Dynamic Modelling of the Brain-electrode Interface for Stereoelectroencephalography

The modelling of the brain-electrode interface has many useful applications whenever patients are implanted with intra-cranial electrodes. In general, extensive research has been performed using electrode impedance spectroscopy in cases of deep brain stimulation for patients suffering from motor disorders. The brain-electrode interface model is also useful when it comes to epileptic patients, not only for diagnosis purpose, artefact correction, but also potentially for brain tissue classification. However, for this latter application, the interface model has never been studied considering typical stereoelectroencephalography signals. In this scope, the aim of this paper is to propose a brain-electrode interface model adapted to be used considering only voltages recorded during pre-surgical investigations. Various models are discussed based on a previous study on a phantom EEG device, and compared to the expected dynamics between pairs of consecutive contacts also presented in a previous non-parametric transfer function study. From this, two new models are proposed: one based in monopolar derivation, and the other on bipolar derivation. The models are simulated, and identified for three pairs of contacts inserted in one patient. The model based on bipolar derivation had better results.