View all news

MRC IEU Researchers develop ‘Activity Bigrams’ - a novel approach to represent how activity changes from one moment to the next.

10 July 2017

In a new study MRC IEU researchers have developed a novel method called activity bigrams to analyse how physical activity changes from one moment to the next and how this relates to other traits and disease.

Physical activity is an important epidemiological factor and is related to many diseases such as type II diabetes and heart disease.

In this study researchers tested how the activity bigrams were related to body mass index (BMI), using data from the Avon Longitudinal Study of Parents and Children (ALSPAC). They found several associations between changes in the frequency of particular activity bigrams and BMI, although the study was unable to infer if this link is causal. Bigrams may be an important method to determine how activity patterns relate to disease in future research.

Previous studies of physical activity have used few statistics to represent physical activity, such as the average activity level overall. Activity bigrams allow the pattern of activity, such as switching between low and moderate intensity, or moderate and vigorous intensity, to be investigated. This technique may be able to provide more precise public health advice in the future.

Dr Louise Millard of the MRC IEU says “This study demonstrates how a novel representation of physical activity called activity bigrams, can be used to assess how changes in activity from one moment to the next relates to traits and disease. We found that some of the activity bigrams were associated with BMI, showing that these characteristics of activity may be important for health. Further work is needed to replicate these identified associations, determine if these associations are causal, and investigate how activity bigrams relate to other traits and disease”.

You can read the study in full in the International Journal of Epidemiology here: https://academic.oup.com/ije/article/3902974/Physical-activity-phenotyping-with-activity

Edit this page