View Full Version : Re: Marker-set independent gait analysis

07-06-2007, 06:28 AM
I would like to thank Nancy and the respondents for initiating a fruitful
and overdue discussion on a very clinically relevant topic.

My two cents:

Technically, a marker set is really just a reference system to which joint
centers and bone segments are defined. Far too often, I hear marker sets
being inappropriately referred to as biomechanical models when in fact they
make up only a part of the "model." In its simplest form, a model used to
measure kinematics from surface markers include:

1) Reference Systems: local "intermediate" coordinate systems to which joint
centers and subsequent segments are defined. This is the so-called "marker
set" and the major differences between sets involve how these reference
systems are defined (ie, wands in Helen Hayes, triads in Cleveland Clinic)
and their susceptibility to surface motion (Castagno et al. Gait & Posture,
3(2), 87, 1995).

2) Joint Centers and Bone Coordinate Systems: locations of joint
centers/axes with reference to marker-based coordinate systems based on
anthropometric ratios (standard in most gait analysis software) or
functional determination (ie, sphere fit); bone segments are then defined
between joint locations and tracked using the attached markers

3) Angle Conventions: calculations used to estimate joint angles based on
movement of the marker-defined bone segments (ie, Euler, JCS, etc).

Thus two labs that use the SAME marker set (ie, Helen Hayes, CCF) can in
fact output different results if the joint/segment definitions and/or angle
conventions used in their software are different. These definitions make up
the heart of a kinematic model, regardless of the marker set used to define

In most cases, the users have little control on how the "black box" software
implement these estimations although I'm discovering that newer versions
allow users to input more subject-specific parameters (ie, functional hip
center in MAC's Orthotrak). As these functions become more readily
available, it is only fitting (pardon the pun) that more clinicians use them
as they are relatively straightforward to implement.

For the hip center, we reported that it doesn't take a significant amount of
ROM to functionally determine its location in patients with hip pathologies,
where standard anthropometric ratios no longer apply (Aguinaldo et al, GCMAS
2003). However, be careful when selecting a marker trajectory for fitting
the sphere as it's been shown that certain optimization methods are more
sensitive to random noise (Hicks and Richards, Gait & Posture, 22(2),
138-45, 2005).

Bottomline, it really shouldn't matter what marker set is being used as long
as the other elements of the kinematic model are known and ideally can be
modified or controlled depending on the patient population. I understand
this can be a nightmare for those of us pushing for standardization but I
also believe we shouldn't blindly accept the current "standards" when the
tools to improve the kinematic engine are there for us to use.

Arnel Aguinaldo, MA, ATC
Director, Center for Human Performance
3020 Childrens Way 5054
San Diego, CA 92123

Assistant Professor, Biomechanics
Department of Exercise and Nutritional Sciences
San Diego State University