Human running speed is optimized for efficient energy use

Experiments on humans have long demonstrated that minimizing energy expenditure is an important objective while walking along and that people adjust aspects of their walking style constantly to keep energy costs at a minimum. When it comes to running, however, this relationship has not been tested. In the laboratory, human runners have been shown to choose preferred running speeds that correspond to a minimization of energy expenditure, but it is not known whether runners in natural settings do the same. 

The proliferation of digital movement tracking devices has given researchers an opportunity to find out whether runners have preferred speeds in free free-living environments or whether they vary their speed depending on the distances they run, as hypothesized previously. By combining data from runners monitored in a lab with data from over 37,000 runs recorded on wearable fitness trackers, scientists have now been able to answer this question.

The research group, made up of scientists from Queens University in Ontario and Stanford University in California, has been studying the mechanics of running in labs for 15 years but had not investigated running “in the wild” before now. 

“We were able to fuse the two datasets to gain new insights and combine the more messy wearable data with the gold standard lab experiments to learn about how people run, out in the world,” said study co-author Jennifer Hicks, deputy director of Stanford’s Wu Tsai Human Performance Alliance.

The free-living data collected by wearable digital devices included recordings of distance and elevation, derived from GPS data at 3-s intervals. A total of 4,645 runners self-reported details of their sex, age, height, and weight, and their running data were recorded over a 14-month period between August 2016 and October 2017. This added up to a total of 37,201 runs, or more than 28,022 hours (3.20 years) worth of running data.  

To find out whether runners in a free-living environment choose different speeds for different distance tasks, the researchers selected runners in the database who ran at least one run at each of three distances: 3.22, 6.44, and 9.66 ± 0.80 km (2, 4, and 6 ± 0.5 mi). They tested for differences in each runner’s average speed over these three distances and found there were no significant differences. This indicated that runners choose a preferred running speed and stick to it, irrespective of the distance they plan to cover.

The scientists then compared these preferred running speeds to experimentally determined speeds that are optimal in terms of minimizing the energy cost. The results of the analysis showed a remarkable consistency across the combined datasets – one derived from laboratory experiments and the other derived from runners with wearable tracking devices. 

“We intuitively assume that people run faster for shorter distances and then would slow their pace for longer distances,” said study first author Jessica Selinger. “But this wasn’t the case. Most of the runners analyzed stuck with the same speed, whether they were going for a short run or a long haul over ten kilometers.”

Intuitively, one might assume that people run faster over shorter distances and slow their pace for longer distances. This is what is found in competitive races where minimizing time is the explicit goal. But the current study shows that the average person out for some running exercise has a preferred speed that is largely independent of running distance and is also consistent with the objective of minimizing energy expenditure. 

The irony of these findings, published today in the journal Current Biology, is that while we may be motivated to go for a run in order to burn up calories, it seems we are choosing a running speed that actually minimizes the calories we use. This makes sense from an evolutionary perspective, and is consistent with the locomotion preferences recorded for other, nonhuman animals living in natural environments. If we wish to increase the calories we use up while running, we will need to increase our speed by using strategies, such as running with faster companions or listening to music with a fast beat as our feet pound the sidewalks.

Selinger and Hicks hope that having large pools of fitness data from wearable movement tracking devices will help researchers to gain insights about populations. 

“You can look at connections with the built environment and access to recreation resources and start to layer all of that data to really understand how to improve physical activity and health more broadly,” said Hicks.

—

By Alison Bosman, Earth.com Staff Writer