A New Look at Posturographic Analysis in the Clinical Context: Sway-Density versus Other Parameterization Techniques

In order to identify useful guidelines for the clinical practitioner as regards the use of static posturographic analysis, we collected a set of posturograms from 3 groups of participants (normal participants, Parkinsonian patients, and osteo-porotic patients), according to the Romberg test. From each posturogram, we extracted global parameters (in the time domain and frequency domain) and structural parameters (based on diffusion plots and sway-density plots), with a total of 38 parameters. The discriminative power of each parameter was evaluated by means of statistical analysis in relation to the condition effect (open vs. closed eyes) and the pathology effect (normal participants vs. patients). The initial set of 38 parameters was reduced to 24 by identifying clear redundancies, and then to 18 by eliminating the parameters that did not pass the condition effect with normal participants. These parameters were analyzed for reliability and discriminative power in the general framework of a biomechanic model of postural stabilization. At the end of this analysis, we suggested that a set of 4 parameters is particularly valuable in the clinical practice: 2 global parameters (sway-path and frequency band of the posturogram) and 2 structural parameters (mean value of peaks and mean inter-peak distance in the sway-density plots).