Smartwatch that can predict Parkinson’s disease

“The reason people with diabetes live shorter lives is because they are much more likely to develop heart problems,” says Vamil. “We hope that in the future, this data can be used to detect early signs to alert the patient and physician to the future risk of heart attacks and strokes.”

Predicting neurological problems

But many uses for smartwatches can go far beyond the heart. In July 2023, researchers from Cardiff University published a study that used data from more than 100,000 people who were given a smartwatch to wear for a week. The results showed that it was possible to identify people with signs of Parkinson’s disease seven years before their clinical diagnosis. This was done by detecting slight variations in their gait, measured by the watch’s motion sensors.

Cynthia Sandor, who led the study, believes these signs could be detected even earlier by combining motion data with other smartwatch measurements, such as sleep quality, which is known to be disrupted in people who develop the disease.

“In Parkinson’s disease, there is a long phase before diagnosis when signs such as subtle motor changes become apparent,” Sandor says. “We found that the most predictable sign was slower movement during light physical activity, too slight to be noticed by people themselves.”

Sandor believes this information could soon be used to recruit people into clinical trials. One theory for why effective treatments for Parkinson’s disease have proven so elusive is that patients are diagnosed at a stage when significant brain damage has already occurred, and it may be easier to slow or even reverse the disease at an earlier stage. “We hope that early screening tools based on smartwatch data will be able to identify people at an early stage, potentially allowing for successful trials of neuroprotective treatments,” she says.

There is also hope that smartwatches could one day help people living with chronic conditions such as epilepsy by giving them early warning signs that a seizure is about to occur. Falls and serious accidents as a result of seizures are known risk factors for people living with epilepsy.

“The uncertainty of when seizures may occur is one of the most difficult aspects of living with epilepsy,” says Eileen McGonigal from the Queensland Brain Institute. “However, forecasting seizures is still in the early stages.”

McGonigal wonders whether a special Empatica smartwatch prototype developed for research could help predict seizures. As part of her current research project, she applies artificial intelligence algorithms to a combination of data streams. These include heart rate variability, skin temperature, body movement patterns, and changes in skin conductance due to sweating, which reflect changes in the body’s autonomic nervous system. Each of these parameters can be measured using a watch.

“We aim to analyze patterns in the hours leading up to attacks,” says McGonigal. “Ideally, epilepsy researchers and clinicians would like patients to be able to predict when seizures are more likely, which could allow treatment to be tailored, including variable medication doses and adaptations of daily activities, to reduce the risk of falls and seizure-related injuries. “, she says.

But while there is considerable enthusiasm about what the combination of powerful artificial intelligence algorithms and increasingly precise measurements from wearable devices can achieve, some doctors are also wary about the possibility of false positives. There are concerns that excessive use of smartwatches could lead to increased anxiety among patients, as well as testing the resources of already stretched healthcare systems.