Plate Nº 30 · recorded November 11, 2026
Health & Medicine ResearchReported finding
A Wrist Sensor Spotted Parkinson's Signs Nearly 7 Years Early
Wrist sensor data from 73,413 UK Biobank volunteers revealed walking and arm-swing differences up to nearly seven years before a Parkinson's diagnosis — but only at the group level.
By James Calloway4 min read739 words
In brief
- Arm-movement differences appeared in wrist recordings nearly 7 years before Parkinson's diagnosis
- Study analyzed 73,413 UK Biobank participants; 314 were diagnosed with Parkinson's during up to a decade of follow-up
- Those diagnosed within 40 months averaged ~4,000 steps/day vs ~6,400 for undiagnosed participants
- Findings are group-level only and cannot predict an individual's risk
- Published in the Journal of Neural Transmission; three authors may earn royalties from commercializing the Watch Walk algorithm
A week of wrist recordings flagged differences in arm movement nearly seven years before some people received a Parkinson's diagnosis. That's the headline number from a new study out of the University of New South Wales in Sydney, published in the Journal of Neural Transmission. Impressive. Also easy to overread, and we'll get to why.
What did the researchers actually find?
Parkinson's disrupts a piece of walking most of us never think about. When your right leg steps forward, your left arm swings forward with it, countering the twist of your body. As the disease kills off nerve cells that produce dopamine — the chemical that helps control movement — muscles stiffen and movements slow, and that arm-leg coordination frays.
The team, led by joint first authors Paulo Pelicioni of UNSW's School of Health Sciences and Sophie Waller of its School of Clinical Medicine and Prince of Wales Hospital's neurology department, turned to the UK Biobank. That long-running health study asked more than 100,000 volunteers to wear a wrist sensor for a week. The researchers analyzed usable recordings from 73,413 people and followed their medical records for up to a decade. During that window, 314 people received a Parkinson's diagnosis after wearing the sensor.
Here's where it gets concrete. A wrist sensor can't tell a swinging arm from a hand grabbing a coffee cup on its own. So the team used software called Watch Walk to isolate the walking stretches of each recording, estimating step counts, walking speed, and how regularly steps fell. The direction the wrist pointed, and how it moved, helped the software separate a natural arm swing from a hand held near the head or in front of the body.
The numbers that emerged:
- People diagnosed within 40 months of wearing the sensor averaged about 4,000 steps a day.
- People with no recorded Parkinson's diagnosis averaged about 6,400.
- The gap shrank for those diagnosed later — but it didn't vanish.
- Those later diagnosed also hit lower top walking speeds, with different step rhythm and less walking done in long stretches.
- They spent less of their walking time swinging their arms naturally, a difference visible in recordings made nearly seven years before diagnosis.
That last point matters. A step counter alone would miss it. The wrist tells a story the pedometer can't.
What does this change for diagnosis?
Honestly? Not much yet — and the researchers say so themselves. The results describe differences between groups. They cannot tell you, individually, whether you'll develop Parkinson's. Pain, medication, and other health conditions all bend how a person walks, and none of them have anything to do with Parkinson's.
There's a structural limitation too. Each person was recorded during a single monitoring period, so the team couldn't pinpoint when anyone's arm swing actually started to change. A follow-up study tracking the same people at regular intervals could answer that question. Until someone runs it, the timeline stays fuzzy.
If the findings hold up in other groups, the recordings could help researchers find candidates for trials of treatments aimed at slowing the disease in its earliest stages. Note the chain of conditionals there. Could help. If they hold up. That's a trial-recruitment tool, not a screening test.
One more thing worth flagging: three authors disclosed an agreement with SpaceTalk Inc. to commercialize the Watch Walk algorithm, with possible royalties attached. Research groups can still use the algorithm for free, but a financial stake in a diagnostic tool is the kind of detail that deserves daylight, not a footnote.
Where does this fit?
This isn't the only trail scientists are following toward earlier Parkinson's detection. Other work has hunted for early signs in blood proteins. Wrist recordings add something those can't: a view of how people move in ordinary life, outside the exam room, over a full week rather than a twenty-minute appointment.
That's the real promise here — passive, cheap, continuous movement data from devices millions already wear. It's also the real risk. Movement differences are not disease-specific, and a wrist sensor that whispers "maybe Parkinson's" seven years out could do harm if it's wrong. The study's own caveats — group-level differences, single recordings, confounding conditions — are the ones to keep in mind before anyone calls this a breakthrough.
via sciencealert.com (Source)
More from James Calloway
Show full bio
Staff writer covering regulation, policy and their impact on general science news at SciBeat.
346 articles
Nearby plates
- Wrist Sensor Beats Clinicians at Tracking Stroke Recovery
- Lower walking activity linked to later neurological diagnoses
- Doubling Lifespan in Parkinson's Mice: How One Protein Stopped Toxic Clumps
- More Daily Steps Linked to Fewer Teen Mental Health Problems
- Letting Your Arm Hang During a Blood Pressure Check Can Add 6.5 mmHg