CONICET | Buscador de Institutos y Recursos Humanos

Purpose: Intracranial ERPs from epilepsy patients are a great opportunity to study the brain activity of cognitive processes with a unique spatial and temporal resolution. In this cases is of critical importance to know the exact localization of the electrodes. Here we present an open source Matlab toolbox (https://sourceforge.net/projects/ielectrodes/ ) to obtain the coordinates and anatomical labels of each electrode in a semiautomatic way with minimal user intervention.Method: Six patients implanted with grids and depth electrodes were studied. Post implantation T1 MRI and CT were coregistered using an affine transformation with SPM8 toolbox. Subject-specific cortical segmentation labels were obtained using Freesurfer software. Brain masks were obtained using FSL-BET software. T1 MRI and CT were coregistered to the MNI-152 space using a brain mask and nonlinear warping deformations with SPM8. Images were then processed in the toolbox (see figure).Electrode voxels were detected by using a dilated brain mask and thresholding high intensity CT voxels in a 3D reconstruction [1]. User intervention was needed to determine the threshold and dilation levels. Then, these voxels were automatically clustered and the electrode coordinates were obtained. Electrodes were numbered and an anatomical label was assigned automatically to each one. A 3D view of electrodes over a semitransparent brain was shown for visual interpretation. CT and MRI sagittal, coronal, and axial views of electrode coordinates were visually checked by an expert. Electrode coordinates are finally exported to EEGLAB.Results: Six patients were studied with 104/332 SEEG/ECoG electrodes. In all cases, electrode coordinates were successfully located within visualized electrodes artifact in CT and MRI images.Conclusions: The proposed tool is a useful instrument to achieve a fast and robust localization and labeling of electrodes.