INFANT’s Dr John O’Toole will lead a team of international researchers to accelerate the development of AI Technologies that detect brain injuries in infants.
Working alongside a team of scientists, clinicians and technical experts from 14 different European countries, Dr John O’Toole aims to build capacity and strengthen cooperation among international research groups, with the goal of developing algorithms that will minimise the risk of babies developing catastrophic life-long neonatal brain injuries.
Insufficient oxygen around the time of birth can cause brain injury. For babies born prematurely, the heart and lungs may struggle to adapt to the new environment which can lead to brain injury too. Brain monitoring of a tiny infant in an intensive care unit is challenging.
It can be difficult and slow to interpret the complex brain-wave patterns. AI systems are a perfect fit to this problem, as they can be designed to automatically recognise signs of brain injury.
Funded by the European Cooperation in Science and Technology, the researchers involved in the AI-4-NICU project plan to build on existing cot-side technologies, such as devices that measure brain waves, by including AI algorithms to detect markers of brain injury.
This, Dr O’Toole anticipates, will lead to the development of decision-support tools that will help clinicians in neonatal intensive care units to quickly identify potential brain injuries that can result in death, cerebral palsy, or delayed development.
Reading and interpreting the brain-wave signals is a notoriously difficult task which requires highly specialised expertise. AI systems can be designed to mimic the human expert, by shifting through enormous amounts of data to automatically find signs of brain injury.
These AI systems, unlike the human expert, can then run around the clock for all at-risk infants to provide a continuous assessment of brain health.
To develop the device, Dr O’Toole and his team will first develop the tools necessary to acquire, pool, share, and manage neuro-physiological data sets.
They will then create a framework to develop, test, and compare algorithms that they hope will act as decision-support tools in neonatal intensive care units.