A polygon model of the functional base-of-support during standing improves the accuracy of balance analysis

Mathematical balance models have the potential to identify people at risk of falling. However, most balance models depend on a model of the base-of-support (BOS) of the feet to calculate how well someone is balancing. Here, we evaluate the functional base-of-support (fBOS) during standing: the convex polygon on the bottom of the foot that can support a large fraction of the body's weight. First, we develop a geometric model of the fBOS by measuring the center-of-pressure (COP) and kinematic data of the feet of 27 younger adults instructed to move their body in large loops without taking a step. Next, we extract a planar convex polygon that contains the COP data. Finally, we compare the area of this fBOS model to a marker-based BOS model before evaluating if the fBOS differs across four common conditions: footwear, stance-width, foot dominance, and during single and double-stance. We found that the fBOS is much smaller (23% the size) than a marker-based BOS model. Our analysis suggests that using the fBOS, rather than a marker-based BOS, can improve the accuracy of the margin-of-stability by 20% of foot width and 16% of the length. In addition, we found that the fBOS area does not differ across footwear (p=0.88), stance-width (p=0.88), and foot dominance (p=0.68), but during single stance, the fBOS is 17% (p=0.0003) larger than during double-stance. The variability of the fBOS area suggests that future studies should establish the repeatability and reliability of the assessment and systematically study the effects of different types of footwear. We have put the fBOS models, example data, and code in the public domain to help others build on our work.