Completed Brain & Nervous System Pregnancy, Children & Inherited Conditions

Multicentre Clinical evaluation of a neonatal seizure detection algorithm.

Funder Wellcome Charity-funded
Recipient organisation University College Cork
Funding £2.6M
Period Nov 2012 — Jun 2018

In plain English

AI plain-English summary

Every year, thousands of newborn babies in intensive care have seizures that go undetected by the naked eye. These seizures can damage a developing brain, but the standard monitoring tool—the EEG—requires a specialist to interpret, and those specialists are rarely available around the clock. The researchers have built a computer algorithm that can spot neonatal seizures automatically from EEG data, and they have already tested it on recordings from babies in Cork and London. Now they need to prove it works reliably in the real world. This project will run a formal clinical trial across multiple hospitals, using different brands of EEG machines, to measure how accurately the algorithm detects seizures. The team will also interview doctors, nurses, and parents to design a screen display that is clear and easy to use in a busy neonatal unit. If the trial succeeds, the algorithm could be licensed to any medical device company and embedded directly into cotside monitors. That would give every intensive care nursery—not just those with a neurologist on call—a tool to catch seizures early, guide treatment, and improve long-term outcomes for vulnerable newborns.

View original technical description
We have developed an automated seizure detection algorithm for neonates, which has been trained and tested on data acquired from newborns in Cork and London. To ensure that this development can benefit all babies it is essential that we conduct multicentre clinical trials, conduct these trial in a regulatory compliant manner so that it is possible for any company to obtain a product licence at the end of this project. We need to test our algorithm on data from different centres using different EEG machines, and in order to do this we have recruited a number of international collaborators. A key part of this project will be a detailed clinical evaluation study, designed as a clinical trial of the algorithm and powered to test its performance in seizure detection, but which will also evaluate the views of users and parents. We will use the experience with end users to establish the best possible interface (taking into account ease of use, interpretation and presentation) for the seizure detection algorithm in the NICU. Our goal is to have a commercially viable, clinically tested and validated product that can be embedded into cotside monitoring systems worldwide independent of the manufacturer, in order to detect seizures in newborns and monitor treatments. Prompt and accurate treatment of seizures will facilitate neuroprotective intensive care and improve long term outcomes.

View the original record at the funder ↗

Researchers

Geraldine Boylan (EPMC Awardee)Liam Marnane (EPMC Awardee)

Related Research

Grants with similar aims, by meaning.

Real World Testing and Cost-effectiveness Analysis of Subcutaneous EEG (REAL-ASE)
BabEEG: Implementing a dry, easy-setup EEG in neonatal units
EAGLET: EEG vs aEEG to improve the diagnosis of neonataL seizures and Epilepsy – a randomised Trial
Bridging the gap: Accessible Electroencephalography for Epilepsy diagnosis and management
Development of a cot-side optical biomarker of brain tissue health following neonatal hypoxic-ischaemic brain injury.

Original classification

Strategic Translation Award

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.