It appears you don't have support to open PDFs in this web browser. To view this file, Open with your PDF reader
Abstract
In neuroscience, simulating electric current in the head of a subject is of main interest for both electroencephalography (EEG) and transcranial direct current stimulation(tDCS). EEG is used to reconstruct the electric activity of the brain based on the measured electric potential on the scalp. On the other hand, tDCS consists in injecting a small electric current through the head of a subject to modulate the activity of a specific brain region.
Such simulations rely heavily on the electric conductivity of the biological tissues composing the head. Unfortunately, there is currently no effective and non-invasive method to measure it accurately for each individual. Consequently, researchers and practitioners have to set arbitrary values chosen from the literature, despite the fact that this property has been shown to vary widely both inter- and intra-subject. The simulations also depend on the geometry of the tissues and on how they are modelled.
In this thesis, we studied the influence of different skull models and of the electrical conductivity of the tissues on the EEG forward problem. We also analysed the effect of the uncertainty in the conductivity on the electric field induced in different regions of the brain by several stimulating electrode montages in tDCS.
To support these experiments, we developed a python package named Shamo which provides the user with tools to perform mesh generation, current simulation, surrogate modelling and sensitivity and uncertainty analyses with a user-friendly API. It interfaces with industrial grade software to perform the computationally intensive tasks and is easy to use on distributed architectures.
The present work describes both Shamoand the results that it helped to obtain for the different experiments.