Welcome to Watchwalk

Register for an Account

Frequently Asked Questions

What is WatchWalk?

WatchWalk is a method designed to interpret motion sensor data obtained from smartwatches to quantify a person's walking and sleeping behaviours. It measures various aspects such as walking speed, quality, distribution, and step count.

Why are day-to-day walking and sleeping behaviours important?

Walking performance and sleep patterns are linked to the risk of injuries from falls and the potential development of dementia and depression. Additionally, understanding whether these behaviours are improving or deteriorating helps clinicians decide on appropriate treatment methods.These metrics can be used as risk factors, trial endpoints, confounders and mediators in health studies.

Can't we evaluate sleep and walking through self-reporting?

While self-reporting is a quick method, it can be subjective and biased by memory and perception, making it difficult to quantify accurately.

What data does WatchWalk need?

WatchWalk requires at least 3 days of motion sensor (accelerometric) data from a smartwatch, with at least 12 hours of data per day. For a comprehensive view, seven days of 24-hour data is recommended. This data can be retrieved from devices like Axivity (e.g., AX3/AX6), ActiGraph (e.g., GT3X+), or other smartwatches providing continuous tri-axial acceleration signals at 100Hz with timestamps. Additionally, information on a person's height, sex, and age is needed to estimate day-to-day walking speed and facilitate age-sex-matched comparisons.

Is WatchWalk accurate?

WatchWalk has been trained and validated with 101 participants aged between 18 and 81. Its concurrent validity and test-retest reliability have been confirmed with an external dataset of over 70,000 individuals aged over 50.

How much does it cost to use WatchWalk?

Our goal is to make WatchWalk accessible for all research purposes. Currently, access to the WatchWalk Platform is free of charge. This free access is available until our server capacity reaches its limits.

How much budget do I need to use WatchWalk measurements in our study?

In 2025, the Axivity AX3 sensor with a wristband costs £139 each, plus shipping. This will be the main cost required for using the WatchWalk method. https://axivity.com/product/ax3

How does WatchWalk process the data?

Motion sensor data from smartwatches is segmented into 4-second windows, merged and filtered. We then extract features using frequency-domain analysis and other BioSignal analyses. Subsequently, we classify the windows into different behaviours ( walking, running, stationary etc.) and different hand position patterns ( arm-swing, hands in pockets etc.) with SVM Classification. We finally apply SVM regression and biosignal processing techniques to measure walking performances and sleep behaviours.

How is the data secured?

All data processed by WatchWalk is anonymized and encrypted to ensure the privacy and security of the participants. Our secured severs are situated in Sydney, NSW.

Who can benefit from using WatchWalk?

Researchers, clinicians, and healthcare professionals can benefit from using WatchWalk to monitor and analyze the walking and sleeping behaviours of their subjects or patients. It is particularly useful in studies related to mobility, fall prevention, and the impact of physical activity on health outcomes.

What are the main limitations of WatchWalk?

Due to server capacity, the Watch Walk Platform currently supports automated processing for accelerometric data of 7 days (168 hours per recording) or less. Please contact our team at l.chan@neura.edu.au if you intend to process data collected over more than 7 days.

Not all WatchWalk measures apply to walking aid users. Further, WatchWalk has not been validated in children and teenagers.

What measures does WatchWalk include?

Here are a list of digital biomarkers we generate:

  • Sleep
  • Sleep duration (hour)
  • Bedtime (hour of the day)
  • Gait Quantity
  • Steps per day
  • Walking duration (hour)
  • Walk Length distribution
  • Longest continuous walk duration, (second)
  • Walk ≥ 8 second (%)
  • Walk ≥ 60 second (%)
  • Gait Speed and Intensity
  • Maximal walking speed (centimetre per second)
  • Usual walking speed (centimetre per second)
  • Cadence median (spm)
  • Gait Quality
  • Step-time variability (millisecond)
  • Step regularity (%)
  • Stride regularity (%)
  • Walk hand positions
  • Walk with Arm Swings (%)
  • % Texting
  • % Phonecall
  • % Hands in pockets
  • % Shoulder bag
  • % Briefcase
  • Physical activity level
  • Time spent doing sedentary activity (minute)
  • Time spent doing light physical activity (minute)
  • Time spent doing moderate physical activity (minute)
  • Time spent doing vigorous physical activity (minute)
  • Others
  • Non-wear duration (hour)

What studies has WatchWalk been used in?

