Résumé |
A boom in scientific interest in sports science has moved the field forward
substantially, from an isolated sub-discipline of physiology to a shining example
of the impact of multidisciplinary science. The rapidly growing amount of
available exercise data, that is currently almost unexplored, holds a great
potential for new quantitative research. While traditional studies are limited by
a small sample of participants, big data collections make it possible to bring the
laboratory to the field, and study millions of subjects under real world
conditions.
In this talk I shall give two examples for physics inspired research in exercise
physiology:
(1) The analysis of a large dataset obtained by runners from wearable exercise
trackers (covering about 20 million running kilometres), guided by a mathematical
model for running performance. From this we obtained a better understanding of
the complex interplay between training and performance. Identification of key
performance parameters allows accurate race time prediction and quantification of
correlations with training volume and intensity. Additionally, application of the
model to the last 100 years of running world records provides novel insights into
the evolution of physiological characteristics of runners, the detection of
doping, and the effect of new training approaches and of new technological
advances.
(2) The study of fluctuations of the human heart beat during physical exercise. I
shall show that running across various training and racing events changes the
scaling and correlations of beat-to-beat intervals (BBIs), using methods for the
analysis of non-stationary time series. These changes can be related to the
exercise intensity quantified by the heart rate. BBIs how multiscale
anticorrelations with both universal and individual scale-dependent structure.
|