Pre-conference training courses
Cartool is a Windows-based software providing a comprehensive spatio-temporal analysis of multichannel EEG/MEG/iEEG data. Emphasis is put on advanced user interfaces with presets and the interactive 3D-visualization of the data in all analysis steps (waveforms, maps, source localization).
The course will lead through the main features of Cartool that include:
- pre-processing (filtering, interpolation, spatial smoothing)
- statistical analysis using bootstrapping
- microstate segmentation for EEG and ERP
- time-frequency analysis and phase-amplitude coupling
- MRI preprocessing (brain extraction, re-alignment, filtering)
- interactive electrode co-registration
- Individual LSMAC head model
- source localization (MMN, LORETA, LAURA)
- co-registration of intracranial electrodes
Teaching materials will be based on the MNE tutorials: http://martinos.org/mne/stable/tutorials.html
When aiming at the identification of cortical regions involved in electric (and magnetic) brain activities, the restriction to a channel level analysis can provide only limited information. Source level analysis allows to overcome these limitations and gain a deeper insight into cortical activity.
We then propose a full day workshop on source reconstruction using Fieldtrip Toolbox starting from EEG raw data. Three topics will be analyzed: 1. pre-processing and channel analysis, 2. forward modeling, 3. inverse solutions. Theoretical introductions will be delivered, followed by hands-on sessions based on FieldTrip pipelines
Requirements: Familiarity with Matlab
Teachers: Marios Antonakakis, Simon Homölle, Maria Carla Piastra
Duration: 1 day
Requirements: Familiarity with resting state EEG data and EEGLAB.
Teacher: Thomas Koenig
Duration: 1/2 day
Platform: Matlab / EEGLAB
The information contained in multichannel event-related potential data can give precise information about the spatial and temporal distribution of brain information steps. However, for the statistical analysis of such data, the high dimensionality and the unknown correlation structure of the data make an adequate control of type II errors a challenge for statistics.The course will introduce the topographic consistency test, the topographic analyses of variance (TANOVA), and microstate statistics in ERPs. These tests are custom-tailored for ERP analyses and employ randomization procedures to produce statistical evidence. They include all sensors and work on the entire set of measurements and yield robust, physiologically meaningful conclusions with a minimal number of a-priori choices. We will also introduce Ragu, a Matlab based open-source user interface that allows the efficient application and visualization of the test results. The course will consist of theory primers, hands-on analyses of sample datasets, and time to discuss individual questions.
Requirements: Familiarity with multichannel evoked potentials
Teacher: Thomas Koenig & Mara Kottlow
Duration: 1/2 day