As far as I know, there are some methods for this purpose, such as UCM, GEM, and TNC. However, because these methods are complex, please let me know, are there a more simple methods for this purpose?
Thanks in advance,
Esmaeel.
Maybe you could try analysing through principle component analysis (pca). This is also heavy on the mathematics but thisxcan even be done in spss.
I'd suggest you to try to replicate Temprado et al. (1997) Human Movement Science, 16(5), 653-676. In this paper, rooted in the dynamical pattern theories (Kelso, 1995), researches used video caméras to record limbs trajectories. Then they off-line computed the angular time series of upper arm joints. Knowing the task they used, really similar to yours, and the power of their findings. I guess that it could a useful paper to address your question and maybe to find tricky ways to design your experiment.
GEM is difficult to use but UCM is relatively simple. You should try it. The best instance for your study is to use the "Clint Eastwood quick-draw shoot-out." article from JP Scholz, G Schöner, ML Latash.
I recommend you take notice of a formal approach introduced by Hermann Mueller and Dagmar Sternad a couple of years ago. See e.g.
Müller, H., & Sternad, D. (2003). A randomization method for the calculation
of covariation in multiple nonlinear relations: Illustrated with the
example of goal-directed movements. Biological Cybernetics, 39, 22–
33.
Müller, H., & Sternad, D. (2004a). Accuracy and variability in goaloriented
movements: Decomposition of gender differences in children.
Journal of Human Kinetics, 12, 31–50.
Müller, H., & Sternad, D. (2004b). Decomposition of variability in the
execution of goal-oriented tasks: Three components of skill improvement.
Journal of Experimental Psychology: Human Perception and
Performance, 30, 212–233.
Müller, H., & Sternad, D. (2007). Variability and learning. In D. Sternad
(Ed.), Progress in motor control: A multidisciplinary perspective (pp.
130–161). New York: Springer.
Their analytical model requires some mathematics, too, but it specifically addresses your problem stated above, and has been successully applied to tasks like dart throwing, basketball free shots, and boole.
For an intro, I'd specifically recommend to start with Müller & Sternad, 2004b, where they try to disentangle movement related variability components like noise, covariation (of, e.g., release velocity and release angel in dart throwing), and equifinality of movement variations with respect to goal achievement.
Since you have a discrete task (if you consider the success of shooting as goal and the release position, orientation and velocity as body level states) the GEM will be the best approach. I agree with Klaus Blischke that the CR (correlation by randomization) approach introduced by Muller also will be a good choice. but the sensitivity analysis an generality of GEM makes it the best choice. maybe it seems a little complex but really it is not that complex. I re command you to take a look at recent paper of Joseph Cusumano and Jonathan Dingwell. " Movement variability near goal equivalent manifolds: Fluctuation, control, and model-based analysis". specially section 4.4, general experimental hypotheses. they are really interesting. although some of them are not testable for your task (like DFA analysis that needs quite large number of trials) you can give some help from engineering groups working on the subject in Iran.
the UCM analysis will be easier to run. It is appropriate for analyzing the continues task of throwing too (in a identical initial position until the ball release moment).
the more easier approach will be PCA which you can run in MATLAB with a single function like princomp.
Generally it depends on your hypotheses and level of analysis and interpretation you want to have.
My question would be "WHY WOULD YOU WANT TO REMOVE VARIABILITY??" its inherent in all movement and useful!!
See dynamical systems and uncontrolled manifold theory for motor learning approaches.
See all animals and humans in efficient motion for a more 'global'/holistic answer
Thanks Harvey for your response but I don't want to remove variability but I want to decompose that for more analysis.
Yes, I would like to decompose variability into simpler compounds for more analysis and find relationship between body level and goal level.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics