Smartwatch data help detect Huntington’s disease progression
Study finds AI tool consistently registers severity of involuntary movements
Written by |
Motion data collected by a smartwatch may offer a more objective way to monitor worsening involuntary movements in people with Huntington’s disease than clinic visits alone, a study found.
Researchers in the U.K. and Switzerland developed a digital tool to estimate the severity and progression of chorea, a hallmark Huntington’s symptom characterized by involuntary, jerky movements. They found that their Digital Passive-monitoring Chorea Score (DCPS), which uses wrist-movement data collected from a smartwatch to estimate upper-limb chorea severity, measured chorea more consistently than standard clinic assessments.
“High-resolution accelerometer data collected passively during daily life can improve the measurement of chorea” in Huntington’s disease, the researchers wrote.
The study, “Longitudinal assessment of chorea in Huntington’s disease using digital passive monitoring,” was published in npj Digital Medicine.
Chorea affects more than 90% of people with Huntington’s, a neurodegenerative condition that causes movement problems, cognitive decline, and behavioral changes. These involuntary movements often first affect the face or extremities before spreading to other body parts, progressively interfering with daily life and contributing to stigma and reduced quality of life.
Clinicians may miss subtle changes
Doctors usually assess chorea during clinic visits using the Total Motor Score (TMS), a component of the validated Unified Huntington’s Disease Rating Scale that evaluates motor function in people with Huntington’s. While widely used, the assessment relies on doctors’ judgment, provides only a snapshot of the symptom between visits, and may miss subtle or infrequent movements.
Digital tools, such as smartwatches and other wearables that collect movement data in daily life, have increasingly been explored as a means to complement clinic-based assessments. However, “further innovation is … needed to develop a digital score capable of sensitively tracking the progression of chorea,” the researchers wrote.
To develop the DPCS, the team drew on data from 958 participants enrolled in four studies. These included a natural history study (NCT03664804) that followed people with early-stage Huntington’s over 15 months, and the Phase 3 GENERATION HD1 clinical trial (NCT03761849), which tested Roche’s now-discontinued experimental treatment tominersen against a placebo in people with established Huntington’s.
Participants wore smartwatches for as long as two years, collecting an average of 7.85 hours of wrist movement data per day, nearly six days per week. About once every three months, they visited clinics where trained clinicians rated upper-limb chorea using the TMS on a scale from 0 (no chorea) to 4 (marked and prolonged involuntary movements).
The researchers used artificial intelligence (AI) to compare clinician ratings with smartwatch movement data and build the DPCS. The AI learned which wrist movement patterns tended to match milder or more severe chorea by analyzing features linked to the intensity, instability, and unpredictability of movement, allowing it to generate a chorea severity score from smartwatch data alone.
When tested in a separate group of 310 trial participants, DPCS produced the same chorea rating as clinicians in 53.1% of assessments. In another 43.7%, the difference was one point. Larger discrepancies were rare, occurring in 3.2% of assessments. This performance approximated “levels of agreement expected when different clinicians evaluate chorea in the same patients,” the researchers wrote.
Researchers next tested whether DPCS could track worsening chorea over time using data from 181 GENERATION HD1 participants who received a placebo and were not included in DPCS training.
The smartwatch-based score increased steadily, corresponding to an estimated 0.13 point-per-year worsening. A comparable rate — 0.15 points annually — was observed in the natural history study, with no significant differences between studies.
Compared with standard clinic ratings, which showed less consistent progression during GENERATION HD1, DPCS produced a steadier pattern of change, suggesting it may be less affected by factors that can complicate symptom assessment during clinical trials, including visit-to-visit variability and placebo-related influences.
“The comparable progression of the DPCS in observational versus interventional studies confirms its robustness to such biases that [influenced] in-clinic motor assessments,” the team wrote.
“With this proof-of-concept study, we show that passive-monitoring allows the estimation of chorea during the everyday life of people with HD, in a way that is robust to in-clinic biases,” the researchers concluded. “Wearables offer promising and low burden options to evaluate the effects of signs of HD and other neuromuscular diseases, and could allow a better understanding of the effects of potential disease modifying or symptomatic treatments.”
Leave a comment
Fill in the required fields to post. Your email address will not be published.