Researchers have managed to develop a revolutionary AI method able to identify Parkinson’s disease. Here’s how it works!
According to a new study published in Nature Medicine, the system measures the patient’s breathing patterns while they’re asleep while also measuring and recording the disease’s severity and progression.
After the data is gathered, it gets processed through a so-called neural network – a computer system made to act like the human brain.
The team who worked on this study points out that identifying early signs of Parkinson’s disease is not easy since many symptoms appear only years after the illness had already started taking over the body.
With that being said, this is not the first time that researchers are looking into methods of detecting the disease in time, before it can develop.
They’ve tried brain imaging and brain fluid examinations before but such methods not only require special medical centers but they are also quite expensive.
However, this new AI system is much more accessible and simple, all the while being able to efficiently detect the illness early on in its development.
This is great news since the World Health Organization says Parkinson’s disease is the fastest growing neurological condition on the planet!
In fact, the number of cases worldwide has doubled in the last 25 years with WHO estimating that there are over 8.5 million people suffering from Parkinson’s all over the world.
Hopefully, the AI method developed by scientists from the Massachusetts Institute of Technology (MIT) and other health and educational organizations, will help with identifying the disease.
As for how it works, professor of electrical engineering and computer science at MIT, Dina Katabi, explains that being able to examine breathing patterns is really important.
“A relationship between Parkinson’s and breathing was noted as early as 1817, in the work of Dr. James Parkinson,” Katabi states, adding that this was what led the team to “consider the potential of detecting the disease from one’s breathing without looking at movements.”
After all, past studies have also shown that breathing patterns linked to this disease can show up even years before all other symptoms!
This new AI algorithm has already been tested on 7,687 people 757 of them being Parkinson’s patients as well.
The system was able to detect Parkinson’s disease with up to 86% accuracy after only one night of gathering sleep data.
After 12 nights monitored, however, the accuracy went up to 95%!
