Wearables track Parkinson’s better

C&I Issue 11, 2023

Read time: 2 mins

By Bárbara Pinho | 8 November 2023

Smartwatches could help to detect Parkinson's early.

Scientists have shown that a combination of digital wearables and machine learning can detect subtle movement changes in Parkinson’s patients better than standard rating scales. The findings could help clinicians better track the progression of the disease.

Assessing changes in patients’ symptoms, such as their gait, cognition and general muscle movements over time helps clinicians understand how the disease progresses. One way of doing this is through clinical rating scales — clinical questionnaires in which clinicians ask patients to perform tasks and rate the movement involved.

Despite being used around the world, these assessments can be subjective. ‘I can be trained, you could be trained, and we could be having variations between the two of us, but also within ourselves,’ says Chrystalina Antoniades, co-author of the study from the University of Oxford, UK.

To see if sensors could help better track the onset of symptoms, Antoniades’ team used a combination of wearable sensor data and machine learning algorithms to estimate each patient’s clinical rating scale.

They attached wearables to 74 patients who visited the lab seven times. By analysing more than 100 characteristics recorded by the sensors, the models were able to pick up signals of disease progression much earlier than typical rating scales. ‘It doesn’t mean that the [standard] clinical rating scale is not good. It means that it’s not built to operate as early as this,’ explains Antoniades.

Alberto Espay, from the University of Cincinnati, US, who wasn’t involved in the research, says it’s an interesting study, but believes that we should look at the individuality of each Parkinson’s patient: ‘We often say that no two patients with Parkinson’s are alike, but do not act as if we mean it when it comes to our research efforts.

‘We insist on trying to find a grand model of Parkinson’s progression, now assisted by AI, missing the unique signatures at the individual level that are needed to launch precision medicine for patients with Parkinson’s,’ he added.

Become an SCI Member to receive benefits and discounts

Join SCI