6fkrroo/hyhoo5hfrjqlwlrqiurp&kloguhq ¶v'udzlqjvdqg:ulwlqjv à Ã

Studies have shown that some children labelled as presenting $ $ ! saw their access to the written language strongly slowing down by graphomotor difficulties [1]. However, no objective tool allowing the detection of such difficulties is available. Spatio-temporal and kinematics characteristics of the layout probably describe the automation level of the graphic activity, as pointed out in [2]. Therefore, we have defined a protocol containing exercises, like copying figures or writing sentences under different conditions, to underline simple up to complex aspects of graphomotor skill. The skills required during layout production change according to the proposed task. We aimed at analyzing layouts produced by children from a dynamic and a static point of view. To this end, many features were on-line recorded. We present the method we used to select features that significantly describe the automation level of graphic activity. We show that, in most cases, these descriptors allow to recognize the school level of children. We point out that the dynamic characteristics of graphic activity can discriminate children with scholar difficulties as well.