Science & Enterprise subscription

Follow us on Twitter

  • Cashing out ... Gene therapy pioneer Kathy High has left Spark after completing $4.3B union with Roche ... via Endp… https://t.co/9wCsBVL2U2
    about 15 hours ago
  • Amazon Care, the company's virtual medical clinic, is now live for Seattle employees https://t.co/yaXClfiY6K
    about 1 day ago
  • A survey of institutional investors shows these money managers for financial institutions are starting to take clim… https://t.co/jMNeawHlpR
    about 2 days ago
  • New post on Science and Enterprise: Institutional Investors Factoring in Climate Change https://t.co/26SjKQRHch #Science #Business
    about 2 days ago
  • ‘The disruption is enormous.’ Coronavirus epidemic snarls science worldwide https://t.co/J3Nyf9wXnG
    about 3 days ago

Please share Science & Enterprise

Algorithm Improves Activity Tracking for Wellness Apps

iPhone (William Hook/Flickr)

(William Hook/Flickr)

Engineers and physiologists at Northwestern University in Chicago developed an algorithm to improve the way health and wellness apps on smartphones track a user’s physical movements. Professor of physical medicine and rehabilitation Konrad Kording, with colleagues Stephen Antos and Mark Albert, published an advance version of their findings online in the Journal of Neuroscience Methods (paid subscription required).

The Northwestern team tackled the problem of smartphone users carrying their phones in different places during the day, depending on their activity at the time or what they’re wearing. The place owners carry their phones, however, can have an impact on the effectiveness of health and wellness apps that use built-in accelerometers for tracking physical activity.

In the study, researchers asked 12 healthy volunteers to carry a smartphone with an activity tracking app in designated carrying places (e.g. belt, pocket, or purse), while engaging in specified activities. The data were collected to highlight the variations in recording physical activity among those conditions second-by-second during the day. That same method was used to record physical activity with smartphones by two people with Parkinson’s disease.

The data collected were used to construct an algorithm to adjust the apps readings for different phone-carrying locations and activities. That algorithm is based on a hidden Markov model, a statistical technique used to associate heterogeneous sources with specified outcomes; “hidden” in this case refers to the changes from one state or condition to another, not the properties of the model.

Tests of the algorithm show it can help improve activity tracking of healthy individuals, and has the potential at least, to be used with Parkinson’s patients. “I believe we will have apps running on smart phones,” says Kording in a university statement, “that will know exactly what we’re doing activity-wise and will warn us of diseases before we even know that we have those diseases.”

Read more:

*     *     *

Please share Science & Enterprise ...

Comments are closed.