Headband Delivers Clinical-Grade Sleep, Brain Health Intelligence

Muse has launched Muse S Athena, the first wearable to integrate clinical-grade EEG (electroencephalogram), fNIRS, and SpO₂ in a lightweight, consumer-friendly form. This development could help Muse transform its wearable into a scalable, artificial intelligence (AI)-powered platform for brain-based diagnostics.

By combining real-time biosignal data with foundation models inspired by the architecture behind GPTs, Athena delivers expert-level sleep staging and insights from home. This enables researchers, clinicians, and consumers to access precise, contextual, and longitudinal insights that were previously only available in a lab.

“We’re seeing powerful evidence that foundation models can transform how we interpret brain data,” Muse Co-Founder/Chief Innovation Officer Chris Aimone said. “What GPTs did for language, this architecture is beginning to do for neural signals—revealing patterns that are both highly individualized and deeply generalizable.”

Through its latest device, Muse S Athena, high-performance SpO₂ monitoring adds a vital respiratory layer to Muse’s neurotechnology platform. Unlike traditional fingertip sensors, the forehead placement enables faster response times and more stable readings throughout the night—especially important for detecting rapid changes in oxygen linked to sleep-disordered breathing.

This feature is now rolling out for research partners and, when paired with EEG and fNIRS, gives researchers unprecedented visibility into the brain, circulatory, and respiratory systems—all captured in real time, from a single wearable, at scale.

Muse Trains AI to Understand the Brain’s Language

Muse claims to have built the world’s largest longitudinal EEG dataset—over 1 billion minutes of brain activity from more than 16 million sessions—spanning meditation, sleep, and cognitive use cases. This data underpins the Muse Foundational Brain Model (FBM): a deep-learning architecture akin to a large language model, but trained to decode the brain.

The FBM–a transformer-based neural network model like those that underlie large language models like ChatGPT–is trained on vast, contextual, multimodal datasets. Translation layers allow the FBM to interface with the latest generative AI intelligences to augment their capabilities with neural insights. This integrated AI stack enables:

• Clustering individuals based on neural patterns that may reflect cognitive resilience or early decline
• Detection of subtle anomalies in recovery that may indicate emerging neurological issues
• Interpreting real-time brain states using multimodal data streams including EEG, SpO₂, and fNIRS

Because the FBM has been trained on such a wide variety of real-world brain sessions, the FBM can be fine-tuned for new applications—accelerating development in areas like cognitive decline, mental health, and digital therapeutics without costly new datasets, according to Muse.

“We’re no longer looking at EEG as static metrics,” Muse Chief Marketing Officer Nadia Kumentas stated. “To us, brain activity is a living, contextual language—one that foundation models can learn, interpret, and translate into actionable insight.”

In a major clinical milestone, Muse has demonstrated sleep staging accuracy on par with expert sleep scorers using PSG-labeled datasets, validated across both retrospective and real-time testing. Going beyond human capabilities, Muse’s proprietary mEEG signal and Foundational Brain Model now deliver not only lab-grade scoring precision in a fully wearable form factor, but also deeper insights—such as staging macro and microarchitecture of sleep even in the presence of environmental noise.

“What we’ve achieved with Athena confirms that Muse is more than a consumer device—it’s a clinical-grade platform ready to support real-world research, decentralized trials, and the future of at-home diagnostics,” Muse CEO Jean-Michel Fournier said.

This level of accuracy enables frictionless, longitudinal sleep scoring in home environments—reducing cost and complexity in clinical trials while providing reliable, precision insights to end-users.

Muse now makes it possible to measure blood oxygen levels (SpO₂) from the forehead—bringing a critical tool for sleep apnea and respiratory research out of the lab and into the home.

Already in use across studies in cognitive decline, long COVID-19, stroke recovery, and digital therapeutics, Muse’s infrastructure supports rapid experimentation, population-level insights, and plug-in potential for pharmaceuticals, research, and health tech.

Muse is pioneering brain health through its mEEG platform, built on advanced AI algorithms. Muse spearheads decentralized research initiatives focused on enhancing mental health, optimizing sleep, and advancing cognitive performance through innovative neurotechnology. Muse’s AI-driven neurotech tools empower researchers, developers, and consumers worldwide, and are underpinned by more than 200 third-party-led research studies from renowned institutions including the Mayo Clinic, MIT, and Harvard. Muse is headquartered in Toronto.