Séminaire SCALab 23/01/2025
SéminaireJean-Julien Aucouturier, DR CNRS, Dept. of Automation & Robotics,
FEMTO-ST Institute, Besançon
Title: Cracking the code of the social cognition of face and voice with
reverse-correlation methods
Abstract: Reverse correlation is a powerful psychophysical method able
to uncover what stimulus features are used by observers in perceptual
decisions. Although reverse correlation was traditionally restricted to
low-level stimulus dimension (e.g. edge detection in abstract images),
recent signal-processing advances have extended the approach to the
perception of speech and faces, sparking interest in the affective and
social-cognitive sciences. In this talk, I present a set of recent
studies where we used these methods to explore (1) the perception of
speech prosody in healthy (L1 and L2 listeners) and patient populations
(right-hemisphere stroke survivors); and (2) the perception of facial
features of emotional expressions in humans and machines (i.e. AI
explainability). In addition, I will introduce CLEESE, PALIN and JONES
(©2018-2024), a set of open-source, and interoperable, Python toolboxes
we designed to facilitate the design, conduction and analysis of such
experiments - in hope they can foster ideas for your own research
questions.
Bio: JJ Aucouturier is Directeur de recherche CNRS at the Dept. of
Automation & Robotics, FEMTO-ST Institute in Besançon, France. JJA was
trained in electrical engineering (Supélec, 2001), has a doctorate in
computer science (Université Pierre et Marie Curie, 2006) and has held
several postdoctoral positions in cognitive neuroscience in Japan
(University of Tokyo, 2006-2008; RIKEN Brain Science Institute,
2008-2011) and in France (LEAD UMR5022, Université de Bourgogne,
2011-2012). From 2014 to 2020, he was the PI for the CREAM and ACTIVATE
ERC projects and led the music neuroscience team at the Sciences et
Technologies de la Musique et du Son lab (STMS UMR 9912), IRCAM in
Paris. In Jan. 2021, he joined the FEMTO-ST institute in Besançon, to
develop a research project on the interaction between control
engineering and neuroscience, with clinical applications in the fields
of neurology neurology (stroke, coma) and psychiatry.
https://neuro-team-femto.github.io
Partager sur X Partager sur Facebook