Faites connaître cette offre !
Reference : UMR5105-GUYOMN0-018
Workplace : GRENOBLE
Date of publication : Tuesday, September 10, 2019
Scientific Responsible name : Caroline Jolly
Type of Contract : PhD Student contract / Thesis offer
Contract Period : 36 months
Start date of the thesis : 1 November 2019
Proportion of work : Full time
Remuneration : 2 135,00 € gross monthly
Description of the thesis topic
-------Handwriting deficits, also termed dysgraphia, concern 5 to 10% of children (Smits-Engelsman et al, 2001; Danna et al, 2016; Jolly 2017). The diagnosis of dysgraphia in France is based on a relatively subjective test, the BHK (Charles et al., 2003). If not handled properly, handwriting deficits can seriously impact both children's behavioural and academical development, eventually leading to academic failure. Early diagnosis of dysgraphia is thus crucial.
In the literature, dysgraphia are well described from a motor perspective (Danna et al, 2013; Smits-Engelsman & van Galen, 1997; Hamstra-Bletz & Blöte, 1993), but little is known about the cerebral and oculomotor activities involved. Recently, an algorithm allowing the identification of dysgraphic children based on writing samples collected on a graphic tablet has been developed (Asselborn et al, 2018). Although very promising, this automated tool presents some technological limitations preventing its use for the diagnosis (Deschamps et al., 2019). In the context of a former grant obtained from the Bottom Up program of CEA, we collected a large database of handwriting and drawing samples from typical and dysgraphic children, and we used machine learning approaches to identify the kinematic features of handwriting which allow the better discrimination between typical and dysgraphic children. Our algorithms allow a correct classification of children with a precision of about 85%.
The objective of the thesis project is to analyse handwriting in typical and dysgraphic children using three concomittant measures: kinematic parameters of handwriting, cortical activity measured using electroencephalography, and oculomotor activity measured using eye-tracking. A first objective will focus on analyzing the evolution of handwriting using these 3 measures from a developmental perspective, in the population of children. A second objective will consist in evaluating the benefit of EEG and eye-tracking data in supervised models for the diagnosis of dysgraphia. The final goal of this project is to develop a new, fully-automated and reliable tool for the diagnosis of dysgraphia.
* Missions :
The candidate will have to collect handwriting, EEG and eye-tracking data from the 2 populations of children. Using these data, the candidate will then use machine learning approaches (supervised models) to analyse handwriting development in typical and atypical children, and to develop an algorithm for the diagnosis of dysgraphia.
- master 2 or engineer level, with a good expertise in statistical modelisation and machine learning
- skills in Python programming are recommended
- a former expertise in neurosciences would be beneficial
* working place :
The thesis will take place on 2 sites, the LPNC laboratory (UMR CNRS 5105) located at the University campus, and the CEA LETI.
* thesis direction :
The thesis will be co-directed by Dr Caroline Jolly (LPNC) for all Cognition and Neurosciences aspects of the project, et by Dr Etienne Labyt (CEA, LETI) for Machine Learning aspects. The candidate will benefit from the expertise of Dr Caroline Jolly in the field of typical and atypical child development, namely handwriting, and from the expertise of Dr Etienne Labyt in EEG et machine learning. Dr Vincent Brault from the Laboratory Jean Kuntzmann (IMAG – UGA) will also be involved in the project, and will bring his expertise in the domain of statistical modelisation.
* école doctorale :
Ingénierie pour la santé la Cognition et l'Environnement
Spécialité BIS - Biotechnologie, instrumentation, signal et imagerie pour la biologie, la médecine et l'environnement
This project is at the border of three research fields : typical and atypical child development, cognitive neurosciences, and health engineering.
We talk about it on Twitter!