A model of artificial intelligence for the brain comes to ICU

Cleveland Clinic To create a large -scale artificial intelligence model used to monitor the health of patients in intensive care sectors, it works with a San Francisco -based startup.

Instead of teaching on the text, this system is based on electronsfall data (EEG) data collected through the electrodes placed on the scalp and then read by a computer on a series of wave lines. EEG records the electrical activity of the brain and a change in this activity can indicate a problem. In an ICU adjustment, doctors scan EEG data following evidence of seizures, altered awareness, or reduced brain function.

Currently, doctors rely on constant EEG supervision to diagnose abnormal brain activity in an ICU patient, but they cannot control any patient in real time. Instead, EEG reports are typically produced every 12 or 24 hours and then analyzed to determine if a patient is experiencing a neurological problem. It can take two to four hours to manually check the brain wave data.

“These kinds of things are time -consuming. It is mental and dependent on experience and expertise,” says Amad Najm, a neuroscientist and director of the Center for Epilepsy at the Cleveland Clinic Institute.

The system that is developing the Cleveland and Dyramidal Clinic is designed to interpret the continuous EEG data and flag malformations in a matter of seconds so that physicians can intervene sooner.

“Our model considers the role of continuous monitoring of patients in the intensive care department in the intensive care department and to doctors what happens to the patient and how their brain health is evolving in real time,” says Chris Pahuja, a senior product manager.

Pahuja and CEO of Dimitris Fotis Sakellariou Piramidal in 2023 aimed at building a brain -based model – an artificial intelligence system that can read and interpret neural signals widely among different individuals. Earlier, Sakellariou spent 15 years as an neurological engineer and artificial intelligence scientist conducting EEG research. Pahuja worked on product strategy in Google and Spotify. Launched, supported by Y Combinator, raised $ 6 million last year.

The company has developed its ICU brain model using the EEG data collection available as well as dedicated EEG data from the Cleveland Clinic and other partnerships. Sakellariou says the model covers almost a million hours of EEG monitoring data from “tens of thousands” patients, nervous and unhealthy. Patterns of brain activity are very variable from person to person, so building a brain -based model for recording patterns and common features requires large amounts of data.

“The beauty of a basic model is the same way that the chatter can generalize the text, it can adapt to your tone, it can be compatible with your writing method. Our model is capable of compatibility with different people’s brains,” says Sackerio.

Currently, the Cleveland Clinic and the Pyramidal team are using retrospective hospital data to accurately adjust the model. In the next six to eight months, they intend to test the model in a fully controlled ICU environment with live patient data and a limited number of beds and doctors. From there, they aim to slowly rotate the software into the entire ICU. Finally, this software allows the hospital system to control hundreds of patients at the same time.