Deep Recurrent Encoder: an end-to-end network to model magnetoencephalography at scale
Article dans une revue avec comité de lecture
Date
2022-10Journal
Neurons, Behavior, Data analysis, and TheoryAbstract
Understanding how the brain responds to sensory inputs from non-invasive brain recordings like magnetoencephalography (MEG) can be particularly challenging: (i) the high-dimensional dynamics of mass neuronal activity are notoriously difficult to model, (ii) signals can greatly vary across subjects and trials and (iii) the relationship between these brain responses and the stimulus features is non-trivial. These challenges have led the community to develop a variety of preprocessing and analytical (almost exclusively linear) methods, each designed to tackle one of these issues. Instead, we propose to address these challenges through a specific end-to-end deep learning architecture, trained to predict the MEG responses of multiple subjects at once. We successfully test this approach on a large cohort of MEG recordings acquired during a one-hour reading task. Our Deep Recurrent Encoder (DRE) reliably predicts MEG responses to words with a three-fold improvement over classic linear methods. We further describe a simple variable importance analysis to investigate the MEG representations learnt by our model and recover the expected evoked responses to word length and word frequency. Last, we show that, contrary to linear encoders, our model captures modulations of the brain response in relation to baseline fluctuations in the alpha frequency band. The quantitative improvement of the present deep learning approach paves the way to a better characterization of the complex dynamics of brain activity from large MEG datasets.
Files in this item
- Name:
- DYNFLUID_NBDT_2022_LOISEAU.pdf
- Size:
- 9.060Mb
- Format:
- Description:
- Deep Recurrent Encoder: an ...
- Embargoed until:
- 2023-05-12
Collections
Related items
Showing items related by title, author, creator and subject.
-
Article dans une revue avec comité de lectureSANSICA, Andrea; KANAMORI, Masashi; HASHIMOTO, Atsushi; LOISEAU, Jean-Christophe; ROBINET, Jean-Christophe (American Institute of Aeronautics and Astronautics (AIAA), 2022-02)When modeled within the unsteady Reynolds-Averaged Navier-Stokes framework, the shock-wave dynamics on a two-dimensional aerofoil at transonic buffet conditions is char- acterized by time-periodic oscillations. Given the ...
-
Article dans une revue avec comité de lectureNASTRO, Gabriele; ROBINET, Jean-Christophe; LOISEAU, Jean-Christophe; PASSAGGIA, Pierre-Yves; MAZELLIER, Nicolas (Cambridge University Press, 2023-01)The stability and sensitivity of two- and three-dimensional global modes developing on steady spanwise-homogeneous laminar separated flows around NACA 4412 swept wings are numerically investigated for different Reynolds ...
-
Chapitre d'ouvrage scientifiqueBUCCI, Michele Alessandro; CHERUBINI, Stefania; ROBINET, Jean-Christophe; LOISEAU, Jean-Christophe (Springer, 2018)With the ever increasing computational power available and the development of high-performances computing, investigating the properties of realistic very large-scale nonlinear dynamical systems has become reachable. It ...
-
Article dans une revue avec comité de lectureCHERUBINI, Stefania; LERICHE, Emmanuel; ROBINET, Jean-Christophe; LOISEAU, Jean-Christophe (Cambridge University Press (CUP), 2014)The linear global instability and resulting transition to turbulence induced by an isolated cylindrical roughness element of height h and diameter d immersed within an incompressible boundary layer flow along a flat plate ...
-
Article dans une revue avec comité de lectureBENGANA, Y.; TUCKERMAN, L. S.; ROBINET, Jean-Christophe; LOISEAU, Jean-Christophe (Cambridge University Press (CUP), 2019)A comprehensive study of the two-dimensional incompressible shear-driven flow in an open square cavity is carried out. Two successive bifurcations lead to two limit cycles with different frequencies and different numbers ...