The relationship between structure and function of the human foot has attracted many investigators over the years, and has led to the development of many theories. In most cases, these theories have not been quantitatively verified due, in part, to the complex structure and the variable function of the human foot. In the present study the concept of load distribution under the sole of the foot has been used to present conceptually, and validate experimentally, the relationship between structure and function of the human foot.

Although plantar pressure distribution measurements during gait are used extensively and successfully as a screening tool, the origins of variation in plantar pressure distribution are not fully understood. To date, there is no model that can fully explain why load is distributed in a certain pattern in one person, and in different pattern in another. Furthermore, the exact reasons for the development of areas of high pressure in various clinical entities are not clear. The purpose of this study was to develop a comprehensive model which describes the relationship between structure and function of the human foot. In particular, this model was designed to explore how a combination of structural and functional mechanisms influence load distribution underneath the human foot during walking.

Attempts to predict plantar pressure parameters have been described since the mid 1970s. These provided regression equations to predict plantar pressure based on age, weight, walking speed, and angle of 'toe-out', but could not explain much of the variance in plantar pressure based on the given predictors. Later attempts to predict peak pressures based on a single factor also met with only limited success. It was hypothesized that both structural and dynamic variables affect regional peak pressure, and that a comprehensive study is needed to identify the significant factors.

In the present study, a number of structural and functional outcome measures (potential independent variables) were identified from a total of five structure and three function data sets as follows:​ (1) physical characteristics (age, height and weight), (2) anthropometric data, (3) passive range of motion at selected foot joints, (4) standardized measurements from weight bearing plain radiographs of the foot and ankle, and (5) mechanical properties of the soft tissue under the heel and the first and second metatarsal heads (MTH1 and MTH2). The function data were obtained from five walking trials in each subject at an average speed of 0.78 statures per second. Each walking trial was recorded by three Mac Reflex cameras, and gait analysis was performed leading to (6) stride foot parameters, (7) 3D joint kinematics, and (8) EMG. Plantar pressures in the right foot only were recorded simultaneously and subsequent analysis provided peak pressure values under the rearfoot, midfoot, MTH1 and MTH2.

Selected potential predictors from each data set were identified, based on their anatomical relevance and their relationship with peak pressure. Regression analysis was performed to explore what portion of the variance in plantar pressure under each region during walking could be explained by both structural and functional characteristics of the foot and lower extremity, using only the selected variables (from the list of all the independent variables) that were identified as potential predictors.

The final regression models for the five foot regions consisted of five to seven variables. Each one of the models introduced a unique combination of structural and functional variables. For example, heel peak pressure was a function of:​ the inclination of the calcaneus (+), the amount of unloaded soft tissue thickness under the heel (-), contact time during stance (-), the horizontal velocity of the foot at heel strike (+), the vertical velocity of the foot at one frame after heel strike (+), and age (-). The model accounted for 49.2% of the variance in peak heel pressure. In comparison, peak pressure under MTH1 was a function of:​ the inclination of the calcaneus (+), the Chopart's angle (+), the angle between the horizon and the proximal first phalanx shaft (-), the distance between floor a?id lower surface of the MTH1 sesamoids (-), Morton's Index (-), the average activity of the gastrocnemius over the third quarter of stance (+), and the talocrural dynamic range of motion (+). The model accounted for 48.6% of the variance in peak MTH1 pressure. A (+) sign represents direct association between pressure and a given variable, while a (-) sign represents inverse association between the two.

It was concluded based on these models that foot structure and function predict approximately 50% of the variance in peak pressure. Despite the similarity in predictive ability for the different regions, the contribution of structure and function for each anatomical region varied dramatically. Structure proved to be very dominant in predicting pressure under the midfoot and MTH1, while function accounted for the majority of the variance in peak pressure for the heel region.

The present study was not designed to provide immediate clinical application. However, it is well known clinically that mechanical factors and gait abnormalities contribute to the development of foot disorders. The results presented here provide quantitative data in the form of a multifactorial predictive model that provides insight into the potential etiological factors that are associated with elevated plantar pressure.