Digital biomarkers for Parkinson's Disease
Motor symptoms that produce mechanical signals, fluctuate with every dose, and are still graded by eye during a single appointment.
.png)
Parkinson's disease is the condition digital measurement was arguably built for. Its cardinal symptoms are motor, which means they produce mechanical signals a sensor can detect. They fluctuate across the day in response to medication, which means a single clinic assessment samples an arbitrary point in that cycle. And the established rating scales depend on trained observation during an appointment, which is exactly the setting least likely to capture what the rest of the day looks like.
The measures used in Parkinson's research therefore fall into two groups. The first quantifies specific symptoms directly: tremor, slowness of movement, dyskinesia and freezing of gait. The second describes the functional consequences: how fast the person walks in daily life, how their walking is distributed across the day, how often they rise from a chair, and how they sleep.
The second group is often more informative than it first appears. Gait fragmentation and reduced sit to stand activity change while average measures still look normal, and they describe the practical loss of independence that patients and carers report as the thing that matters most.
Digital biomarkers used in Parkinson's disease research
How these measures are used
The central research application is quantifying on and off time, meaning how much of the day a person spends with adequate symptom control. That is the outcome medication titration actually targets, it is currently estimated from patient diaries that are known to be unreliable, and it is measurable continuously with wrist worn sensors. Dedicated cleared devices exist specifically for this purpose.
Beyond symptom control, digital measures are used to detect progression at higher resolution than annual rating scale assessments allow, and to test whether an intervention improves function in daily life rather than only in a testing room.
These measures are reported alongside the standard clinical rating scale rather than instead of it, since regulators and clinicians read the scale and it encodes disease specific meaning that a sensor output does not yet carry. In practice, protocols pair continuous monitoring with periodic in person assessment and treat disagreement between them as informative rather than as a problem to resolve.
What the evidence supports today
Parkinson's disease has one of the deeper digital measurement evidence bases, with several strands worth separating. Continuous monitoring of motor fluctuation has dedicated hardware that has passed regulatory review, and a literature comparing its output against clinical assessment and patient diaries. Free living gait analysis has been validated against laboratory reference systems, including through the Mobilise-D consortium, which included Parkinson's disease among the conditions in its technical validation programme.
Freezing of gait remains harder. It is episodic, brief and context dependent, and detection performance in the home falls short of what laboratory studies suggest. Automated detection is usable for research description but is not yet a dependable clinical alarm.
The unresolved question across the condition is clinical meaningfulness. Sensors detect change smaller than rating scales can, but how much change matters to a person is still being established for most of these measures, and a statistically detectable difference is not automatically a difference worth treating.
Common questions
Which digital biomarkers are most used in Parkinson's research?
Continuous motor symptom monitoring covering tremor, slowness and dyskinesia, and free living gait measures including walking speed, bout structure and sit to stand transitions. Sleep measures are also common, since sleep disturbance is a frequent and burdensome non motor feature of the condition.
Can wearables replace the clinical rating scale?
Not currently, and that is unlikely to change soon. The scale encodes clinical meaning and regulatory familiarity that a sensor output does not carry on its own. Digital measures are reported alongside it, contributing frequency and objectivity, particularly for symptoms that fluctuate between appointments.
What is on and off time, and why measure it digitally?
On time is the period when medication is controlling symptoms adequately; off time is when it is not. Treatment adjustment targets that balance directly. It is conventionally estimated from patient completed diaries, which are burdensome and imprecise, and continuous wrist monitoring offers an objective alternative.
Is freezing of gait detectable by a wearable?
Partially. Episodes are short, irregular and strongly context dependent, and algorithms trained in controlled conditions perform worse in the home. Detection is useful for characterising how often freezing occurs across a monitoring period, but it is not reliable enough to function as a real time alert.
Run a study on these measures
WeGuide captures wearable data and patient reported outcomes in one platform, from screening through to analysis.
