Cleveland Clinic Partners with Piramidal to Deploy AI Brain Monitoring in ICUs

Image may contain EEG Scan Accessories Adult Person EEG electrodes Brain scans Computer Hardware Electronics and Hardware

Background and Motivation

The Cleveland Clinic has teamed up with Piramidal, a startup founded in 2023 by neuroengineer Dimitris Fotis Sakellariou and product strategist Kris Pahuja, to create a large‑scale AI model for monitoring brain activity in intensive care units. Traditional ICU monitoring relies on periodic EEG reports generated every 12 or 24 hours, a process that can take two to four hours to review and is highly dependent on specialist expertise. As Imad Najm, director of the Epilepsy Center at the Cleveland Clinic’s Neurological Institute, explains, “This type of thing is time‑consistent. It is subjective, and it is experience- and expertise-dependent.” The new AI system seeks to replace this bottleneck with continuous, automated analysis.

The AI Model and Its Training Data

Piramidal’s model functions as a “foundation model” for the brain, analogous to large language models for text. It was trained on publicly available EEG datasets as well as proprietary recordings from the Cleveland Clinic and other partners, incorporating nearly a million hours of EEG monitoring data from “dozens of thousands” of patients, both neurologically healthy and unhealthy. This massive data volume is intended to capture the wide variability in brain‑wave patterns across individuals, enabling the model to generalize across diverse patient populations.

How the System Works

The AI continuously ingests raw EEG streams and uses learned patterns to identify signs of seizures, altered consciousness, or declining brain function within seconds. According to Piramidal’s chief product officer Kris Pahuja, “Our model plays that role of constantly monitoring patients in the ICU and letting the doctors know what’s happening with the patient and how their brain health is evolving in real time.” This rapid flagging is designed to give clinicians a chance to intervene sooner than the current manual review process allows.

Testing, Validation, and Deployment Plans

After initial development using retrospective patient data, the partnership intends to conduct a tightly controlled live‑patient trial in the next six to eight months, initially involving a limited number of ICU beds and physicians. The rollout will be gradual to manage false‑positive and false‑negative rates, which Najm describes as a “big problem that keeps us awake at night.” While Piramidal has not disclosed specific accuracy metrics, it claims the model has achieved “humanlike” performance when evaluated against a network of doctors and plans to publish detailed results later.

Broader Applications and Future Directions

Beyond ICU monitoring, Piramidal aims to adapt the foundation model for epilepsy management and sleep monitoring. The company’s founders also note that other firms, such as brain‑computer‑interface developer Synchron, are pursuing similar foundation‑model approaches to broaden the usability of neural‑signal technologies. Consumer‑oriented applications, like EEG earbuds that gauge emotional states, are also on the horizon.

Ethical and Societal Considerations

The expansion of AI‑driven brain monitoring raises significant questions about data privacy, consent, and the appropriate contexts for deployment. Caroline Montojo, president and CEO of the Dana Foundation, emphasizes that “Advancements like this one highlight the need for anticipatory ethical frameworks that support responsible development and use of these technologies.” She calls for early involvement of ethicists, social scientists, legal scholars, and patients in the design process.