ORBITAL STABILITY ANALYSIS OF VOLUNTARILY ALTERED GAIT PATTERN

INTRODUCTION Falls are the primary aetiology of accidental deaths in persons over the age of 65 years in the United States [1]. The occurrence of fall is generally associated to an alteration of the stability in motion, although no consensus exists regarding how this stability characteristic can be quantified. Stability analyses of human gait range from mere observation by a trained physician or physical therapist to variability quantification of gait parameters, to the application of nonlinear techniques, designed to quantify the stability of a mechanical system. Among these, orbital stability quantified by means of Floquet Multipliers (FM) was found promising [2] in assessing how a periodic system (e.g. the neuro-muscolo-skeletal system during locomotion) responds to small perturbations (e.g. perturbations in control and/or environment determining gait pattern variability). This knowledge could lead to a better understanding of the control mechanisms determining the stability of a variable gait pattern, supporting the development of more effective rehabilitative procedures. Moreover, the eventual quantification of orbital stability exploiting a minimal measurement wearable set-up could contribute to the development of effective devices for the early detection/alerting of fall. The aim of the present preliminary study is to determine if alterations coming from voluntary changes in gait pattern during over-ground walking generate orbital stability alterations detectable from the acceleration data of a single inertial sensor. MATERIALS AND METHODS 5 healthy subjects (3 males, 2 females; age 27-35) performed an overground walking task at their preferred speed on a 20m hallway in 3 different walking conditions: normal walking (No), walking with narrower steps (N) and walking with wider steps (W). A portable triaxial inertial sensor (Xsens Technology, Enschede, Netherlands) was placed on the subjects, at the level of L5. State spaces were created using 5-dimensional delay embedding of the accelerations of the L5 sensor in the medio-lateral (ML), anterior-posterior (AP) and vertical (VT) directions [3]. Maximum FM (maxFM) were then computed for each time series. Acceleration data were normalized with respect to the height of the subjects. RESULTS MaxFM in the double support phase showed values < 1 for all the gait conditions and all the measurement directions. These values were lower than the mean values of the maxFM calculated through all the step cycle (Fig.1). In the N condition, maxFM values in double support phase resulted higher than in the other walking conditions, for all the measurement directions. Figure 1 - Values of maxFM calculated in the double support phase (Grey) compared to the mean values of maxFM through all the step cycle (Black) in the medio-lateral (ML), anterior-posterior (AP) and vertical (VT) directions, in the normal (No), narrow steps (N) and wide steps (W) gait condition. DISCUSSION MaxFM showed stable values (< 1) for all the gait conditions and measurement directions, indicating that all the subjects maintained orbital stability during the execution of the task. Lower values showed by maxFM during double support phase pointed out the capability of the method to identify the best stability condition of this phase of the step cycle with respect to the mean stability during the cycle. Higher values of maxFM of the double support phase in the N condition revealed a decrease in stability during that phase. In conclusion, acceleration data coming from one sensor placed at the level of L5 confirmed stability of walking normally and with voluntary changes in gait patterns in young subjects, but was just partially able to discriminate between the different walking condition. Future studies will involve orbital stability analysis of motor tasks performed by elderly or pathologic subjects.