Among the components of postural control, sensorial integration is of crucial importance. The most important sensorial inputs for postural maintenance are the vestibular system which provides information regarding accelerations of the head in space (HORAK 1994), the somatosensory system which provides proprioceptive information that are used to determine changes in body position (INGLIS 1994) and the visual system which provides information for self motion in the environment. The receptors and the integration channel of each sensorial system have inner characteristics which lead to different performance of the different input in regards to the task and the environmental context (FITZGERALD 1994). Most of time, sensorial information are congruent but sometimes they are conflicted. This is the case when being in a stationary train, the train beside going on, giving the illusory sensation of movement even if the vestibular system is not activated. A sensory weighting process is then necessary for subjects to control balance. The Central Nervous System (CNS) is thought to adjust the relative contribution of sensory input to control stance depending on environmental conditions (CENCIANNI 2006) and the reliability of the sensory input (OIE 2002, KESHNER 2004) in order to maintain or achieve the desired orientation in space and to provide postural stability. Nevertheless the interconnection of the multiple sensorial feedback involved in the postural control is not yet completely understood (CHIARI 2000).

Today, there is no tool available to evaluate the individual use of the sensorial information to postural control. It seems interesting to have such a tool to better understand the sensorial preference of subjects. It would be of particular importance for patients with various pathologies in the aim to design individualized balance rehabilitation programs.

The aim of the study was to test a tool built to evaluate the sensorial preference of subjects by studying their postural reaction related to the 3 main sensorial perturbations. Normal subjects will be first tested to assess the repeatability of the protocol and to collect normal values. Then, patients with post stroke hemiplegia, vestibular disease, neuropathy and fallers will be studied in order to test the feasibility of the protocol and to have preliminary data of sensorial preference among these populations.

Subjects Subjects with balance disorder related to a first hemispheric stroke, or vestibular trouble, or neuropathy and or aging

Apparatus Balance evaluation Subjects stand barefoot on a force platform (FeeTest Technoconcept®) consisting of two separate aluminium plates, distant of 12 cm, each placed on 2 force transducers that recorded the vertical ground reaction forces. The position of the center of pression (COP) was calculated from the ground reaction force. The data are collected with a sampling frequency of 40 Hz.

Sensory disturbance Sensory information (vision, proprioception and vestibular information) are successively disturbed. For each sensory information, different perturbations are tested in order to induce a sway in a determined direction (anterior, posterior, right and left).

Vision is disturbed by an optokinetic stimulation in a dark room, an optokinetic bowl (OPTOTEST, Technconcept®) placed behind the subject projecting moving luminous dots on the wall in front of the subject at the speed of 60/°s. The subjects are instructed to stare straight ahead at the stimulus pattern without attempting to follow the moving dots. Four trials are conducted for the visual stimulation in which the direction of the stimulus pattern differs: the direction of the displacement of the luminous dots is first top to bottom up, then bottom to top down, right to left and finally left to right.

Proprioceptive information is disturbed by a vibratory stimulation: vibration is delivered by two mechanical vibrators (VB 115, TECHNOCONCEPT ®). The vibration is of 1mm amplitude and the frequency is 50 Hz. Vibratory stimulation is first applied on the muscle tendon of the triceps sural, afterwards on the tendon of the tibialis anterior.

Then lateral stimulation is applied with a frequency of 90 Hz on the gluteus medius.

Vestibular information Is disturbed by a galvanic stimulation: a binaural galvanic stimulation of 2 mA is applied using a constant current stimulator (World Precision Instruments®). Surface electrodes are placed over the mastoid processes. Four conditions are tested: cathode placed on the right mastoid process then on the left mastoid process with the head looking straight ahead ; cathode placed on the side of the cerebral lesion (or on the right side for the healthy subjects), and head turned to the right then the left.

Procedure Each trial begins with a 15 second baseline period without stimulation, followed by a 35 second period of stimulation and a final 20 second period of observation without stimulation.

Subjects are asked to stand as steel as possible with their arms along the body looking straight ahead (except for the last two trials of the galvanic stimulation where the head is tilted). A trial without stimulation is registered in order to use the subject to the procedure and for each condition sensory perturbation is tested by the subject before starting the recording.

The subjects are tested as soon as possible after the onset of their disease and are retested 4 to 6 weeks later as they completed rehabilitation program.

Healthy subjects are also tested. Data analyzed For each sensory perturbation, we analyze two characteristics of the behavior of the subjects during the stimulation: the displacement induced by the stimulation and the intensity of sway during the stimulation ·Because the position of all the subjects can not be rigorously identical among all the subjects, the displacement induced by the stimulation is calculated by the mean of the displacement (in millimeters)(MD) of the COP (15 to 50 seconds) in the plane of the attended effect (for example anterior-posterior for a triceps sural trial and lateral for an optokinetic trial from the right to the left) minus the mean position of the CoP during the prestimulation period (2 to 13 seconds). A sensorial score is obtained in each direction for the successive sensory stimulation. From these values, and for each sensory perturbation (visual, proprioceptive and vestibular) a composite score is calculated as the mean of all the trials done for one sensorial stimulation. Thus, we obtain three scores: a visual composite score, a proprioceptive composite score and a vestibular composite score.

The number of falls induced by the stimulations is also registered.

·The intensity of sway during the stimulation is evaluated by a fast fourier transform Statistical analysis The reproducibility of our test is assessed by the intra-class correlation coefficient (ICC). Reproducibility is considered as correct if the ICC is superior at 0.6 and excellent if superior at 0.8.

Descriptive data are reported as means with the (standard deviation) SDs. Correlation between the force platform values and the subjects characteristics are evaluated by non parametric statistics: Wilcoxon signed-rank test, Mann-Whitney test and Kruskal-Wallis test. P values are chosen at 0.05.

Inclusion Criteria: depending on the group

• Healthy subjects without history of balance disorder
• Vestibular patients with a unique history of peripheric vestibular disorder
• Post stroke patients after a unique hemispheric stroke, being able to stand without any help
• Patient with proprioception disorder related to peripheral neuropathy.
• Old subjects with a history of falls (one or more during the last year).

Exclusion Criteria:

• For each category, other disorder than the one studied leading to balance trouble.
• Neuropsychological troubles preventing comprehension of the tests.

Healthy subjects Vestibular patients Post stroke patients Patient with proprioception disorder Old subjects with a history of falls