Statistical foot shape-function modelling
A comprehensive dataset (N = 100) consisting of 3D foot scans under varying loads (captured as described elsewhere) as well as 3D motion tracking and ground reaction forces during various walking and running tasks was used to assess how accurately foot mechanics can be predicted from foot shape characteristics.
` `
To construct a foot shape-function model (SFM) 3D vertex positions from registered and scaled foot scans were combined with time normalised foot joint kinematics and kinetics. A principal component analysis was then performed to determine the main sources of covarying shape and function parameters.
` `
The main sources of shape variation were the longitudinal (PC1) and transverse arches, foot width, relative lengths of the forefoot and rearfoot (PC2) and orientation of the forefoot relative to the rearfoot (PC3) along with their associated joint mechanics.
` `
Amongst the largest sources of variation in joint mechanics were midtarsal joint range of motion (ROM) throughout stance, and differences in tarsometatarsal and MTP joint moments during midstance when walking (PC1). Further variations included differences in midtarsal plantarflexion during the first 10 and last 40-50% of stance, and ankle, subtalar, midtarsal, tarsometatarsal and MTP joint moments regardless of locomotor task (PC2). Differences in tarsometatarsal plantarflexion during short periods of the stance phase and ankle joint moments during mid to late stance in all walking tasks (PC3) were also sources of substantial variation.
` `
A leave-one-out analysis showed that the SFM can be used to accurately predict foot joint angles and moments from external foot shape and deformation for most feet. However, for some feet the model failed to produce accurate predictions.
` `
Despite accounting for the largest amount of the overall variation in foot shape and joint mechanics, the three largest SFM PCs only accounted for a small proportion each (14.7, 8.6 and 7.1%, respectively), most likely due to the high structural complexity and variability of the foot.