Bienvenue dans Adobe GoLive 6

SAMPLE FREQUENCY

in clinical stabilometry.

What standardized rate ?

Pierre-Marie GAGEY

At the last ISPGR meeting in Bologna, Bastiaan Bloem****, ISPGR President, has decided to form the Standardization Committee for Clinical Stabilometry with Fabio Scoppa and Kimberly Edginton Bigelow, as Committee Co-Chairs and Lorenzo Chiari, as ISPGR Board Member Representative for Committee.

A discussion about all the problems raised by this standardization (specifications of manufactured force plates, paradigms to calculations of reported postural sway parameters, etc.) will start on July 15, 2011 and it is hopped that we can have a consensus meeting in June 2012, at the ISPGR meeting in Norway... We must not lose time! That's why I think we can start discussing now between us, before discussing with the Committee. And I propose this discussion on sample frequency.

As Maurice OUAKNINE says : "The ideal would be an infinite rate !" Unfortunately that is impossible. Therefore all chosen rate won't ever be the ideal rate...

Metrological characteristics of the gauges, as the 'maximum permissible errors' (5.21 VIM * ISO 1999) of assessment of the distances between two successive sampled positions of the center of pressure, are not a basis of decision. Indeed:

- If one takes a rate that takes into account these characteristics so that one doesn't collect any random value, then one makes a systematic error, impossible to avoid, that comes from neglecting everything that is not measured.

- If one takes a rate superior to the metrological characteristics of the gauges, then one makes a systematic error that comes from all random values.

- However one can show that these two unavoidable mistakes are practically equivalent.

The insufficiencies of precision of the statistical parameters (X-means, Y-means, Ellipsis of position of center of pressure or of center of gravity) don't permit to fix a rate. The calculations have been made between a rate of 5Hz and a rate of 40HZ, the inaccuracy doesn't exceed some for cents**.

The insufficiencies of precision on the amplitude spectrum don't permit to fix a rate. The calculations have been made between a rate of 5Hz and a rate of 40HZ, the inaccuracy doesn't exceed some for cents**.

The insufficiencies of precision of the dynamic parameters (Speed, Acceleration) depend on the sampling rate, of course. Therefore, they could help to fix a rate. Unfortunately these insufficiencies are insufficiently studied nowadays!!!

The insufficiencies of precision of the nonlinear dynamic parameters (prediction power, Lyapounov coefficient, Le Van Quyen, etc. ) are again insufficiently studied.

We must recognize that, at least for the moment, we are unable to fix a sampling rate on rational data bound to signal processing and/or analysis.

That gives a real weight to the convenient arguments: let's choose a rate coherent with the conditions of clinical practice. Stabilometric examination, USELESS to find the right postural treatment, but nevertheless PRECIOUS, must remain POSSIBLE, even in these ambivalent conditions. That is to say neither too long, nor too expensive.

These kinds of arguments only of convenience guided our choice in 1985 in favor of the II Apple * * * against the Lisa... And I have the weakness to think that this choice made the success of the standardization because it was applicable.

BUT... it doesn't tell us what rate! 40? 80? 100? 500? More? The "builders" probably have an idea on the kind of material that is "marketable."...

** Gagey PM, Baudin B., Bizzo G., Scheibel A. Weber B. Should we increase the sampling cadence in stabilometry ? http://pierremarie.gagey.perso.sfr.fr/Cadence-a.htm

** Gagey P.M. Ouaknine M., Sasaki OH. (2001) to show the dynamics of the stabilization the AFP40 platform /16. In M. Lacour (Ed.) Stance and balance. Novelties 2001, conceptual, instrumental and clinical. Solal, Marseille, 95 -106.

*** Apple II cost about 10.000,00 Frs at the time (1983), and Lisa 100.000,00 Frs.