Model used for the biomechanical analysis
The biomechanical model used to in these examples is the simplest possible, consisting of a two-dimensional (planar) rigid body motion, with markers placed on the 5th metatarsal head, lateral malleolus, lateral knee-joint line, greater trochanter and gleno-humeral joint.
Whilst this very crude model relies on several assumptions, it gives relatively quick results suitable for a basic analysis of joint moments and powers by inverse dynamics. Gaits which have a lot of non-planar motion will not be adequately modelled, of course, and there will consequently be some inaccuracies. Unfortunately, 3D methods are still rather time-consuming with current technology, though this situation is expected to improve. Methods for estimating the true joint centres (in contrast to the assumption that they lie beneath skin markers) are also improving.
The intention here is to provide the most basic type of analysis which everyone can understand, so as to facilitate discussion of the cases, and the common problems encountered in clinical gait analysis. The videos are useful in themselves for observational analysis, but note that occassionally it is desirable to keep the light level low to avoid identification of the patients.
I hope that gait analyses made with more sophisticated systems will also be submitted to CGA, and if a standard can be agreed upon, this will provide an opportunity for assessment of inter-laboratory accuracy.
Format of the data files
Each analysis is derived from two raw kinematic data files (the horizontal, x, and the vertical, y, positions of each marker) and one force-plate record file during a single gait cycle (one stride). The format of the kinematics files is as follows:
Column 1: Toe (5th metatarsal head)
Column 2: Ankle joint (lateral malleolus)
Column 4: Knee-joint
Column 6: Hip-joint (greater trochanter)
Column 7: Trunk (gleno-humeral joint)
Each row represents data from a single frame (picture) of the video, which is recorded at 25 frames per second. Note that columns 3 and 5 are left blank because they represent the shank and thigh segments in subsequent analyses of the raw data. In the analysis these raw coordinates were filtered at the 6th harmonic of the stride period.
The format of the force-plate recordings is as follows:
Column 1: Trigger
Column 2: Blank
Column 3: Mediolateral force
Column 4: Anteroposterior force
Column 5: Load
Column 6: Sagittal-plane Moment (about mediolateral axis)
Column 7: Frontal-plane Moment (about anteroposterior axis)
Column 8: Transverse-plane Moment (about vertical axis)
Note that the following 6 columns are the recordings from a second force plate, which is rarely used.
Anthropometry has been obtained from David Winter's book, Biomechanics & Motor Control of Human Movement, and normalisation of the kinematic and kinetic variables calculated has been done by the method outlined in the FAQ.
Like to know more, or criticise my methods? Email me! 
kirtley@cua.edu
Back to Clinical Gait Analysis home page