Scientific direction Development of key enabling technologies
Transfer of knowledge to industry

PhD : selection by topics

Engineering science >> Physiopathology
1 proposition(s).

Combined analysis of kinematic, brain activity and oculometric parameters related to handwriting with supervised machine learning models for dysgraphia diagnosis tool in children

Département Systèmes

Laboratoire Signaux et Systèmes de Capteurs

01-10-2019

SL-DRT-19-0443

etienne.labyt@cea.fr

Nearly a third of school-aged children fail to develop the efficient handwriting performance required to cope at school (Smits-Engelsman et al, 2001; Danna et al, 2016). Among this population, 5 to 10% of children are dysgraphic. Currently, the diagnosis of dysgraphia is based on the BHK test which is relatively subjective. Most of the time, the dysgraphia diagnosis is done lately and lead to serious consequences on children scholar achievements. It is thus crucial to diagnose and handle these deficits as early as possible. The most investigated aspects of handwriting is the motor level (Danna et al, 2013; Smits-Engelsman & Galen, 1997; Hamstra-Bletz & Blöte, 1993), but brain or oculomotor activities associated to handwriting have been poorly investigated in children. Recently, a first algorithm for automatic detection of dysgraphia has been developed (Asselborn et al, 2018), but technological improvements are required for its use in the dysgraphia diagnosis. In a previous study, an important database has been collected in typical and dysgraphic children and handwriting parameters specific to dysgraphic children have been identified and used to develop a first algorithm. Performances achieved in terms of dysgraphia diagnosis are around 85%. The current PhD position aims at analyzing the handwriting in typical and dysgraphic children by using 3 simultaneous measurements: handwriting kinematic parameters, brain activity recorded by EEG and oculomotor activity recorded by eye tracking. From these data, contribution of EEG and oculomotor features in supervised machine learning models will be assessed. The final goal is to develop a new tool, automatic and reliable, for dysgraphia diagnosis.

Voir toutes nos offres