GeoSourceElectrical Source Estimation GeoSource is a set of tools to model the neural sources of the brain’s electrical fields measured with the electroencephalogram (EEG) at the head surface. Unlike first-generation source analysis packages, the goal for GeoSource is not one of basic science exploration, nor merely to prove that EEG sources can be characterized accurately. With the advances of 256-channel whole head EEG recordings, exact sensor position registration, and accurate conductivity models of head tissues, the accuracy of distributed source models can now be established within reasonable confidence limits. The goal of GeoSource, therefore, is to facilitate ease of use and thus efficient workflow. Seamless Integration with EGI’s Net Station Software • Move from scalp space to brain space with the touch of a button. • View sample-by-sample playback of scalp data and source data side by side. • Link the Net Station topo map view with the GeoSource MRI and/or flat map views. • Run source estimation on the data of any number of subjects, files, or categories with a single tool specification. • Export source waveform data for statistical analysis. • Use standard Net Station waveform tools (such as baseline correction, combine files, difference wave, file export, filtering, t-test, and wavelet) on your source waveform data. Source Estimation Methods• Choose an inverse method (minimum norm least squares (MNLS), EPIFOCUS, or FOCUSS). • Counteract the biases inherent to all inverse methods by selecting a weighting technique. • Constrain the source space solution by selecting an inverse solution (LAURA or LORETA). • Account for the variability in signal-to-noise by selecting a regularization method (Tikhonov or truncated singular value decomposition (TSVD)). Dipole Sets• Sparse dipole set. Contains 57 dipoles that are distributed across the functional regions of the cortex. • Dense dipole set. Contains 2,447 dipoles that are distributed in 7 mm voxels across the cortex, with each dipole being represented by a triple regional source in the orthogonal orientations (x,y,z). Sensor Registration• Automatically load the appropriate Geodesic Sensor Net 3D coordinates. • Improve source estimation with subject-specific 3D EEG sensor locations from EGI’s Geodesic Photogrammetry System. Head Models• Finite difference model (FDM). Based on EGI’s detailed analysis of the geometry and conductivity of the tissues of a typical head (the MNI “atlas man”) using detailed MRIs (for soft tissue) and CTs (for skull). • Isotropic Sun-Stok four-shell spherical model. Includes preset radii and conductivity values. Flat Map View • View modeled source activity across the entire cortical surface in a single glance. • Inspect the progression of activity across regions over time. • Identify concurrently active regions. • Show or hide the cortical folds for any flat map. MRI View• Visualize dipoles overlaid on sagittal, coronal, and axial MRI slices. • View important file information, including original EEG filename, experimental condition descriptor, source specification name, and time. • Standardize dipole amplitudes across multiple windows.  • Navigate through MRI slices with sliders, text entry, or arrow buttons. • Jump to maximal source activity for individual EEG time points. • Sort dipoles by intensity and navigate directly to the MRI dipole location from the Intensity pane. • View the orientation vector for each dipole in the dense dipole set. Visualization (MRI and flat map views)• Automatically display the ID number, location (in Cartesian coordinates and in Talairach space), intensity (in nanoamps), Brodmann area, gyrus name, and lobe of dipoles. • Link crosshairs at a dipole location across multiple MRI and flat map windows. • Integrate with Net Station Viewer for simultaneous playback of scalp waveform and dipole activation. Source Waveforms • Generate source waveforms for any of the 2,447 dipoles or for predefined and user-defined montages. • View three orthogonal orientations (x,y,z) or the single RMS vector. • Display source waveforms as continuous data, as categories, or overlaid segments to compare source activity across experimental conditions. • Extract time and amplitude from source waveforms for statistical analysis.
Source Montages • View total source intensity in MRI view (alongside activity across the cortex in flat map view) for predefined montages based on Brodmann areas, gyrii, and lobes. • Create source montages based on MRI-derived ROIs, Talairach coordinates, or any combination of voxels. • Generate user-defined montages with dipole groups by selecting individual voxels, voxels within a specified radius of an individual voxel, or voxels within a specified threshold range of an individual voxel. • Export statistics for total source intensity for predefined and user-defined source montages. Handle Multiple Conditions and Groups• Open multiple MRI and flat map windows to compare source activity between conditions or subject groups. • Link multiple conditions or subject group windows to navigate through MRI slices and flat maps. • Jump to maximal activity in each MRI and flat map window, or lock all displays to a region of interest. Training and Documentation
For further details, Contact EGI for GeoSource-related white papers. GeoSource 2.0 is now cleared for clinical use in the US and the EU. All clinical use is strictly limited to the approved intended uses as stated in the regulatory filings.
|
GeoSource users have a thorough user's manual that illustrates the basic principles of EEG source estimation, and that explains the specific parameters for use that will give valid results. For instance, as with any linear inverse method, artifacts in the recording (e.g., movements and eye blinks) will distort the source estimates. The user must also understand the degree of regularization that is required for the noise level of the recording. In addition to the GeoSource manual, EGI provides regular training sessions for GeoSource, both at the EGI Summer School and at scientific meetings.