The analysis of human

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The analysis of human gait frequently necessitates the identification of foot-strike and toe-off events. This is characteristically achieved with the use of force plates. However, when force plate information is not available, alternative methods are necessary. Several kinematic algorithms for the determination of foot-strike and toe-off have emerged in the literature, but the effectiveness of these methods have yet to be compared to one another. The rationale for this investigation was therefore to compare these methods with a force platform.

The validity of these algorithms was compared with the results obtained using synchronized vertical ground reaction force (GRF) recordings of eleven participants running at 4.0 ms-1 for a total of 5 trials.

The results indicate that the most accurate method for the determination of heel-strike was the Alto et al (1998) algorithm producing a mean error of 0.016(s), toe-off most accurately determined via the Dingwell et al (2001) algorithm which produced a mean error of 0.011(s).

Thus, a strong argument is presented for the utilization of these algorithms during gait analysis.


Gait analysis necessitates identification of both heel-strike and toe-off to define key components of the gait cycle. This is most accurately quantified using force plates where a threshold is defined to determine heel-strike and toe-off [1]. For studies of running kinematics force data is not always available, collecting force data relies on the ability of the participants to consistently make contact with the platform without altering their natural gait pattern. Whilst footswitches and pressure sensors are often utilized in they may reduce the number of available analogue channels or introduce an additional source of error to the data. Thus, it is necessary to identify alternative methods of quantifying foot strike and toe-off. Several kinematic methods are available for gait event determination, but comparisons of their accuracy for defining stance phase events has, yet to be reported.

Mickelborough et al.,[2] developed a method of determining gait events during walking, heel-strike was associated with the second of the W shaped troughs of the foot vertical velocity curve in the Z (vertical) axis, toe-off was determined as the minimum position of the toe-markers in the Z axis. Alton et al., [3] used the minimum position of the distal heel marker in the Z axis in order to determine footstrike. Toe-off was defined using the same method as Mickelborough via the position of the metatarsal markers in the Z axis. Similarly, Zeni et al., [4] proposed two methods of identifying gait events. The first used the difference in displacement of the peaks and troughs of sacral and foot markers in the sagittal plane. The second method is a velocity based technique. The velocity of the heel marker in the sagittal plane changes from positive to a negative direction at each heel strike. The frame at which backward movement of the foot is initiated is termed heel-strike. At the initiation of swing phase the velocity of the heel or toe markers alters from negative to positive and is thus labelled toe-off.

Hreljac and Stergiou [5] utilized shank and foot motion in the sagittal plane. They determined foot strike as the time that coincided with the minimum sagittal plane foot angular velocity, and toe-off as the local minimum of the shank angular velocity. Schace et al., [6] utilized the vertical velocity and displacement of the foot markers to identify gait events for overground and treadmill running. Heel strike was deemed to be the time of the downward spike of the vertical velocity of the 1st metatarsal and the plateau in the displacement of the lateral malleoli marker in the Z axes. Toe-off was deemed to be the onset of the rise in vertical displacement and velocity of the 1st metatarsal marker. Finally, Dingwell et al., [7] provided a kinematic method designed specifically for treadmill running. Foot strike was deemed to be the first time when peak knee extension occurred and toe-off was determined as the second occurrence of peak knee extension.

The overall objective of this investigation was to illustrate the most accurate means of predicting heel strike and toe-off, by contrasting the computationally predicted events to those detected using force data.



Eleven male participants volunteered to take part in this investigation (age 19 + 1 years; Height 176.5 + 5.2 cm; Mass 78.4 + 9.0 kg). Ethecal approval for this project was obtained from the School of Psycology ethics commnttee, University of Central Lancashire and each participant provided verbal concent.


Participants ran at 4.0 m s-1 along a 20 m runway striking the centre of a force plate with an onset of 20 N (Kistler, Kistler Instruments Ltd., Alton, Hampshire, UK; Model 9281CA), sampling at 1000Hz. Timing gates were used to control velocity, a maximum deviation of +5% from the specified target was allowed. Kinematic data was obtained via an eight camera infra red motion analysis system (Qualisys Medical AB, Goteburg, Sweden) operating at 350Hz. The marker set used for the study was based on the CAST technique (Cappozo et al [8]. Retro-reflective markers were attached to the 1st and 5th metatarsal heads, medial and lateral maleoli, medial and lateral epicondyle of the femur, greater trochanter, iliac crest, anterior superior iliac spines and posterior superior iliac spines with tracking clusters positioned on the shank and thigh of left and right legs. A static trial was captured to define the pelvis, thigh, foot and tibial segments.

