Efficient volume-based localization and automatic labeling of intracranial depth electrodes.

Background: The accurate localization and anatomical labeling of intracranial depth electrodes are crucial for stereoelectroencephalography (SEEG) recordings and the interpretation of results in patients with epilepsy. The laborious electrode localization procedure requires an efficient and easy-to-use pipeline. Thus, we developed a useful tool, which we called the depth electrode localizer (DELLO), to automatically identify and label depth electrode contacts with ease. Methods: The DELLO is an open-source package developed in MATLAB (MathWorks). It was specifically fine-tuned to expedite the localization of depth electrodes. The basic procedures include preoperative magnetic resonance imaging (MRI) and postoperative computed tomography coregistration, intensity threshold electrode spatial sampling, the hierarchical clustering of electrode samples, and gray-matter and automatic anatomical labeling (AAL). The DELLO also has a graphical user interface (GUI) that can be used to review the results. The only manual intervention procedures are the identification of the target (tip) and entry point of each electrode and the naming of the clustered electrode contact groups, which generally take ~5 min per case. The coordinates of each contact were recorded in individual spaces and were also transformed in standard space by applying a volume-based deformation field. To validate the performance of the current method, 7 patients with epilepsy were retrospectively included in the analysis. Results: A total of 80 depth electrodes, including 1,030 contacts from the 7 patients with epilepsy, were localized. All the procedures functioned well, and the entire process was robust and intuitive. Among the 1,030 contacts, 746 (72.43%) were labeled as inside the gray matter. The gray-matter and AAL accuracy rates were 95.83% and 90.78%, respectively, over all contacts. Conclusions: The DELLO is an integrated tool that was designed to semi-automatically localize and label intracranial depth electrodes. It is open source and freely available. Given its high accuracy and efficiency, the DELLO could facilitate SEEG interpretation and be used in SEEG-based cognitive neuroscience studies.