Presentation of Neural Communication

There is surprisingly little literature on the promising connection between information theory, in particular coding theory and digital communications, and brain signal processing. The CominLabs “Neural Coding” project laid the foundation for the first joint work linking mental information theory and brain electrical signal through ElectroEncephaloGraphy (EEG). This collaboration allowed mixing top-down and bottom-up approaches for better analysis of signals measured on human brains. This great success paved the way for many open questions, which this project entitled “Neural Communication” aims at exploring.
The team NeuCod led by Claude Berrou in lab-STICC (Télécom Bretagne) and the team SESAME led by Fabrice Wendling at LTSI (University of Rennes 1) want to pursue their joint work and push towards new aspects. Listed from those closest to the signal to most abstract ones, these aspects include heterogeneous approaches to signal processing based on graphs, connections between digital communication and information in the brain and dynamic models for mental information. “Neural Communication” does not propose a substitution for the “Neural Coding” project, but rather aims at giving new dimensions to this fruitful collaboration.
Recently, the collaboration between the two previously cited teams led to exciting results showing that EEG techniques can lead to very good time resolution (of the order of several milliseconds). With this accuracy, it is possible to observe the processing of information as a set of active graphs going from brain regions to others. These interpretations are new and take full benefit of EEG techniques that allow observing correlation between sources, even if those have limited amplitude, contrary to BOLD or simple source localization methods. The automatic dynamic detection of changes in correlation graphs could lead to important valorisations such as tracking brain dynamics at millisecond time scale. This motivation converges with that of adding processing to the static mental information theory developed at Télécom Bretagne.
In one sentence, the “Neural Communication” project aims at exploring the dynamic aspects of mental information theory at the millisecond scale.

The consortium involves the following people and teams:
- Télécom-Bretagne, Lab-STICC: Vincent Gripon, Claude Berrou and Olivier Dufor
- University of Rennes I, LTSI: Fabrice Wendling and Mahmoud Hassan
- B-Com: Jean-Marc Diverrez
- Orange Labs: Rozenn Nicol and Lætitia Gros