No announcement yet.

Re: BioNet: Problems in myoskeletal inverse dynamics

This topic is closed.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Re: BioNet: Problems in myoskeletal inverse dynamics

    __On models and supermodels__
    The discussion started by Hatze and Leardini is getting interesting. Three
    messages in particular got my attention -- and forced me to think -- those
    posted by Georg Duda, Paul Ostric and Ilse Jonkers. I may get back on
    Duda's positions in another message with more detail, but here I would like
    to spend my two cents on the concept of modelling the reality.

    So, first of all this will not be about those gorgeous ladies that fly
    around the world showing their kilometric legs as the title may suggest.
    This is about how, in my humble opinion, we describing the physical reality
    around us. Now that I dug my grave deep enough let me jump in it.

    The physical reality, the nature, is infinitely complex. The human mind is
    not able to understand it as it is, because it is too complicated. But the
    humanity wanted to understand the nature. Why? In the beginning I bet it
    was to provide us the ability to predict natural events. When is going to
    rain? How that building may collapse? What is the best time to harvest
    my crops? The man who knew these answers could pretend to control the
    nature. This sense of control is in my opinion a very basic instinct for a
    human been, and we feel pleasure when we achieve it. So after a while we
    started to search for this pleasure with no practical reason, just for the
    pleasure we got from that. The pleasure of knowing was born.

    What all this has to do with musculo-skeletal modelling? Well, in my
    opinion from the very beginning of human history we always tried to
    understand (which for me means be able to predict) nature for two reasons:
    to solve practical problems and because we just enjoyed it. I call the
    first engineering and the second physics. You may call it John and Mary if
    you like. I am not interesting in the disputes of the names of science
    branches. The point is that we all investigate (model) nature for two very
    different reasons.

    You may disagree with the some of the speculations I made, but I am quite
    sure most of you will see these two different purposes: engineering and
    physics. Now the provocations. I believe that:

    __We do not understand nature, only the models we made of it__
    Every time we try to understand nature, in practice we are collecting a
    limited number of observations, use them to create a model of the nature,
    check it is able to predict correctly, and then use it for our purposes.
    Thus, modelling is the ONLY way we use to investigate nature. The model
    may not be explicit, but it is always there. The reason is that we can
    understand only our models of the nature, not the nature itself.

    __Every time we make a model we must state the purpose__
    Sticking with my definitions every time we say model we should state if it
    is an engineering model or a physics model. With the first we aim to
    predict a natural phenomenon in one very specific instance. With the second
    we aim to predict a natural phenomenon in all (or most) of its instances.
    Useless to say results of one exercise can be reused in the other one, so
    they are very contiguous and the boundaries sometime tend to blur. However,
    when is time to talk about models, we need to agree on their purpose.

    __Good engineering models MUST be limited in scope__
    Physics models of the musculo-skeletal apparatus, as I defined them, seem a
    huge challenge. However, if you young hearts want to go, fine, the road is
    given: it is called scientific method. So, forget about all those excuses
    we have in biomedical sciences, such as uncontrollable variability,
    impossibility to measure directly most parameters, etc. Engineers must
    cope with them, but you have to solve them. Because you are looking for a
    supermodel, one that can predict most instances of the natural phenomenon
    we call musculo-skeletal model. For the rest of us, those who are a
    looking for the ability to predict the behaviour of the musculo-skeletal
    model in one specific instance defined by a practical purpose, engineering
    models must be limited in scope. And not because our computer are not
    powerful enough but because using a physics approach to solve an
    engineering problem is wrong.

    Please understand that I am not making any hierarchy here. Without
    scientists doing research just for the joy of knowing more, we would simply
    have no science. We call it basic research, sometime, and it is a pity to
    see how little money is available for it in these days. And without
    engineers, who solve problems before all the answers are available,
    technology would not grow so quickly. As research engineers many of us had
    the joy to switch between these two roles in their carriers, and it is pure
    fun. But please, let us do not confuse them.

    In a philosophical mood,

    Marco Viceconti

    Laboratorio di Tecnologia Medica tel. 39-051-6366865
    Istituti Ortopedici Rizzoli fax. 39-051-6366863
    via di barbiano 1/10, 40136 - Bologna, Italy

    Tiger! Tiger! Burning bright in the forest of the night,
    what immortal hand or eye could frame thy fearful symmetry?
    Opinions expressed here do not necessarly reflect those of my employer

    To unsubscribe send SIGNOFF BIOMCH-L to
    For information and archives: