Stabilogram phase estimation

In order to maintain balance, the central nervous system processes information from visual, vestibular and proprioceptive origin and coordinates an appropriate muscle response. A deficiency of sensory inputs or in the central nervous system usually results in balance disorders. An assessment to the efficiency of the human balance system is important, for example, in the diagnosis of balance disorders and in the monitoring of medical treatment procedures. Stabilograms are obtained from the measures of the displacement of the Center-of-Pressure (ground reaction forces) acting on the subject's feet. and recorded during either static or dynamic conditions. A variety of measures have been introduced to quantify the postural control system especially from registration of the trajectory of the center-of-pressure time series during quiet standing. Another measure is concentrated on the trajectory of the Center-of-Mass. Recent approaches based on concepts taken from statistical mechanics have demonstrated that the Center-of-Pressure displacement is of stochastic origin, and gave a description of postural sway based on the model of bounded, correlated random walk. The phase of a stabilogram trajectory is often ignored. After a short presentation on recent stabilogram analysis methods, we introduce the two approaches, i.e. the empirical mode decomposition and the Hilbert transform, used in this paper for stabilogram phase fluctuations analysis. Hence, we first present evidence that, in general, a stabilogram trace is practically composed Of a Small number of intrinsic modes of proper rotation from which the phase can be computed via the Hilbert transform. Secondly we show that fractional Brownian random processes can describe the fluctuations of the phase about that of a uniform rotation. Finally, we present first results. Implications to nonlinear stabilogram analysis are pointed out in the context of clinical applications and for human postural control system understanding.