Linear dependence of peak, mean, and pressure–time integral values in plantar pressure images

Abstract

Dynamic plantar pressure images are routinely used in clinical gait assessment, and peak pressure, mean pressure, and pressure–time integral are the most frequently used parameters to summarize these images. Many studies report only one parameter, but other studies report all three. The interdependency of these variables has not been explicitly studied previously. The purpose of this study was to describe the linear relation between these three pressure parameters. 327 subjects walked normally over a pressure plate. Peak pressure, mean pressure and pressure–time integral were calculated for 10 different anatomical areas and, after applying a previously described spatial normalization procedure, these variables were also calculated for each pixel. Mean pressure was highly correlated with peak pressure (r = 0.90 ± 0.09) and pressure–time integral (r = 0.81 ± 0.13) for pixels. Peak pressure and pressure–time integral showed a linear correlation coefficient of r = 0.78 ± 0.21. The pressure parameters of the forefoot pixels were more highly correlated than the heel pixels. The current results have two major implications: (1) plantar pressure parameters (peak, mean, and impulse) can be reasonably compared across studies, even across parameters, and (2) the variables most commonly used to characterize plantar pressures are highly inter-correlated, implying that a smaller set of parameters may more efficiently capture the biomechanical behavior of interest.

Introduction

Plantar pressure measurement has been widely used to study the foot. The foot is typically divided into anatomical areas (masking process) to calculate plantar pressure parameters (PPP) [1] including, most often: peak pressure, pressure–time integral and mean pressure. Many studies report the results of one parameter only, which makes it difficult to compare results from various studies. Therefore, knowledge about the relation between the parameters is very helpful in comparing the results of studies describing different PPPs.

Masking processes can affect PPP values [2], as the shape and number of the masking areas differ between studies. Recently, new methods were developed that spatially normalize plantar pressure images for foot size and foot progression angle [3], [4]. The main advantage of these methods is that plantar pressure can be studied at a sensor level. This technology enables a variety of PPP analyses, including the sensor-level relations amongst PPPs. The aim of this study was to compare the sensor-level and region-level relations amongst PPPs.