The COPHY team relies on an original and unique methodological expertise that we intend to extend further. This concerns not only our expertise in computational modelling and human electrophysiology, but also our developments for advanced data acquisition and analysis. In particular, we are currently introducing the use of individual headcast for laminar MEG source reconstruction. We are further developing new analysis tools for combining feature specific laminar fMRI signals and EEG signals at the source level. We are also establishing collaborations (already funded by a regional grant “Pack ambition recherche”) with CERMEP and CEA LETI in Grenoble as well as international connections for promoting and contributing to the development of a new generation of portable MEG sensors at ambient temperature (Optically Pumped Magnetometers, OPMs).

​

Regarding data analysis, an original line of research consists in using real-time approaches, like in BCI, to optimize the experimental design, online, at the individual level (so-called adaptive design optimization). This closes the loop between our BCI developments and our fundamental studies as such an optimization aims at reaching the highest sensitivity in hypothesis testing, whatever the models to be compared (e.g. Bayesian learning, DCM, brain-behaviour relationship…). In this aim, an important aspect will be to develop efficient and robust artefact correction methods for online experiments with EEG and/or MEG.

​

We are also developing artificial intelligence networks (including spike nets) as realistic generative models of both behavioural and neurophysiological observations in humans.

​

We also started using non-invasive neuromodulation approaches, namely tACS, in order to explicitely test the causality between inter-areal oscillatory synchronization (in combination with flickering stimuli) and perceptual performance (ERC funded).

​

Besides, we will work on a head-mounted, virtual reality version of our EEG-based BCI games for rehabilitation (already funded by Labex, 30k€). This will also allow more immersive, more ecological experiments, outside the laboratory.

​

Importantly, we will go on sharing our expertise in statistics, machine learning, Bayesian modelling and signal processing within the CRNL, through the sharing of codes (via a gitlab repository) and through the animation of a weekly methods meeting.