The UK Biobank

https://www.ukbiobank.ac.uk/

The UK Biobank is a large-scale biomedical database and research resource containing in-depth genetic, health, and lifestyle information. Watch Walk has been applied to extract real-world walking performance from seven-day wrist accelerometry data in over 92,000 participants aged 45 to 79. This analysis was conducted under project application IDs #56109 and #103840, with dataset upload #ID 4789.

The Sydney Memory Ageing Studies (MAS2)

https://www.mas2.org/

The Sydney Memory and Ageing Study (MAS2) is a five-year longitudinal study investigating cognitive ageing, dementia risk factors, and health trajectories in older adults. It follows community-dwelling individuals aged 70–90 from the Sydney region through comprehensive assessments.

The Higashiura study

https://www.ncgg.go.jp/ri/report/20240415.html

The Higashiura Study is a population-based longitudinal study conducted by the National Center for Geriatrics and Gerontology in Japan, in collaboration with industrial partners, starting in 2023. Watch Walk will be applied to wrist accelerometry data collected from over 2,300 older adults.

Are there any publications where I can learn more about Watch Walk?

  1. Chan, L. L. Y., Choi, T. C. M., Lord, S. R., & Brodie, M. A. (2022). Development and large-scale validation of the Watch Walk wrist-worn digital gait biomarkers. Sci Rep, 12(1), 16211. https://doi.org/10.1038/s41598-022-20327-z
  2. Chan, L. L. Y., Arbona, C. H., Brodie, M. A., & Lord, S. R. (2023). Prediction of injurious falls in older adults using digital gait biomarkers extracted from large-scale wrist sensor data. Age Ageing, 52(9). https://doi.org/10.1093/ageing/afad179
  3. Chan, L. L. Y., Brodie, M. A., & Lord, S. R. (2023). Prediction of Incident Depression in Middle-Aged and Older Adults using Digital Gait Biomarkers Extracted from Large-Scale Wrist Sensor Data. Journal of the American Medical Directors Association, 24(8), 1106-1113.e1111. https://doi.org/https://doi.org/10.1016/j.jamda.2023.04.008
  4. Chan, L. L. Y., Lord, S. R., & Brodie, M. A. (2024). Daily-Life Walking Speed, Quality and Quantity Derived from a Wrist Motion Sensor: Large-Scale Normative Data for Middle-Aged and Older Adults. Sensors, 24(16), 5159. https://www.mdpi.com/1424-8220/24/16/5159
  5. Chan, L. L. Y., Yang, S., Aswani, M., Kark, L., Henderson, E., Lord, S. R., & Brodie, M. A. (2024). Development, Validation, and Limits of Freezing of Gait Detection Using a Single Waist-Worn Device. IEEE Transactions on Biomedical Engineering, 71(10), 3024-3031. https://doi.org/10.1109/TBME.2024.3407059
  6. Osuka, Y., Chan, L. L. Y., Brodie, M. A., Okubo, Y., & Lord, S. R. (2024). A Wrist-Worn Wearable Device Can Identify Frailty in Middle-Aged and Older Adults: The UK Biobank Study. J Am Med Dir Assoc, 25(10), 105196. https://doi.org/10.1016/j.jamda.2024.105196
  7. Chan, L. L. Y., Espinoza Cerda, M. T., Brodie, M. A., Lord, S. R., & Taylor, M. E. (2025). Daily-life walking speed, running duration and bedtime from wrist-worn sensors predict incident dementia: A watch walk – UK biobank study. International Psychogeriatrics, 100031. https://doi.org/https://doi.org/10.1016/j.inpsyc.2024.100031

How can I get started with WatchWalk?

To get started with WatchWalk, researchers need to register on the WatchWalk Platform, obtain the necessary motion sensors, and follow the setup instructions provided.

Detailed guidelines and support are available to help with the setup and data collection process.