Robotized Transcranial Magnetic Stimulation : from automatized protocols towards new approaches in functional neuroimaging

Abstract

Transcranial Magnetic Stimulation (TMS) is a non invasive cortical stimulation tool.Major technologicalevolution has continuously increased the spatial reliability and reproducibility of TMS since its beginningin the middle of the 80’s, by minimizing the influence of human and experimental factors. Therefore, TMSestablished itself as a powerful technique for probing and treating the human brain. The aim of this thesisis to study the methodological and basics contribution of robotized TMS, as being the last technologicaladvance to date. By means of the automatic handling of the TMS coil, robotized TMS opens new avenuesfor the automation of stimulation protocols, and to new approaches in functional neuroimaging. Thetwo first studies of this work aim at developing two tools that are still needed to achieve the automationof set-up procedures of TMS protocols : CortExTool and AutoHS. CortExTool is a toolbox allowing theautomatic analysis of electromyographic signals, while AutoHS is a Bayesian model aiming at automaticallyfinding the motor hotspot, which are two critical ingredients used during such procedures.We validatedour automatic set-up procedure on both virtual and real data, during an experimental comparison againstmanual set-up procedures on 19 healthy volunteers. Results showed that the automatic procedure was atleast as reliable as the manual one, while being faster and more reproducible. The third and last study of thisthesis aims at exploring new basics approaches offered by robotized TMS.We developed a protocol allowingthe extensive mapping of evoked electroencephalographic responses on 18 cortical targets covering thewhole neocortex, and tested it on 22 healthy volunteers. The analysis of the dynamical properties of theseresponses revealed regional specificities as well as cortical networks sharing similar properties. Our resultsprovide the proof of concept of functional cytoarchitectonics, that would guide the parcellation of thehuman cortex in vivo based on its intrinsic responses to local perturbations. The results of this thesis arepromising regarding the new possibilities offered by robotized TMS. Its use could decrease the experimentalvariability, facilitate the handling of TMS protocols used for research and clinical routine, and finally offernew functional exploration approaches that could allow a better diagnosis of psychiatric and neurologicalpathologies.