Étude méthodologique de l'imagerie optique diffuse couplée à l'électroencéphalographie

Diffuse Optical Imaging (DOI) and electroencephalography (EEG) are two complementary functional brain imaging modalities: they do not image the same physiological phenomenon, but their spatial and temporal resolutions are of the same order of magnitude. Because the relationship between electrical (as measured by EEG) and hemodynamic (as measured by DOI) activities is still poorly understood, being able to record EEG and DOI data simultaneously appears necessary.

On the experimental side, the aim of this project was to adapt an EEG cap to introduce optical fibers and thus, record DOI and EEG simultaneously. A study of the activations in the premotor and motor cortices, involved in movement preparation, allowed us to validate the acquisition setup, which nevertheless needs to be further improved.

The theoretical part of this work focused on the use of regions reconstructed from DOI data as a spatial prior for the resolution of the inverse problem in EEG (i.e. the source location) via a Bayesian model. Simulation data gave us an accurate assessment of the contribution of DOI data and showed in particular that the use of a prior improved the amplitude of the reconstructed EEG sources. This study also showed that, regardless of the spatial coherence between DOI and EEG activities, the use of DOI prior did not worsen the EEG reconstruction.

Although both the theoretical model and the acquisition system remain to be improved, this work showed the interest and feasibility of simultaneous EEG and DOI acquisition.

Key words: diffuse optical imaging, near infrared spectroscopy, electroencephalography, movement preparation, Bayesian estimation