Speaking at the Mayo AI conference is Prof Geraldine Boylan, director of INFANT and she spoke with the conference chair and ITN journalist prior to the event.
A.I. is needed for Medical Health
Prof Geraldine Boylan is director at the INFANT Research Centre at University College Cork. INFANT unit conducts research focusing on improving outcomes for women and children. Seizures in babies are one research area at INFANT. They are very difficult to detect, as there may be no obvious outward signs, making intervention and treatment very difficult.
The only accurate tool for diagnosing seizures is Electroencephalographic (EEG) monitoring – a measure of electrical brain activity but this is challenging to interpret without specific expertise.
But the INFANT team has developed a tool that uses artificial intelligence to read the electrical impulses and warn staff when intervention is required.
“I am not an expert in A.I.,” confesses Prof Boylan. “But I know some areas where it is really needed in healthcare, particularly the neonatal intensive care unit (NICU) and A.I. is a huge part of our research and development here at INFANT.”
Prof Boylan leads a team of engineers and data scientists that develop software to monitor the data extracted in the NICU.
“A.I. is helping us predict and detect problems in newborn babies. It is crucial for babies as they cannot tell us what is wrong, yet we often need to act very quickly.
“We monitor many data feeds – blood pressure, heart data and of course brain activity.” Prof Boylan is a neurophysiologist and monitoring electrical brain activity is core to her research.
“As a neurophysiologist, I can monitor brain activity and look for specific patterns that may signal a problem. This is fine but babies are born day and night and won’t wait for office hours. If a neurophysiologist is not available in the unit, then who would track the patterns?
“This is where A.I. comes in and is of superb assistance to the medical team. A.I. is very good at objectively and consistently viewing data and looking for patterns quickly. A.I. does not get cranky or have to sleep.”
Prof Boylan also dismisses some of the perceived threats of A.I. “There are fears that it will take over or displace jobs – but in neonatal care, that is not the case as it does something that humans can’t do and overall is providing vital support for clinical decision making, not a replacement for the medical team.
“I personally love what A.I. is doing to help newborn infants and there is just so much more to do.”
Prof Boylan will be speaking at the A.I. conference in GMIT on November 20.
Recent news from INFANT
INFANT RESEARCH CENTRE SECURES FUNDING OF €570,000 TO DEVELOP SMART, NEWBORN BRAIN MONITORING SYSTEM
Putting machine learning at the heart of medical research
Machine learning and artificial intelligence aren’t just the latest tech fads – they are already showing their potential in healthcare.
With the latest imaging techniques, more sophisticated probes and accurate tests, doctors and
researchers can gather an almost overwhelming amount of data about how our bodies work. But this vast amount of data can take years to analyse.
Luckily, computer systems can be trained to analyse this data and, ultimately, teach themselves how to piece this information together to make predictions that can improve healthcare.
As Wellcome launches a new funding scheme to support researchers to transform their great ideas into digital healthcare innovations that could have a significant impact on human health.
To highlight the difference this research could make, two Wellcome funded projects are training
computers to improve how we treat brain conditions – brain injuries in new born babies and strokes
How can you ask a baby if it’s not well?
Time is critical for doctors treating new born babies with suspected brain injuries, the sooner that treatments like whole body cooling can be used the better the outcomes are likely to be.
But doctors rely on the results of EEG brain scans to give them crucial information about whether a new born baby has a brain injury. These scans are so detailed though that they can only be interpreted by an expert, and there aren’t enough experts to be at every cot side analysing the data in real time.
To tackle this, Wellcome funded researchers at the University of Cork are developing the first ‘smart’ system that will recognise patterns in electrical brain activity in babies with brain injuries and can identify those that need treatment.
To develop their system Professor Geraldine Boylan and her team are training a computer to learn the different EEG patterns and how they relate to the extent of the brain injury. If the computer can identify the warning signs, it could help doctors give treatments as soon as possible, helping more babies survive and cutting the risk of permanent disabilities such as epilepsy, cerebral palsy or learning difficulties.
The potential for digital technology in healthcare is vast, helping to free up the expertise of doctors to focus on patient care. As Professor Boylan says, “Machine learning has given us a new tool to improve clinical decision making. We already had the data from scans like EEGs, but we couldn’t quickly and accurately extract the information that we needed. These new technologies can transform how we care for our patients.”
The Wellcome Innovator awards will support teams who have innovative ideas that can challenge the status quo and address an unmet healthcare need or challenge. For more details of the scheme visit: https://wellcome.ac.uk/funding/innovator-awards-digital-technologies.