Journée des actions de l’axe 2016
lundi 13 juin 2016
Lundi 20 Juin, salle 1 du LJAD
- 9h15-9h30 : Quelques mots d’accueil
- 9h30-10h00 : Tobias Scheer (BCL) Non-invasive replication of Sahin et al.
Sahin et al. (2009) used deep electrodes (implated for medical purposes) to show in a production experiment (silent pronunciation) that, as predicted by modularity, lexical access, morpho-syntactic and phonological computation constitute three distinct computations whose occurrence in the timeline is located at about 200ms, 320ms and 450ms upon stimulus, respectively. Using the same stimuli as Sahin et al. (English singular - plural alternations), we have successfully replicated their results based on non-invasive scalp EEG. Relevant data are presented and the perspective for future work is outlined : knowing that the experimental setup works and what we have to look for in scalp EEG data, we can move beyond this pilot to test our actual linguistic target where the existence of phonological activity is contentious (the "electricity" paradigm, also called velar softening).
A bonus in our current EEG data is that they also allow us to test more fine-grained linguistic predictions (regarding the voicing of obstruents and sonorants), which Sahin et al. did not address. Results confirm by and large what is expected, and also show that, again as predicted by the standard theory, the voicing of sonorants and obstruents is not the same thing.
- 10h00-10h30 : Pascal Chossat (LJAD) Dynamiques hétéroclines en Neurosciences
- 10h30-11h : Pause café
- 11h- 11h30 : Elisabetta de Maria (I3S) Verification of Temporal Properties of Neuronal Archetypes Modeled as Synchronous Reactive Systems
There exists many ways to connect two, three or more neurons
together to form different graphs. We call archetypes only the graphs
whose properties can be associated to specific classes of biologically relevant
structures and behaviors. These archetypes are supposed to be the
basis of typical instances of neuronal information processing. In this projet,
we model different representative archetypes and express their temporal properties
using a synchronous programming language dedicated to reactive systems
(Lustre). The properties are then automatically validated thanks
to several model checkers supporting data types. The respective results
are finally compared and depend on their underlying abstraction methods.
- 11h30 -12h00 : Patricia Reynaud-Bouret (LJAD) : Synchrony detection in neuroscience
- 12h00-14h00 : Pause déjeuner
- 14h00- 14h30 : Benoit Miramond (LEAT) Les systèmes neuromorphiques, un défi interdisciplinaire
- 14h30 - 15h00 : Ismaël Rafaï / Dino Borie (GREDEG) Définition théorique d’une mesure expérimentale de l’attention
- 15h00 - 15h30 : Pause Café
- 15h30 - 16h00 : Frédéric Precioso (I3S) Electromyographic signals classification for the analysis of inner language by Transfer Learning
- 16h00 - 16h30 : Etienne Tanré (INRIA) Des processus de PDMP en neuroscience, pourquoi, comment.