Skip to content


Journal of Foot and Ankle Research

Open Access

New insights into stance phase foot biomechanics using pedobarographic statistical parametric mapping

  • Todd C Pataky1Email author,
  • John Y Goulermas2,
  • Paolo Caravaggi1 and
  • Robin H Crompton1
Journal of Foot and Ankle Research20081(Suppl 1):O25

Published: 26 September 2008


Stance PhaseStatistical Parametric MappingPlantar PressureLinear Statistical ModelLongitudinal Arch


There is disagreement in the literature regarding the peak plantar pressure correlates of walking speed. Some studies report generalized increased pressures across the entire plantar surface [1] while others report decreases in lateral forefoot pressures and infer a medial shift in load with increased walking speeds [2]. The purposes of this study were: (a) To use statistical parametric mapping (SPM), a high resolution statistical technique, to clarify the pressure correlates of walking speed and (b) To corroborate SPM results with those of a traditional ten-region subsampling technique.


Ten subjects performed twenty trials of each of slow, normal, and fast walking in a fully randomized design. Plantar pressure data were collected using a Footscan 3D system (RSscan, Belgium). Walking speed was recorded with a six camera ProReflex system (Qualisys, Sweden). Peak pressure images were registered [3] using an optimal rigid body transform, and then between-subjects registration was performed using an optimum affine transform to ensure homologous structure overlap. The registered images were analyzed using pedobarographic SPM (pSPM), an adaptation of an established cerebral fMRI technique [4]. A parametric mass univariate general linear statistical model blocked SUBJECT effects and yielded a generalized SPEED regression statistic having the Student's t distribution. The original peak pressure images were also subsampled over ten anatomical regions using commercial software (Footscan 7, RSscan). These data were analyzed using the same linear model.


pSPM analyses produced smooth and continuous statistical maps that exhibited positive correlation with walking speed over the heel and distal forefoot, but which also exhibited broad negative correlation over the midfoot and proximal forefoot (Figure 1A). The significance of these trends was confirmed across subjects (Figure 1B). Subsampling obscured these data, exhibiting negative correlation only in the lateral forefoot, and reversing the midfoot trend (Figure 1C).
Figure 1

(A) pSPM results for an example subject, unmasked. (B) Between-subjects pSPM results, masked at p < 0.001; data were qualitatively identical even after a Bonferroni correction of p < 5 × 10-5. (C) Subsampling results, masked at a Bonferroni-corrected p < 0.005.


pSPM analyses revealed novel information regarding stance phase foot biomechanics, suggesting that longitudinal arch collapse is actively reduced as a function of walking speed and that such prevention may be beneficial to propulsion, possibly through increased plantar aponeurosis tension. The data also demonstrate, critically, that traditional subsampling techniques can distort or reverse statistical trends due to regional conflation, procuring erroneous conclusions regarding foot function. Foot pressure image analyses should incorporate all pixel data wherever possible.

Authors’ Affiliations

School of Biomedical Sciences, University of Liverpool, UK
Department of Electrical Engineering and Electronics, University of Liverpool, UK


  1. Taylor AJ, et al: Foot. 2004, 14: 49-55. 10.1016/j.foot.2003.09.004.View ArticleGoogle Scholar
  2. Rosenbaum D, et al: Foot Ankle Surg. 1997, 3: 1-24. 10.1046/j.1460-9584.1997.00043.x.View ArticleGoogle Scholar
  3. Maintz JBA, et al: Med Image Anal. 1998, 2 (1): 1-37.View ArticlePubMedGoogle Scholar
  4. Friston KJ, et al: Hum Brain Mapp. 1995, 2: 189-210. 10.1002/hbm.460020402.View ArticleGoogle Scholar


© Pataky et al; licensee BioMed Central Ltd. 2008

This article is published under license to BioMed Central Ltd.