to compare its overnight mapping outputs with existing diagnostic standards and to refine models on clinically labeled datasets. Beacon also aims to make its analytics interoperable with electronic health records and standard sleep-lab software, enabling clinicians to view mapped signals alongside routine clinical notes and test results.

On the regulatory front, Beacon has described a roadmap toward cleared clinical tools, which would require prospective validation and partnership with health systems or device manufacturers. The company is positioning its platform for both research groups that need long-duration sleep electrophysiology and clinics that want richer overnight diagnostic information without full in-lab stays.

Why it matters

Long-duration brain recordings during sleep could reveal transient or low-frequency phenomena that short tests miss, and automated analytics may surface repeatable patterns tied to disease. If validated, Beacon’s platform could expand access to richer sleep electrophysiology while shifting some diagnostic workflows out of specialized labs and into routine care settings. Clinicians, researchers, and patients with suspected sleep-related or neurological conditions are the most directly affected groups.