The incredible cap that warns people with epilepsy when seizures will strike
When Zoe Cartledge has an epileptic seizure, she doesn't fall to the ground or convulse. Instead, her body freezes up, her eyes glaze over and she begins to overheat.
She has had numerous seizures on Sydney's trains and buses, sometimes finding herself as far as two hours away from her destination once she regains control of her mind and body.
"Before I started taking medications, I had five- to 30-minute seizures about 10 to 30 times a day. Now I have no more than two a week," said Ms Cartledge, 27, from Blacktown.
"Sometimes I can sense it's about to happen and I try to talk myself out of it, but this backfires because I make it a lot worse, I start to overheat and my anxiety skyrockets."
Ms Cartledge is among one-third of people with epilepsy who aren't able to fully control and prevent seizures with medication, meaning they are at constant risk of injury and even death.
A team led by Dr Omid Kavehei of Sydney University's Faculty of Engineering and IT has developed a non-surgical device that can predict, with 80 per cent accuracy, whether a person will have a seizure in the next 30 minutes.
In a paper published in the latest Neural Networks, the researchers said they used advanced artificial intelligence and machine learning to develop a device that could read a patient’s brain signals from a cap on their head.
The device, most likely to be worn on the hip, can trigger an alarm five to 30 minutes before a seizure, giving patients time to find a safe place and reduce stress.
"Artificial intelligence is good in detecting hidden patterns in data. There are patterns that we may not be able to detect with conventional techniques," Dr Kavehei said.
"People think that seizures are random, but there must be some pattern in the immediate time before the seizure takes place."
About 65 million people around the world have epilepsy, a brain disorder, including 250,000 Australians.
At present, seizure detection devices on the market require surgery, which carries the risk of infection and serious complications.
"The worst consequence of that is even if the surgery is successful, there is no guarantee that your seizures have been treated," he said.
"Fifty per cent who have had deep brain or surface brain stimulation say they're not able to suppress seizure onsets."
Dr Kavehei said the final product would be affordable and possibly cost less than $100.
He said remarkable advances in artificial intelligence and nano-electronics had made the project possible.
"Critics might say the technology isn't mature enough, it's noisy, but just look at the past 20 years of electronics, we've gone from big medical sensors to really tiny, wearable sensors, that implants aren't even needed," he said.
The team's next step is to apply the neural networks across much larger data sets of seizure information, improving sensitivity and therefore accuracy.
They are also working with Westmead Hospital to develop a "miniaturised" sensory network.
"We are hoping to put together these two, quality data and miniaturisation, with the hospital to develop a full solution," he said.
In the future, the system might be able to send alerts to the patient's mobile phone.
Carol Ireland, chief executive of Epilepsy Action Australia, said the unpredictable nature of epilepsy caused stress and anxiety for those with the disease and their family and friends.
“Any progress towards reliable seizure prediction will significantly impact the quality of life and freedom of choice for people living with epilepsy," she said.
Ms Cartledge said the seizure prediction device would change her quality of life.
"It means I will no longer end up in the middle of nowhere on a bus or a train," she said.
"I'll be able to swim in a lake and not have to stay onshore, very simple things I would love to do rather than spending time in fear.''