Measure in Posturology

By Pierre-Marie GAGEY* and Guy BIZZO**

* Institut de Posturologie, Paris

** Bureau National de Métrologie, Paris

(4 Janvier 2001)


Aristotle (330 BC) had already understood that the position of the parts of the body relatively to one another as well as their position relatively to the environment, that is to say body posture in its two perspectives, is the expression of superior activities. Among all those activities, Newton's works have focused the Occidentals' attention on those concerning mechanics, the balance of forces, man's struggle against gravity. That field of research has proved fertile, especially as far as measures are concerned - which is what this article will try and explore. Yet it should be borne in mind that posture and balance only represent one aspect of posturology. The Orientals have remarkably succeeded in using posture to express emotion, notably in No theater, and some speech therapists (Grini et al., 1999) nowadays measure the respective position of certain skeletal parts in order to observe vocal expression better.

In any case, posture always leads to what underlies it, and measure, in posturology, is of interest only in reference to the superior activities that posture reveals. No postural measures are possible without a postural concept... That obliges us to propose a few deviations from pure metrology. In order to present measure in posturology, we first have to answer the following question: what exactly do posturologists measure?

Towards the concept of postural system

    When he drew, in his De Motu Animalium, the vertical of gravity in the human body, Borelli (1679) suggested that the laws of mechanics do not exclusively apply to celestial bodies. From that moment, posture and balance have been linked in the mind of men, but it was only in the early XIXth century that Charles Bell (1837) raised the question of postural control: «Obviously, man has a sense through which he knows the bending of his body and he has the ability to readjust it and correct any deviation from the vertical. What could that sense be?» A few years later, Karl von Vierordt (1860) rectified Charles Bell's question as he noticed that postural control is not the effect of one sense but of a whole series of sensitivo-sensorial data: visual, tactile from the sole of the feet, proprioceptives. And Vierordt then had the intuition that the recordings of the posture of man standing at rest are likely to inform us on the functioning of what is not yet called postural system but already exists as a nearly complete concept - he then recorded the very first stabilometric signals.

The actual presentation of the postural system is nothing more than the translation into contemporary language of those progressive advances.


The postural system

Control in retroaction

The human body is mechanically unstable as its mass center is situated above its center of pressure on the ground. As soon as the resultant of the forces of gravity is no more in line with the forces of reaction from the ground, a torque is created and tends to accelerate the falling of the body. The stabilisation of that mechanically unstable body therefore requires a system of control in retroaction, the Inputs of which have to be able to detect the slightest deviation from the position of balance in order to command as early as possible the appropriate reactions for a return to that position of balance.

The Inputs of the system: Exo- and Endo-inputs.

There are three universally accepted exo-inputs in that system of control: the eye, the vestibular apparatus, the foot, and up to now we do not know any other. As they are directly related to the exterior world, they can pick up directly the movements of the body relatively to the environment. They are called the «exo-inputs» of the postural system. Only sensitivo-sensory organs related to the environment can allow a precise stabilisation of man in his environment.
But the eye turns in the socket whereas the vestibule is stuck in the rock. Therefore the position data given by sight cannot be compared to the position data given by the inner ear if the position of the eye in the socket is unknown to the postural system. The oculomotricity giving such data is therefore a necessary input for the postural system even if it does not have any direct relation to the exterior world - it is an «endo-input» of the postural system.

The same logic goes for the rachis - especially for its two more mobile parts, the cervical and the lumbar - as well as for the joints of the inferior members that give the position of the «plantar exo-sensor» relatively to the «cephalic exo-sensors».

The outputs of the system

The appropriate kinematic reactions for the stabilisation of the human body lead to a mobilisation

* either of the pressure center,
* or of the mass center,

in order to bring them nearer the same vertical of gravity. Now those two strategies are very different. The body mass acts as a low-pass filter, limiting the speed of the shifts of the mass center (the own frequency of the human pendulum is of around 0.3 Hz). The shifts of the pressure center, on the contrary, put in motion considerably less important masses, and can therefore be much quicker, therefore more efficient, and use less energy.