Kinematic parameters were quantified using Visual 3-D (C-Motion Inc, Gaithersburg, USA) and filtered using at 6 Hz using a low pass Butterworth 4th order filter following interpolation with a maximum gap fill of 10 frames. Five trials were averaged for each participant. Angles were created about an XYZ rotation cardan sequence referenced to coordinate systems about the proximal end of the segment, where X is flexion-extension; Y is ab-adduction and is Z is internal-external rotation. In order to validate the effectiveness of these methods they were compared to a gold standard, in which event detection is based on force plate data (Hansen et al., [1]. Heel strike was quantified as the first instance at which the vertical component of the GRF was greater than 20N; toe-off was determined to be the first instance in which the vertical GRF fell below 20N. The time (s) in which each event determined from the computational algorithms occurred was contrasted to the equivalent event determined by the vertical ground reaction force. The difference in the time of occurrence was then tabulated in Excel (Microsoft Corp., Redmond, WA, USA). A positive value represented an event defined after the event established from the force plate and a negative value indicates that the computational algorithm defined the event prior to the force plate event.



The aim of this investigation was to identify the most appropriate algorithms for the determination of heel-strike and toe-off using kinematic techniques during overground running. A reliable algorithm must be both reliable and accurate allowing the gait cycle to be separated into phases of stance and swing.

The results suggest that heel-strike and toe-off are most accurately determined using different algorithms. Heel-strike was most accurately determined using the Alton et al., [3] method using position of the distal heel marker, whereas toe-off was most appropriately determined via the Dingwell et al., [7] knee extension method.

The mean errors for event detection appear to correspond to those reported by other studies, with the exception of the Mickelborough et al., [2] method which was confounded by repeatability issues. That is, the vertical velocity of the foot markers often exhibited multiple maxima and/or minima causing gait events to be located incorrectly, although this is common when applying algorithms designed for walking to running data.

In conclusion the Alton et al [3] and Dingwell et al [7] represent simple and robust methods for determining gait events that do not require 3-D analysis to employ. Thus a strong argument is presented for the utilization of these algorithms. Additional work will determine the applicability of these algorithms to treadmill and pathological locomotion.


  1. Hansen, A.H., Childress, D.S., Meier, M.R. (2002). A simple method for determination of gait events. Journal of Biomechanics, 35, 135-8.
  2. Mickelborough, J., Van der Linden, M.L., Richards, J., Ennos, A.R. (2000). Validity and reliability of a kinematic protocol for determining foot contact events. Gait and Posture. 11, 32-37.
  3. Alton, F., Baldey, L., Caplan, S., and Morrissey, M.C. A kinematic comparison of overground and treadmill walking. Clinical Biomechanics, 13, 434-440.
  4. Zeni, J.A., Richards, J.G., and Higginson, J.S. (2008). Two simple methods for determining gait events during treadmill and overground walking using kinematic data. Gait & Posture, 27, 710-71.
  5. Hreljac, A., and Stergiou, N. (2000). Phase determination during normal running using kinematic data. Medical and Biological Engineering and Computing, 38, 503-506.
  6. Schache, A.G., Blanch, P.D., Rath, D.A., Wrigley, T.V., Starr, R. and Bennell, K.L. (2001). A comparison of overground and treadmill running for measuring the three-dimensional kinematics of the lumbo-pelvic-hip complex.Clinical Biomechanics, 16, 667-680.
  7. Dingwell, J.B., Cusumano, J.P., Sternad, D., Cavanagh, P.R. (2001). Local Dynamic Stability Versus Kinematic Variability of Continuous Overground and Treadmill Walking. Journal of Biomechanical Engineering, 123, 27-32.
  8. Cappozzo, A., Catani, F., Leardini, A., Benedeti, M.G., and Della, C.U. (1995). Position and orientation in space of bones during movement: Anatomical frame definition and determination. Clinical Biomechanics, 10, 171-178.