Wearable Device May Detect Parkinson Disease 7 Years Before Symptom Onset

Reduced daytime acceleration over 1 week was associated with a clinical diagnosis of Parkinson disease up to 7 years later.

A wearable device may be able to accurately identify individuals at an elevated risk of developing Parkinson disease (PD), according to study findings published in the journal Nature Medicine.

To date, treatment for PD consists of symptom management, with no current disease-modifying treatments available. Furthermore, at the onset of clinical manifestations, most patients will have undergone significant neuronal degeneration of dopaminergic tracts. There remains a need for reliable biomarkers that may detect the early pathologic changes that are associated with PD.

For the study, researchers aimed to investigate the effectiveness of digital accelerometer data as a prodromal marker for PD.

The researchers utilized a prospective, population-based cohort, termed the UK Biobank (UKBB). Accelerometry data was collected for 103,712 individuals aged 40-69, from 2013-2016. Comparison data was compiled, exploring whether accelerometers data can accurately serve as a prodromal marker for PD, by comparison of those diagnosed with PD to unaffected control individuals. Data was also compared with other current modalities, including genetic information, blood biochemistry, lifestyle, and prodromal symptoms.

Our results suggest that accelerometry collected with wearable devices in the general population could be used to identify those at elevated risk for PD on an unprecedented scale …

During the 2-year time of collection, 273 participants were diagnosed with PD; an additional 196 individuals were diagnosed ≥2 years after the data collection, forming a prodromal group for comparison. When comparing data between the prodromal and diagnosed PD cases to an age- and sex-matched unaffected control individuals (1:1), there was a significantly reduced daytime acceleration profile up to 7 years before diagnosis.

Linear regression models, when adjusted for age, sex and body mass index (BMI) for residual average acceleration, found a significant reduction of acceleration in diagnostic PD and prodromal PD when compared with unaffected control individuals. Additionally, no other investigated neurodegenerative disorders (Alzheimer disease, dystonia, all-cause dementia) were found to have a reduction in acceleration before a diagnosis, as was observed in PD. Depression, however, was the only other disorder that the researchers found to have a reduction in acceleration after diagnosis.

Prodromal and diagnosed PD could also be identified from the general population when using average acceleration with area under the precision recall curve (AUPRC) with values of 0.05±0.04 (prevalence =0.0034) and 0.06±0.05 (prevalence =0.0046), respectively. Comparing previous modalities used to identify prodromal PD (genetics, lifestyle, blood biochemistry), accelerometry data was also shown to have a better predictive value in identifying a future diagnosis of PD.

Of note, time dependent area under the receiver operator curve (AUROC) data identified accelerometry to be able to predict the probability of not receiving a diagnosis of PD better than any other single modality previously used; the findings highlight the ability to predict when a diagnosis of PD can be expected.

Study limitations include accelerometry data collection being restricted to a 7-day period, limiting analysis timeframe. Additional modalities with high predictive power, such as dopamine transporter imaging and motor examinations were also not included in the study.

The researchers concluded, “Our results suggest that accelerometry collected with wearable devices in the general population could be used to identify those at elevated risk for PD on an unprecedented scale and, importantly, individuals who will likely convert within the next few years can be included in studies for neuroprotective treatments.”


Schalkamp AK, Peall KJ, Harrison NA, Sandor C. Wearable movement-tracking data identify Parkinson’s disease years before clinical diagnosis. Nat Med. Published online July 3, 2023. doi:10.1038/s41591-023-02440-2