Obviously, the functioning of the postural system is not univocal.

The upright postural control system

Doctors suggest to isolate, under the name of «upright postural controls system», the postural sub-system controlling the stabilisation of man standing at rest. Actually, the patients themselves taught doctors the difference between dynamic postural control and quasi-static postural control: for instance the patients affected by «vestibular neuritis», at some stage, bump into every door frame when walking, whereas when standing at rest they can behave normally.

Now, that notion of an appropriate system of control for the stabilisation of man standing at rest is coherent with certain fundamental data on the output and the inputs of the system.

«The body is made of the superposition of modules (legs, trunk, head, arms). Each module is linked to the underlying module, on which it rests thanks to a group of muscles that have their own central and peripheral regulation, devoted to maintain the position of reference of the module. The righting reflexes examplify that modular organisation (Rademaker, 1931)» (quoted from Massion, 1997). Now, in the standing at rest position, that modular organisation in multiple pendula modifies itself in order to transform the body into a single inverted pendulum (Winter et al., 1997), which reduces the degrees of freedom that the postural system has to control, then speeds up the appearance of motor reactions and eventually betters the stabilisation so much as to make it surprisingly fine. The tactics of the inverted pendulum is coherent with the tactics of mobilisation of the pressure center, together they realise the tactics of the broom - pendulum held, upside down, in balance on a finger tip - which we have already reckoned as the quickest, the more efficient, and the less onerous in terms of energy.

The inputs of the upright postural control system are also specified by their object. The semi-circular canals intervene in the dynamic postural control, but not in the orthostatic postural control (Fitzpatrick & Mccloskey, 1994), the accelerations of postural sway are then inferior to the threshold of perception of those sensors (Gagey & Toupet, 1988). The gain of the neuromuscular spindles is considerably higher when the muscular stretching is approximately as wide as the range of the orthostatic postural sway (Matthews & Stein, 1969).

What do we have to measure in medical practice?

From that concept of postural system - be it upright or not - it is possible to draw up an inventory of the measures which are likely to interest the therapists (Nashner et al., 1982; Gagey & Weber, 1999):


In search for a measuring apparatus

From Vierordt (1860), who set a feather on the head of his subjects in order to inscribe the body sway on a sheet covered with lampblack, to the appearance - a century later - of force platforms coupled to a computer, the measuring attempts in posturology have met the same difficulties: finding an apparatus that

* does not modify the observed phenomenon,
* gives a readable signal.

Because the doodle of Vierordt's first recordings did not lend itself to analysis and the scratchings of his feather on the sheet set on the ceiling gave a complementary datum which, as we now know, modified the phenomenon (Gurfinkel, 1973, a). We have, on that period of search for the best possible apparatus to make measures in posturology, 25 bibliographical references testifying the vigour of that research and the imagination of the researchers. Miles' ataxiameter (1922) deserves to be quoted as it has been, more than others, used and/or more or less copied by other authors (Fearing, 1924; Hellebrandt, 1938).

Towards the end of the 1960's a student engineer at Cambridge's MIT devoted his technology thesis to the study of the postural system considered as a cybernetic system (Nashner, 1970). He then offered to observe that system of control by opening its various loops of retroaction, following the classical technic in engineering. To this purpose, the subject is set standing in an apparatus that brings under the control of the movements of the subject's center of gravity, the movements of his visual environment and/or of the platform on which his feet are set. The technological challenge of such an apparatus is very important, notably in order to obtain response delays from the mechanical systems that are coherent with the rapidity of response of the human central nervous system. The visual environment of the apparatus has to cover the whole visual field and to be situated several centimeters away from the subject's body (freedom of body movements, problem of visual accommodation). Therefore that environment is necessarily made of a big cabin, the movements of which should be, if possible, instantaneous, regular, silent, without sway, without phase shift from postural sway... It is far from certain that such a tool does not modify the phenomenon it observes, and on the contrary it is certain that it has not helped the authors who use it to understand how fine the control of the othostatic posture is. However, Nashner's apparatus deserves to be quoted for it obviously contributed to the advances of knowledge on postural system in the dynamic conditions of the struggle against exterior destabilisations - it is still used and will certainly be used a long time still by the therapists who work in such conditions.

