Digital biomarkers in neurology
Neurological symptoms fluctuate by the hour, which is exactly what an assessment scheduled once a year cannot see.
.png)
Neurological conditions produce some of the clearest arguments for continuous measurement. Symptoms fluctuate over hours, respond to medication timing, and are frequently absent during the short window of a clinic appointment. A rating scale administered at a scheduled visit captures a sample of the disease, and often an unrepresentative one.
This category collects the digital measures that address that problem: tremor amplitude and duration, movement slowness, freezing and gait disturbance, seizure detection, and the digital cognitive assessments that measure attention, memory and processing speed through repeatable tasks rather than clinician judgement.
The domain also carries a specific risk worth naming. Several of these measures share vocabulary with consumer applications, particularly in cognition, where reaction time tests are widely marketed as brain training. The clinical versions are different instruments with different validation requirements, and conflating them is the most common error in this area. What makes a digital cognitive measure clinically useful is not that it is quick or engaging, but that it has been shown to detect the specific impairment a condition produces and to change when that impairment changes.
Neurology measures in this library
How these measures are used
The primary research use is capturing fluctuation. In movement disorders, continuous recording quantifies how much of a day a person spends with adequate symptom control, which is the outcome medication adjustment actually targets and which no single clinic assessment can estimate. In epilepsy, wearable detection addresses the well documented unreliability of seizure diaries, since many events are unwitnessed or unremembered.
Digital cognitive batteries are used for repeated assessment where conventional testing is limited by practice effects and by the burden of clinic attendance. Their value is in frequency: an assessment that can be administered weekly at home produces a trajectory rather than two distant points.
These measures are almost always reported alongside the established clinical rating scale for the condition, both because regulators expect the comparison and because the scales encode clinical meaning that a sensor output does not yet carry on its own. The digital measure adds resolution and frequency; the scale supplies interpretation.
What the evidence supports today
Evidence here is genuinely mixed and depends heavily on which measure is in question. Wearable detection of generalised convulsive seizures has reached regulatory clearance in several markets and is the most mature application in the domain. Continuous monitoring of movement symptoms in Parkinson's disease has dedicated cleared hardware and a substantial literature.
Other measures are earlier. Automatic detection of non convulsive seizure types remains difficult. Digital cognitive assessment has strong technical validation for many individual tests but a weaker record of demonstrating that change on the test corresponds to change that matters to the person, which is the step that separates a valid measurement from a useful endpoint.
Across the domain, the consistent limitation is population specificity. An algorithm trained on one condition, one device placement or one severity range often performs poorly outside it, and published accuracy figures frequently come from tightly controlled cohorts that do not resemble the intended clinical population.
Common questions
Why measure neurological symptoms continuously rather than at a clinic visit?
Because many of them fluctuate. Tremor, slowness and freezing vary with medication timing, fatigue and stress, so a clinic assessment captures whatever state the person happens to be in at that hour. Continuous recording estimates how the day as a whole was spent, which is closer to the outcome treatment is aiming at.
Can a wearable detect seizures reliably?
For generalised convulsive seizures, detection has reached regulatory clearance and performs well, though false alarms remain a practical burden. Other seizure types, particularly focal and non convulsive events, are much harder to detect from wrist worn sensors and should not be assumed to be covered by a device cleared for convulsive seizures.
Are digital cognitive tests the same as brain training apps?
No. They can look similar, but a clinical digital cognitive assessment is developed and validated to detect specific impairment and to be repeated without excessive practice effects. Consumer brain training products are not validated for that purpose, and treating results from one as equivalent to the other is not defensible.
Do these measures replace clinical rating scales?
Not currently. They are reported alongside the established scale for the condition, which supplies clinical interpretation and regulatory familiarity. The digital measure contributes frequency and objectivity, particularly for symptoms that fluctuate or that occur outside observed settings.
Run a study on these measures
WeGuide captures wearable data and patient reported outcomes in one platform, from screening through to analysis.
