Basic Guidelines for Balance Measurements and Setups

(1)

guidelines_balance.sdw 08/28/02 1/3

Basic Guidelines for Balance Measurements and Setups

The following guidelines are primarily based on the recommendations by the l'Association Française de Posturologie (AFP).

Quiet standing. The feet should be placed with heels 2 cm apart while the feet form a 30 degrees angle.

The markers for the feet should be painted on the force plate. The surface should be smooth and hard. During the quiet standing the subject should keep her arms on the side of the body in a relaxed state with a distance to the visual target of 1 m. The visual field should be a white wall, or a white screen. French reference data are based on a measurement time of 51.2 seconds. (Originally coming from using a sampling rate of 5 S/s and a total number of 256 points per measurement.) An example of the instructions given to a subject, as quoted from [Gagey & Weber 1999: 60]:

Vous restez debout, immobile, décontracté, ce n'est pas le garde à vous, les bras le long de corps, vous regardez dans la direction de la cible visuelle, sans la fixer, et vous comptez lentement et à voix haute jusqu'à ce qu'on vous dise que c'est fini.

Visualizing data. The x-axis of the balance plate is taken to be the medial-lateral direction (left to right), whereas the y-axis is taken to be the posterior-anterior direction (back to forward). The statokinésigramme [posturogram?] shows the trace of the center of pressure (COP) in the xy- coordinate system. The stabilograms are the graphs of x and y as a function of time. Coordinates are expressed in units of mm. A basic analysis tool is the FFT and the power spectra of x(t) and y(t).

The power spectrum is graphed using a logarithmic (dB) scale according to

A( fk)=20⋅log

(

^max^{|ˆx( f}^{|ˆx( f )|}^k^)|

)

which is a number ≤ 0 ("x hat" denotes the Fourier transform of x, etc.). The frequency band of interest in posturology is limited to 0 – 10 Hz; thus the data sampling rate must be at least 20 S/s.

Also of interest are the graphs of the autocorrelations (x-x), (y-y) and the intercoorrelation (x-y).

Frank Borg (borgbros@netti.fi) Chydenius Institute

30º

2 cm y

x

(2)

Other graphs of importance are the histograms of the x and y-coordinates. From the data the velocities dx/dt and dy/dt are also easily calculated. The rms-values of the velocities have been used as parameters as well as the average velocities.

Basic parameters. From the xy-data we can compute a number of basic parameters.

S, surface: the area of the 90% confidence ellipse of the (x,y)-points in the posturogram. Some statistics for the surface S for the French reference data is given in the following table:

Surface (mm²) EO EC

Average 91 225

Lower 95% limit 39 79

Upper 95% limit 210 638

L, length: the length of the posturogram trace. (LFS is a length measure which is the length L divided by the average L obtained for a population. To be replaced by entropy measures?) Due to the fractal geometry of the posturogram the calculated length depends of the time interval between the sampled points. The smaller time interval (higher sampling rate) the longer the calculated length will be (this is the basis of the fractal analysis of the posturogram). The normed length data is supposedly based on using ∆t = 1/5 sec.

QR, quotient du Romberg: defined as QR=100⋅S_ec

S_eo

where Seo is the surface S for a measurement with eyes open, and Sec is the surface S for a measurement with eyes closed. The measurements are to be performed in the order EO and EC.

Some statistics for the QR for the French reference data is given in the following table:

QR

Average 250

Upper 95% limit 677

Lower 95% limit 112

QP, quotient plantaire: defined as QP=100⋅S_m

S_d

where Sdis the surface for a measurement with a hard sole, Smis the surface for a measurement with

(3)

a soft sole (sol mousse: Orthomic 3 mm, density 340 kg m^-3 and hardness of Shore = 37).

AN∅2, amplitude normalisée des oscillations posturales dans le bande de fréquence 0.2 Hz: a parameter that supposedly expresses "en amplitude, le pourcentage des oscillations de la masse corporelle qui partagent le rythme de la ventilation" [Gagey & Weber 1999: 63], and defined by

AN∅2=

∫

0.16 Hz 0.24 Hz

| ˆy ( f )|df

0.04 Hz

∫

0.14 Hz

| ˆy( f )|df 

∫

0.26 Hz 0.6 Hz

| ˆy ( f )|df

This parameter is calculated both for x-data and y-data. AN∅2 is claimed to be a stabile parameter in the sense that it is hard to affect voluntarily.

AN∅∅∅∅2 (%) Along x-axis Along y-axis

Average 95%-limits Average 95%-limits EO 11.39±6.95 -17.89/+17.55 8.36±4.86 -10.55/+11.25 EC 16.57±10.41 -22.34/+21.88 14.65±7.98 -17.46/+21.16

Accuracy of the force plate. The AFP recommendations require that the force plate should measure the COP cooridinates to 1 mm accuracy or better within a 10 cm radius of the reference coordinate point of the force plate. This accuracy can be assessed by putting a weight (10 - 40 kg) and shifting its position on the force plate and compare the positions to the software output. The noise is measured but putting a weight on the force plate and measring the x- and y-displacements; this should be less than 1 mm. The natural frequencies of the force plate should be in excess of 20 Hz as not to interfere with the posturogram spectrum (with range of 0 – 10 Hz). Linearity of the force transducers should be 0.1 % or better, and hysteresis should be less than 0.2 %. The drift should not exceed 1 mm / 30 min.

References:

1. Gagey P.-M., Weber B.: Posturologie. Regulation et Dérèglements de la Station Debout. 2e édition. Masson 1999.

2. Browne J., O'Hare N.: "Recette de plates-forme de force". <http://perso.club-internet.fr/

pmgagey/Browne.htm> . Physiol. Meas. 21 (2000) 515 – 524.