Visual Contribution in Human Postural Control

Abstract: An important function for the human body is the ability to maintain balance. This has been compared to a sixth sense whose function we would have difficulties to be without. The system of balance is a complex system which uses several different sensors to get a satisfactory image of the position and motion of the body. One of the sensors is vision. Our assignment has been to make an equipment for stimulating the system of balance visually and to study the effect of that simulation.

The stimulus consists of a vertical patterned screen rotating horizontally. The speed of the screen is controlled by a computer and a number of motion patterns can be programmed. The movements of the test subject are measured by a force platform.

We have found that the equipment provides visual postural stimulation leading to moment responses and also that the test subjects feel that their balance is affected. It is likely that other receptor systems give contribution even during powerful visual domination. However, nonparametric identification shows high coherence in frequency ranges around 0.05-0.2Hz. The results from the analyzes within this range can therefore be considered significant. The analysis within the above frequency range shows an even proportionality between movements of the surroundings and ankle torque, almost independent of frequency, but the phase may vary with the properties of the stimuli. With a pure sinusoidal stimulus the phase shift was relatively stable at about 180 degrees, while the other test sequence used (such as the PRBS-stimulus) showed a frequency dependent phase shift starting at 0, 90 or 180 degrees. Parametric identification exhibits heavy variations in the time delay between stimulating input signal and response in form of moment changes. Especially active motion patterns with frequent transient elements have short time delays, but continuous motion patterns, like the sinusoid, prolongs the response up to 2-3 s after a while. At occasional times a delay of up to 6 s could be recorded. It is possible to find an ARMAX-model being a good estimation of the real system, giving low residual and correlation values.