# Mapping white matter tracts with SEEG electrodes

## TL;DR

SEEG electrodes, used for epilepsy diagnosis, also sample white matter pathways, making them suitable for studying white matter stimulation.

## Contribution

The study identifies white matter tracts sampled by SEEG electrodes and validates their potential for stimulation in two patients.

## Key findings

- Approximately two-thirds of SEEG electrode contacts sample white matter structures.
- Commissural and association fibers like the corpus callosum and cingulum are nearly universally sampled.
- White matter stimulation induced phosphenes in two patients, showing reproducibility and tract selectivity.

## Abstract

Stereo‐electroencephalography (SEEG) is designed to record gray matter (GM) activity for epileptogenic zone localization. SEEG electrodes, however, traverse white matter (WM) pathways that connect regions involved in seizure networks and cognition. We retrospectively evaluated WM tract sampling in patients with SEEG to identify pathways suitable for prospective WM stimulation studies and validated this in two patients.

This retrospective analysis included 86 individuals who underwent SEEG implantation for drug‐resistant epilepsy between 2014 and 2020. Electrode contacts were localized using postoperative high‐resolution computed tomography (CT), co‐registered to preoperative 3T T1‐weighted magnetic resonance imaging (MRI) and normalized to Montreal Neurological Institute (MNI) space. Tissue classification utilized the Harvard–Oxford atlas, whereas WM tract involvement was assessed using probabilistic WM atlases. Preoperative tractography with dissection of the optic radiation was performed and used for 50 Hz, image‐guided SEEG stimulation of white matter.

Among 86 patients (30 left, 40 right, 16 bilateral hemisphere implantations), 860 electrodes (6372 contacts) were implanted. Of 5853 intraparenchymal contacts (92%), 1826 (31%) were positioned exclusively within WM and 2554 (44%) at the GM/WM boundary. Patients had an average of 10 electrodes (74 contacts). Of intraparenchymal contacts, 4381 (75%) crossed WM pathways (average 21 tracts per patient). The most frequently sampled tracts were commissural fibers (corpus callosum: 100% of patients), followed by association fibers including the inferior fronto‐occipital fasciculus (97.7%), superior longitudinal fasciculus II (96.5%), and cingulum (95.4%). In two separate patients, WM stimulation induced phosphenes at .5–1 mA in different distant electrodes with contacts in the optic radiation, which was reproducible and tract selective.

SEEG electrodes sample WM structures in approximately two‐thirds of contacts, and association pathways demonstrate near‐universal sampling, making them optimal candidates for systematic WM stimulation protocols. SEEG–WM stimulation with preoperative tractography was validated in two different patients.