Wearable devices using machine learning can accurately detect and rate Parkinson’s disease tremors as people go about their normal activities, according to a new study.
Study participants wore sensors on the wrist or ankle. Data was collected while they performed a variety of activities such as walking, resting, eating and getting dressed.
One of the two algorithms tested, called gradient tree boosting, showed high accuracy in estimating total tremor as well as resting tremor sub-score. It was also able to show decline in tremors after patients took their medication.
In most cases, results from the machine learning test matched results of the standard assessment currently used by neurologists. This test, the Unified Parkinson’s Disease Rating Scale, requires patients to perform certain activities during an office visit.
“A single, clinical examination in a doctor’s office often fails to capture a patient’s complete continuum of tremors in his or her routine daily life,” said senior author Behnaz Ghoraani, Ph.D., of Florida Atlantic University. “Wearable sensors, combined with machine-learning algorithms, can be used at home or elsewhere to estimate a patient’s severity rating of tremors based on the way that it manifests itself in movement patterns.”