Abstracts

Rationale: Precise localization of implanted electrodes with respect to preimplant neuroimaging is a prerequisite to establishing the neuroimaging-neurophysiological correspondence in patients undergoing invasive electrode monitoring. We tested a methodology that involves coregistration of preoperative neuroimaging with postoperative MRI and CT scanning for the localization of hand motor cortex using task related functional MRI (t-fMRI) data.Methods: Post-implant CT (post-CT) images were used to manually identify electrode locations in pre-implant MRI (pre-MRI) space. This was performed by obtaining transformation matrices from coregistration of post-CT in post-implant MRI (post-MRI) space (T1) and post-MRI in pre-MRI space (T2), concatenating the two matrices (T1 & T2) to get a third transformation matrix (T3), which was then applied to the post-CT to coregister it to the pre-MRI space. First, pre-MRI data was processed with Freesurfer software tools to obtain a brainmask (i.e. brain extraction ). This mask was then used to coregister the post-CT (performed using 1 mm cuts to the cervical spine) to the pre-MRI space via use of the post-MRI as described above. A software package (BioImage Suite) was used to appropriately threshold then manually identify and mark electrodes to produce an .mgrid file with electrode coordinates within the pre-MRI space. Finally, electrode coordinates in the .mgrid file were processed to radially snap subdural electrodes to the cortical surface created in first step. This is to ensure that subdural contact coordinates do not either lie floating in space or within the surface of the pre-MRI image as a result of shift of brain surfaces between pre-MRI and post-MRI images. The result is a 3D image of electrodes in 3D pre-MRI space that can be easily manipulated for better viewing and a file with electrode coordinates that can be used to create seeds for analyzing preimplantation neuroimaging. The accuracy of this technique was tested by correlating preimplantion t-fMRI data using a finger tapping task with the results of electrical cortical stimulation (ECS) mapping, and intraoperative digital photography of electrodes. Sensitivity, specificity and predictive values (positive and negative) were calculated.Results: This methodology was used for electrode localization in 6 patients admitted for intracranial EEG monitoring during their evaluation for epilepsy surgery. It was found to accurately localize electrodes when correlating with t-fMRI, results of ECS mapping, and images obtained by intraoperative digital photography. High sensitivity, specificity and predictive values were resultant from this analysis.Conclusions: A combined use of post-implantation MRI, CT scanning and a snapping algorithm provides an effective means of localizing implanted electrodes to preoperative neuroimaging. The methodology described accounts for factors such as brain shift and results in good spatial correlation between preimplantation and postimplantation neuroimaging and good localization of electrodes within the coregistered space thus created.