Some comments on 'A Joint Coordinate System for the Ankle
Complex', a proposal submitted to the ISB standardization and
terminology committee (ISB Newsletter 59, pp.6-8).

First of all, I admire the courage of the ankle joint
subcommittee for stepping into this minefield of standardization.
It is always much harder to construct such a proposal than to
criticize it. So, please take these comments as constructive
criticism, with the intention to improve the final outcome of
the process.

1. Anatomical landmarks and anatomical planes
Of the anatomical axes defined in the proposal, only the X-axis
of the tibia/fibula and the z-axis of the calcaneus are actually
used in the joint coordinate system (JCS). The other landmarks
are only required for definition of the neutral configuration.
For this reason, considerable time and effort can be saved in the
data collection protocol when another definition of neutral
(e.g. a reproducible standing position) is acceptable. This, of
course, depends on the purpose of the analysis. The analysis may
become unnecessarily complex when using this standard.

Another point is that the proposed definitions of neutral are
possibly subjective and sensitive to errors in locating the
anatomical landmarks. Specifically:

"plantar aspect of the foot". Is that the same as the ground
surface when the person is standing (full, or half body weight?),
or a plane defined by three anatomical landmarks (which points on
the calcaneus?).

"projection in the transverse plane of line from M5 to O1".
First, why not just use the y-axis of the tibia, as defined
earlier? Second, this projection is very short, which makes the
orientation quite sensitive to errors. This is probably not a
very reproducible definition of neutral for internal/external

2. Choice of rotation sequence
The proposed rotation sequence is (at least in its terminology)
inspired by the widely used standard of Grood and Suntay (1983)
for the knee joint. However, some arguments can be made for an
alternative choice in the ankle, as proposed by Cole et
al. (1993):

(i) If quantification of rotation of the distal segment about its
*long* axis (rather than its "vertical" axis; which is never
really vertical anyway during movement) is found useful (as in
the Grood/Suntay standard for the knee), the rotation sequence
for the ankle complex should be: plantar/dorsiflexion -
internal/external rotation- in/eversion, with the latter axis
fixed in the calcaneus.

(ii) The orientation of the long axis is definitely less
sensitive to errors in locating anatomical landmarks than the
orientation of a (short) "vertical" axis. In the proposal, the
orientation of the z-axis of the calcaneus depends completely on
the definition of neutral, i.e. it is perpendicular to a not very
well defined "plantar aspect" of the foot (see comments above).

(iii) Historically, the horizontal projection of the long (or
A-P) axis of the foot, has been defined by Inman (1976) as the
line from the midpoint of the posterior aspect of the calcaneus
to the gap between the second and third toe. The sagittal
projection of this line still requires some definition of the
"plantar aspect", (the axis can be defined to be parallel to the
plantar aspect) but at least one projection is well defined. In
my opinion, this makes this axis more reliable and more suitable
for use in a joint coordinate system than the z-axis as proposed.

3. Choice of XYZ
The choice of which axis to label as X, Y and Z is somewhat at
odds with the ISB proposal for a global reference frame. It
would be desirable to have global and local reference frames
roughly aligned during a neutral position. This allows easier
interpretation of absolute segment orientation (small angles
corresponds to close to neutral), and less risk for
singularities. If fact, if I'm not mistaken, with the foot in
neutral position, the ISB proposal for absolute orientation can't
be used with this calcaneus reference frame! The absolute
orientation matrix involves a rotation of close to 90 degrees
about the Y-axis (angle beta = 90 degrees, see Part 4 of ISB
proposal), which leads to division by zero when attempting to
calculate the other two angles alfa and gamma.

Rather than change the calcaneus reference frame, I suggest that
this is another reason to reconsider the proposed standard for
the global XYZ directions.

4. Practical use of the standard
The proposal does not make any suggestions for practical
implementation of the standard. Is it proposed that markers are
attached to all of the proposed anatomical landmarks, or are
those landmarks only used during a 'subject calibration'? Some
practical advice may be useful.

5. The use of actual joint axes in a JCS
Movement in the ankle joint complex occurs mainly as rotations in
two joints: the talocrural joint and the subtalar joint. In
unloaded conditions, it has been shown that these joints can be
described as hinge joints with one degree of freedom. It may
therefore be desirable, for some applications, to choose a joint
coordinate system (JCS) in which the tibia-fixed axis coincides
with the talo-crural joint (as in the proposal), and the
calcaneus-fixed axis coincides with the subtalar joint. The
rotation about the floating axis would then become very small,
and a more meaningful interpretation of the other two rotations
becomes possible (Scott and Winter, 1991). Another big bonus of
such a JCS is that joint moments with respect to these axes can
be directly related to muscle forces. This is not without
problems, however. Hontalas and Williams (1995) conclude:
"... the assessment of motion about the subtalar joint for
individuals may not be accurate until valid in vivo methods are
developed for determining subtalar axis orientation."

The subtalar axis orientation can be determined in vivo (Bogert
et al., 1994), but this is not an easy protocol, and the result
is quite sensitive to relative movement between markers and bone.
Therefore I personally feel that a JCS based on anatomical
landmarks, not considering joint axes, is at this point more
practical, and I agree with the Ankle Subcommittee that this
should be standardized. My point is, that it may sometimes be
advantageous to go through a complex procedure to determine a
non-standard JCS which is more appropriate for the purpose of the
analysis. It is therefore important, as already mentioned on
Biomch-L by Cavanagh and Yeadon, that standards are not "cast in
stone" and blindly enforced.


Bogert, A.J. van den, G.D. Smith, and B.M. Nigg (1994) In vivo
determination of the anatomical axes of the ankle joint com-
ples: an optimization approach. J. Biomech. 27,1477-1488.
Cole, G.K., B.M. Nigg, J.L. Ronsky, and M.R. Yeadon (1993) Appli-
cation of the joint coordinate system to three-dimensional
joint attitude and movement representation: A standardiza-
tion proposal. J. Biomech. Engng. 115,344-349.
Grood, E.S. and W.J. Suntay (1983) A joint coordinate system for
the clinical description of three-dimensional motions: ap-
plication to the knee. J. Biomechanical Engng. 105,136-144.
Hontalas, K.L. and K.R. Williams (1995) Talocrural and talocal-
caneal joint rotations during the stance phase of high
heeled gait. Proc. 19th ASB Congress, Stanford CA, pp.223-224.
Inman, V.T. (1976) The Joints of the Ankle. Williams & Wilkins,
Scott, S.H. and D.A. Winter (1991) Talocrural and talocalcaneal
joint kinematics and kinetics during the stance phase of
walking. J. Biomechanics 24,743-752

-- Ton van den Bogert
Human Performance Laboratory
Faculty of Kinesiology
University of Calgary, Canada