The building, in 1952, of the first French force platform by Professor Scherrer (Ranquet, 1953) cleaned the apparatus of any physical modification of the observed phenomenon. The signal analysis by analogical techniques was still a technological feat, but the appearance of computers eased that anaysis which, from the 1980's, even began to be used in current medical practice thanks to the micro-computers.

The stabilometry platforms

The use of stabilometry platforms soon became widespread in research laboratories - except in the USA where Nashner's apparatus has been used almost exclusively for many years. The models of platforms varied according to their builders who used either sensors of forces (pressure gauges or piezo-electric quartz) or sensors of length (electromagnetic plungers) grouped by four or three under the platform, with one platform for each foot or one platform for both feet, etc.. Invention and innovation were utmost in the 1960's/1970's, and that can easily be understood in the research context that ruled over that period.

The fundamentalists then grouped themselves in an international society of posturography funded in Amsterdam in 1969 and the first congress of which was held in 1971 in Madrid (Cf. Agressologie 1972, volume 13, numbers B & C). The few clinicians who took part in the works of that society tried to make their problem understood: a doctor cannot make a stabilometric recording of his patients before they fall ill!... Whereas a fundamentalist can record his experimental subjects before and after the manipulations he submits them to. The doctors then need statistical norms in order to evaluate their patients in a distribution, but not the fundaamentalists. The doctors' request was heard and a normalisation committee was formed under the leadership of Kapteyn (Kapteyn et al., 1983), but at the Houston congress it became clear that it was too late to propose international building norms for a stabilometry platform - three different firms were already selling different stabilometry platforms in Japan, and we considered it impossible to impose arbitrarily one of them as the international norm.

The platform of the Association Française de Posturologie

During the Houston congress, the members of the Association Française de Posturologie, aware of that international failure, expressed their hope that at least in France, if not in Europe, a normalised clinical stabilometry platform could exist (AFP, 1984). They immediately began the writing of the specifications for the building of a standard platform. Several considerations guided their choices:

* The platform was built for the clinicians, not specifically for the fundamentalists,
* Its cost should allow a wide distribution,
* Its achievements would be limited to the clinical study of what had already been studied in a laboratory.

The published norms (Bizzo et al., 1985) set the characteristics of three pressure gauges, their situation at the summit of an equilateral triangle, the rigidity of the plates, the heighth of the gauges, the sampling of the signal at 5 Hz, etc.

The prototype of this platform was used for a first statistical study of the values and of the repeatability of stabilometric parameters obtained in strictly normalised examination conditions (subject's position, environment, instructions). Unluckily, the complete Normes85 make too big a book (270 pages) to have any hope of distribution through the channel of an editor or a review - it is only available as a photocopied document from the Association Posture et Equilibre (new name for the Association Française de Posturologie) and among builders of normalised platforms (AFP, 1985).

The first series of platforms built in Italy following AFP norms gave the opportunity to have a new statistical analysis of the values of stabilometric parameters obtained in AFP normalised conditions (Guidetti, 1989). The coherence of those two statistical analyses with one another and with international publications was thought satisfying enough so as not to start new studies, at least until now.

