A test that uses artificial intelligence to detect early Parkinson’s disease more accurately diagnoses the condition than non-specialist clinicians, a new study says.
Charles Sturt University’s Warwick Adams, a School of Computing and Mathematics PhD student, said his research found a method to analyse the patterns in a person’s finger movement as they type on a computer.
Adams said there are no definitive tests for Parkinson’s disease by non-specialist clinicians, especially in the early disease stages where the symptoms may be subtle and poorly characterised, and added that, until now, diagnosis has relied on observation of a person’s movement where the initial signs can be quite subtle.
In a paper on the study, he said this results in a high misdiagnosis rate and people can have the disease for many years before diagnosis.
“The significance of this new technique is that it’s not only much more accurate, but it can be used in a home environment and does not require supervision by a medical specialist,” he said.
Adams said the test has a 97 per cent accuracy rate, significantly outperforming general practitioners.
Those taking part need to type fewer than 400 words and the technique does not require any specialised equipment or attachments.
Adams looked at keystroke timing information from 103 participants — including 32 with mild Parkinson’s disease severity and those who did not have a diagnosis — as they typed on a computer keyboard over an extended period.
He found that Parkinson’s disease affects various characteristics of hand and finger movement and that these can be detected. He then used a new method to classify participants’ disease status by using a combination of many keystroke features analysed by an ensemble of machine learning classification models.
Adams said his invention will be developed into a full diagnostic suite that can be accessed via the web.
He added that, once developed, it has enormous world-wide commercial potential.
“The benefits will be an earlier detection of Parkinson’s, as well as the ability to monitor the effects of medication and the progression of the disease over time.”Do you have an idea for a story?
Email [email protected]