Context: Among the intrinsic factors to the risk of falling in the elderly, the alteration of visual capacities is central. Paradoxically, this modality remains preferentially used for their postural control as proved their high field dependence [1], known to reduce sensorimotor adaptability [2]. Many studies consider that increased postural variability is a marker of aging [3]. The entropy-based complexity metrics that relates to the information content of a signal by quantifying the degree of regularity allowed detecting an exacerbated degradation of complexity in postural sway in aged people, thus indicating a failure in the dynamic re-organization of control [4]. These metrics may provide a window on the interactions across scale supporting engaged cognitive processes and might offer an original approach to the concept of adaptability [5], notably on the cognitivo-postural behaviour. The aim of this study was twofold: i) to analyse the variability and complexity of the cognitivo-postural behaviour during different suprapostural tasks in young and over 65-y.o adults, and ii) to explore how behavioural fluctuations and field dependence level are linked in elderly. Method: Two 12-healthy-subjects groups were recruited: one over 65 years of age (mean±sd: 75,4±5,1) and a second one of young adults (21,9±2,4). In standing posture on a force plate, each of them performed two 51.2s-suprapostural tasks separately: 1) to focus a dot (VFix); 2) to mentally count C-letters projected in front of them as fast and as efficiently as possible (VDis). During the latter, once the count verbalized, they had to continue on a new C-panel until the end of trial (Lev1). Two other similar and counterbalanced 3-trials-block but with different cognitive difficulty were carried out with same instructions: one with A-letters added to C-letters (Lev2) and another with added A- and G-letters (Lev3). The postural variability will be assessed from the standard deviation of CoP signal (SD). The entropy index (EI) is calculated from the refined composite multiscale entropy curves and after pre-processing time series using empirical mode decomposition [6]. In order to answer to the first objective, statistical comparisons between tasks and within and between groups will be carried out on SD and EI. The choice of tests will depend on the normality of the distributions and the homogeneity of their variance. To respond to the second objective, the elderly group performed a Rod & Frame Test in which they were asked to roll-adjust an initially tilted rod in a 18◦ roll-tilted frame to the direction of gravity. Higher the mean signed error (i.e. in the direction of the frame tilt), the more the subject was field dependent. Factor analyses (PCA, correlations) will be performed to explore associations among FD, SD and EI. Results & discussion: We speculated that the older group will show higher SD, lower EI and higher FD than the younger group. The coordinated analysis of the magnitude and structure of behavioural fluctuations [7] will allow to improve knowledge of the mechanisms of cognitivo-postural adaptability with age. And the potential linking of these fluctuation indicators to perceptual style is to open a large research field around loss of flexibility associated to aging.