[Like the Association Française de Posturologie, Japanese builder «Anima» studied and recently published a statistical work on the values of stabilometric parameters obtained in his configuration, which became, meanwhile, one of the more used in Japan (Imaoka et al., 1997)]

An evolution of the building norms of the AFP platform is currently under study (Gagey et al., 1999; Gagey et al., 2000) because the sampling at 5 Hz is not justified anymore by motives of computer memory, and even if it is still excellent as far as classical stabilometric parameters are concerned (Gagey et al., 1999), it impedes certain analyses, such as the nonlinear dynamic analysis of the stabilometric signal (Gagey & Sasaki, 2000).

What does the stabilometry platform measure?

The stabilometry platform measures, at every sampling instant, the position of the contact point of the reaction forces opposing the shifting of the platform under the impulse of the body mass - in other words, the position of the center of pressure , expressed in a bi-dimensional referential whose plane corresponds with that of the support basis and whose origin is conventionally set at the barycenter of that same basis. That center of pressuredoes nearly never merge with the projection of the subject's center of gravity on the plane of his basis of support because the human body is nearly never in a state of balance, but the center of pressure keeps shifting on both sides of the projection of the center of gravity (Winter et al., 1998; Hugon, 1999) in order to stabilise it, as in the tactics of the broom. The center of pressure therefore shows a quick sway around the slower sway of the mass center. As shown by the frequential analyses statistics (Gagey et al., 1985) and the diffusion analyses (Collins & De Luca, 1993), that quick sway of the center of pressure is not controlled - what is controlled is the slow sway that corresponds to the movements of the mass center (Gurfinkel, 1973, b).

From that signal, it is possible to obtain most of the measures interesting the therapist:

The platform as a measuring instrument

As the stabilometry platform provides numbered evaluations of physical dimensions expressed in the CGS system, it has a potential status of measuring instrument. At the time being, however, no stabilometry platform can really be considered as a measuring instrument because no platform builder provides a study of the uncertainty of his measuring instrument. How certain can we be on the truthfulness of the measures given by those instruments?

The uncertainty, a, of a measuring instrument, is expressed by an undimensional ratio:

a = Di/i

where Di represents the deviation from the true value of a given measure i.

In the absence of thorough studies on the uncertainty of the acquisition chains used in stabilometry, the only piece of information we have is their theoretic resolution, limited to about 0.16 Newtons by the current use of 12 bits digitizers for the reading of a full scale of about 650 N. Knowing that we neglect a series of uncertainties of the measuring instrument (uncertainty of accuracy, of rapidity, of stability, etc.), we can say that the deviation from the true value of a given measure i is, at least, of around 0.16 N, Di = 0.16.

The very approximate expression of the uncertainty of the forces measures thanks to the current chains therefore becomes:

a = 0.16/dF

where dF represents the force variation measured by each of the gauges.

The uncertainty of the stabilometry platforms is therefore very badly known. And we have to admit that the study of the uncertainty of a measuring instrument is a very heavy work for which the builders do not receive any help. Even the mere conformity delivery of a stabilometry platform by an accredited independent laboratory is impossible (at least in France) as no laboratory is accredited by the Cofrac for instruments realised with an assembling of force sensors. «No laboratory is accredited for the delivery of force platforms. Some laboratories are accredited [only] to gauge the force sensors set on those platforms» (Cofrac).

That is why we now think that the stabilometry platform built according to the standards of the AFP, which samples its measures fives times a second and digitizes them on 12 bits, with a required accuracy of about one thousandth of the measure range, represents a measuring instrument imprecise enough so as not to demand more than a regular check of the good functioning of the pressure gauges - for instance a display of the weight of the platform skirt during the gauging preceding each measuring session (that display has been suppressed by some builders, which is a mistake).

But when stabilometry platforms with a required accuracy of about a ten thousandth of the measure range are built and sold, it will be necessary to check, not only at delivery but still regularly, the uncertainty of those new platforms (Browne & O'Hare, 2000).


Fundamental research made a lot for the development of measure in posturology. Clinical research showed that such measures in posturology are not only useful but compulsory. Let us hope that applied research now intervenes so that posturologists can have more and more performant measuring instruments.