04/09/2026
Measuring physical activity in toddlers is challenging, with their unique movement patterns limiting the accuracy of traditional accelerometer methods.
A research team in Canada developed machine learning models that more accurately distinguish physical activity, sedentary time and non-volitional movement in toddlers, outperforming many existing methods.
Importantly, the researchers also developed an open-access, no-coding-required tool, making these models accessible to researchers and clinicians and supporting more accurate links between toddler movement, health outcomes and physical activity guidelines.
Read the article: https://zurl.co/EUThl