Transcranial magnetic stimulation and EEG in studies of brain function

Transcranial magnetic stimulation (TMS) combined with electroencephalography (EEG) is a multimodal technique, with a temporal resolution of submilliseconds, for studying cortical excitability and connectivity. When TMS is combined with neuronavigation, resulting in so-called navigated TMS (nTMS), the technique becomes very powerful. However, despite the potential of TMS–EEG, its use for studying lateral areas has been restricted because the TMS pulse induces strong muscle artifacts, making the EEG data useless for further analyses. In this Thesis, methods for analyzing TMS-evoked EEG data from lateral areas are introduced. First, TMS–EEG is used to study Broca's area and dorsal premotor cortex. Due to the fact that those areas are close to cranial muscles, their stimulation evokes large muscle artifacts in EEG recordings. The behavior of the artifacts is described in detail. Two approaches to deal with large artifacts are presented. In the first approach, independent component analysis (ICA) is used. Here, FastICA algorithm is modified to make the search of the components more robust and easier, allowing one to get more stable results. The second approach presents methods for suppressing the artifacts rather than removing them. These methods were combined with source localization showing that the artifact suppression is efficient. The methods were tested with both real and simulated data, suggesting they are useful for artifact correction. For a better understanding of the effects of repetitive nTMS during naming tasks and the cortical organization of speech in general, here another study is introduced to understand the sensitivity of object and action naming tasks to repetitive nTMS. The distributions of cortical sites, where repetitive nTMS produced naming errors during both tasks, are compared. Thus, it is shown how this study can impact on both cognitive neuroscience and clinical practice. In the last part, the beamformer method is improved to study source localization, which makes it a robust method to study time-correlated sources. In this Thesis, I discuss how all these methods together can contribute to study brain connectivity of language and lateral areas with TMS–EEG, opening new possibilities for basic research and clinical applications.