Paul Ostic
01-10-2002, 03:21 AM
Thank you to Prof. Dr. Hatze for inviting comments from all. I am new to
the field of biomechanics, but I have significant experience in other
scientific areas. I am brave and foolish enough to tender my comments.

First, let me introduce myself: I am a graduate student at Queen's
University in Kingston Canada, working with the Human Mobility Research
Centre. I have undergraduate degrees in both mechanical engineering and
chemical physics, in addition to industrial experience with nylon
polymerisation and yarn production. My current program is related to
computer enhanced wrist fracture fixation. Working with a
multidisciplinary group and being at a very initial stage in my research
have allowed me to explore widely and freely.

I have a grand, and perhaps naive vision. I believe that the time has
arrived to collect all the known data and accepted anatomical and
functional models and incorporate them into a common framework that would
describe and model the entire human musculoskeletal system. The model
should be perpetually maintained and amended as new findings are accepted
by the biomechanics community. Gaps in understanding will become readily
apparent when the whole model is assembled, and perceived gaps may be
solved as soon as all the existing knowledge in incorporated into a
common framework.

The model would contain the bones, muscles, cartilage, kinematic
constraints, reflex loops, physical properties of muscles and connective
tissues, and might even model the changes in muscle performance as a
function of chemical concentrations within the body. Those wishing to
use the model would be able to select the level of detail required in
their simulations. The model will be scalable; a few key measurements
would allow the model to be quickly adjusted to the stature of a given

The anatomical model will be easy to generate, and likely already
exists. There are enough X-ray CT and MRI files in existence to be able
to create a really good anatomical model including soft tissues. I know
of at least one model for the hand and wrist that is commercially
available (The Interactive Hand; Primal Pictures 2000)

Customizing or scaling the anatomical model to the individual is a
relatively easy, and more or less solved problem. The computer enhanced
image guided surgery field has been using "atlas" models of bones for
some time. This process uses several linear measurements to scale a
standard 3D model of a bone to the size and shape of the patient's bone.
Ultrasound, x-ray, or even callipers would suffice for the linear
measurements so that those without capabilities to generate 3D models of
their subjects could have access to 3D models.

Adding kinematic constraints is perhaps a little more difficult, but is
within grasp; note all the discussion regarding gait analysis. The model
would have several levels of detail; at the most basic level of detail,
the joint kinematics might be modelled as having only one or two
rotational degrees of freedom and no translation. As more detailed
understanding on joint function is available, it would be added to the
model until each joint has a full six degrees of freedom. Users might
still prefer to use the simplified form to allow faster computations.
Ultimately, if the properties and functions of muscles, ligaments and
cartilage are known well enough, the kinematics of the joint may be
generated by the model and would match the kinematics that have been
observed and documented already.

Adding reflex loops may seem far fetched, but is also within reach. For
example, Solomonow et al (The Ligamento-Muscular Stabilizing System of
the Spine, SPINE vol. 23 num 23 pp 2552-2562) show a relationship between
tension in a ligament and muscle activity. There may not be enough
quantitative reflex data to incorporate into a functional model yet, but
having it linked to the model would provide an effective way of
cataloguing the data that are available currently, and quickly verifying
whether new data is consistent with the current model.

Ideally, the model would have enough pegs on it to hang all the
information that is currently available about human biomechanics. There
would be a place to include the relative motion between skin and skeletal
landmarks. The user could select whether to treat cartilage as rigid,
elastic, or visco-elastic. The scope of what could be included is beyond
my imagination.

Imagine having a 3D interactive anatomical model that you could display
on your computer screen, and use the mouse to drag the limbs to a
particular position. There would be an interface where you'd put in a
handful of key measurements; length of tibia, width of proximal tibia,
subject height and mass, skin fold thickness and grip strength, for
example, and the computer model would match your subject to within a
certain degree. More measurements would yield a more accurate
correlation. Now, click your mouse on a joint and the computer will tell
you the forces at that joint if the subject were to remain static in the
position you moused the model into. Click on a muscle, and the computer
will tell you the rate of oxygen and nutrient consumption and the changes
in concentration of metabolic products. It might even tell you how long
until the muscle fatigues. This data would be compared with the EMG
frequency parameters that you just collected in your clinical study on
muscle fatigue.

Imagine a high performance athlete with a chronic overuse injury; you
could record his/her technique with a motion tracking system and skin
markers. Based on the same measurements used for the fatigue study
mentioned above, the model would translate the skin marker coordinates
into musculoskeletal movements. Then the forces and loads in the area of
the injury could be analysed. Subtle changes in technique to reduce
loading on the injured tissue could be analysed using the model. The
athlete could then be retrained to the modified technique. The model
would allow you to approach the question in two ways; "If we need to
keep the load below X, what is the restricted speed of the motion?" or
"If we restrict the speed of the motion to Y, what will be the new load?"

I'm certain that each of us, at an early stage in his/her career, has
thought about the idea of a single unifying model. Unfortunately, most
of us have become wise enough to realize that it is too complicated a
system to put into a model. And even if we did, the model would be too
big and cumbersome to use. But maybe the current state of our
understanding and computer technology are sufficient to allow us to build
the first approximation of the ideal model.

A grand new model should not be seen as a threat. It's existence would
not replace or diminish the value of the simple and efficient models we
have now. Newtonian mechanics are still the mainstay of physics, even
though it is accepted that relativity is more complete.

Paul Ostic
MSc candidate
Queen's University
Kingston, Canada.

To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l