Validating computationally predicted TMS stimulation areas using direct electrical stimulation in patients with brain tumors near precentral regions

Abstract

The spatial extent of transcranial magnetic stimulation (TMS) is of paramount interest for all studies employing this method. It is generally assumed that the induced electric field is the crucial parameter to determine which cortical regions are excited. While it is difficult to directly measure the electric field, one usually relies on computational models to estimate the electric field distribution. Direct electrical stimulation (DES) is a local brain stimulation method generally considered the gold standard to map structure-function relationships in the brain. Its application is typically limited to patients undergoing brain surgery. In this study we compare the computationally predicted stimulation area in TMS with the DES area in six patients with tumors near precentral regions. We combine a motor evoked potential (MEP) mapping experiment for both TMS and DES with realistic individual finite element method (FEM) simulations of the electric field distribution during TMS and DES. On average, stimulation areas in TMS and DES show an overlap of up to 80%, thus validating our computational physiology approach to estimate TMS excitation volumes. Our results can help in understanding the spatial spread of TMS effects and in optimizing stimulation protocols to more specifically target certain cortical regions based on computational modeling.

Keywords: Direct electrical stimulation; Finite element method; Motor cortex; Transcranial magnetic stimulation.

Fig. 1

Experimental setup: A) A 5 × 5 grid (1 cm spacing) was placed on the scalp over the primary motor cortex (left panel). Orientation of the TMS coil (indicated by blue arrows) was 45° to midline for each position which is approximately perpendicular to the precentral gyrus (right panel). B) Ten motor evoked potentials (MEPs) were recorded at each position (overlaid potentials for two positions shown at the left panel). Based on the average of the MEP amplitudes a MEP map is calculated.

Fig. 2

Direct electrical stimulation: A) Shown are the DES stimulation points (white squares enhanced in size for better visibility) in one example subject. B) Illustration of the intraoperative stimulation procedure. The position of the stimulation electrode is controlled by a neuronavigation software. The red cross indicates the target point at which the stimulation electrode was aimed (green cross). Different points on the motor cortex were stimulated and the elicited MEP recorded. C) Simulated electric field distribution for the DES for one stimulation point. High electric field strengths are restricted to a confined radius around the stimulation electrode.

Fig. 3

TMS computational models: A) sagittal cut through the head models for both the realistic (left panel) and the spherical (right panel) case. The surfaces of the five different tissue types are shown. The spherical model was fitted to the upper half of the skin surface of the realistic model. B) Exemplary electric field distribution in one patient for one coil position for both the realistic (left panel) and the spherical (right panel) model. While in the realistic model clear effects of tissue boundaries are visible, the electric field distribution of the spherical model is mainly determined by the primary electric field of the TMS coil.

Fig. 4

Computational predicted stimulation areas: Shown is the MEP weighted mean electric field for the A) TMS realistic model, B) TMS spherical model interpolated on the realistic GM surface and C) DES. The stimulation area in the realistic model is restricted to the crowns of the precentral gyrus as well as neighboring gyri. For the spherical model stimulation area is more extended. The stimulation area of the DES is mostly restricted to the primary motor cortex. D) Region of interest (blue area) based on the DES stimulation area (MEP weighted mean electric field > 30% of its maximum field strength).

Fig. 5

DES and TMS comparison: A) percentage of the overlap between the DES stimulation area (3.99 ± 0.46 cm²) and the stimulation area of the TMS for both the realistic model (red line) and the spherical model (blue line). Shown are mean ± standard error of mean over the six patients for the overlap between the DES and TMS stimulation areas for different thresholds of the TMS electric field. For increasing TMS electric field strengths an increasing percentage overlaps with the DES ROI. This effect is more pronounced for the realistic model than for the spherical model. B) Distance between the CoG of the TMS map and the CoG of the DES map for both the realistic and the spherical model. Shown are mean ± standard error of mean over the six patients for the distance between the DES and TMS CoGs for different thresholds of the TMS electric field. With increasing TMS electric field threshold, the distance between the TMS and DES CoGs decreases.