A Standardized Protocol for Functional Motor Mapping Using Navigated Transcranial Magnetic Stimulation

Summary

Here, we describe a standardized protocol for motor mapping using nTMS combined with diffusion tensor imaging (DTI)-based reconstruction of the corticospinal tract (CST). The protocol is reproducible, clinically feasible, and readily integrable into routine clinical workflows, providing a robust and valuable framework for motor pathway assessment, neuroplasticity research, and rehabilitation planning.

Abstract

Navigated transcranial magnetic stimulation (nTMS) is based on the integration of individual brain imaging data to determine the precise positioning of the stimulation coil, thus enabling anatomically guided stimulation of cortical targets. The interest of neuronavigation systems is well recognized in the optimization of coil positioning during repetitive TMS (rTMS) treatments. Furthermore, nTMS is increasingly applied for functional mapping of brain regions in different applications, such as the identification and delineation of eloquent motor and language areas before tumor resection. Besides its usefulness in optimizing neurosurgical procedures, nTMS mapping can also be a tool for monitoring cortical plasticity and quantifying the integrity of the motor system in various neurological diseases. This methodological paper presents a standardized protocol for motor mapping using nTMS, in combination with diffusion tensor imaging (DTI)-based reconstruction of the corticospinal tract (CST). This approach allows for precise delineation of eloquent motor cortical regions and their subcortical projections, and the detection of functional reorganizations in patients with adjacent lesions. When integrated into presurgical planning, this method provides guidance for individualized surgical strategies aimed at maximizing lesion resection while preserving motor function. The protocol presented here is reproducible, clinically applicable, and suitable for integration into routine workflows. It constitutes a promising tool for neuroplasticity research and rehabilitation planning.

Introduction

Maximizing the extent of resection in motor-eloquent brain tumors while minimizing postoperative motor deficits remains a central challenge in neurosurgery. Intraoperative direct electrical stimulation (DES) mapping is the "gold standard" technique to provide reliable anatomo-functional information regarding the cortical and subcortical representation of motor pathways^1,2,3,4,5. However, for preoperative planning, risk stratification, and optimal patient counselling, it is crucial to delineate the individual functional anatomy before surgery. The relationship between anatomy and function in cortical motor areas cannot be inferred from conventional structural brain magnetic resonance imaging (MRI), as brain tumors may induce significant anatomical distortion or plastic reorganization of motor networks.

Transcranial magnetic stimulation (TMS) was introduced as a non-invasive method to probe the motor cortex⁶ and was later adapted for functional mapping of the motor cortex^7,8, including in preoperative testing by recording motor evoked potentials (MEPs) from different muscles with surface electromyography^9,10,11. Early non-navigated TMS protocols were technically demanding and lacked anatomical accuracy. Subsequent integration with individual MRI data and electric-field-based navigation enabled precise guidance of stimulation sites, improving anatomo-functional accuracy^12,13,14 and reproducibility^15,16. By directly eliciting MEPs, navigated TMS (nTMS) provides millisecond-scale temporal resolution and sub-centimeter spatial localization of corticospinal output with good concordance with intraoperative DES^17,18,19. Image-guided nTMS is safe, well tolerated^20,21, and approved by the Food and Drug Administration (FDA) for presurgical functional mapping of the motor cortex for more than 15 years²².

In motor mapping, cortical representations are delineated by sampling MEP amplitudes across targeted stimulation sites to construct patient-specific motor maps²³. Compared with task-based functional MRI (fMRI), nTMS shows closer spatial agreement with intraoperative DES^24,25,26. While intraoperative decisions ultimately rely on DES when lesions abut or invade the motor areas, preoperative nTMS provides valuable complementary information by exporting stimulation-positive sites as seeds for diffusion tensor imaging (DTI) reconstruction of the corticospinal tract (CST). This approach is particularly useful to assess corticospinal integrity when tumors mainly affect the motor tracts in the subcortical white matter^27,28. Moreover, preoperative nTMS motor mapping has shown good positive predictive value^29,30 and high negative predictive value^29,30,31, with improved surgical outcomes^17,18,19,32. It has also recently been proven as an effective tool to assess postoperative motor function^31,33. For these reasons, nTMS motor mapping is increasingly used for both preoperative evaluation and postoperative follow-up in neurosurgery. Methodological recommendations for cortical mapping with nTMS have been published in 2017³⁴. In light of these recent studies and the integration of modern imaging techniques, this methodology can now be refined to provide more accurate guidance for clinical and research practice.

In this paper, we present a standardized protocol for performing motor mapping with nTMS, combining different techniques to evaluate preoperative cortical and subcortical representations of motor pathways for tumor resection planning under real-life clinical conditions.

Protocol

This study was conducted in accordance with national and international ethics guidelines for human research. Retrospective analysis of anonymized patient data collected during routine care was performed with informed consent obtained at the time of care, in accordance with French regulations. Demonstration data from healthy subjects, who are co-authors of the manuscript, were included with written informed consent for participation and publication of data and images. This is the current protocol used in Henri Mondor Hospital (Créteil, France) and Aarhus University Hospital (Denmark) for preoperative planning in brain tumor surgery.

1. Acquisition of neuroimaging data for neuronavigation

1. Verify the absence of contraindication to nTMS and MRI using medical records and patient interview, including intracranial ferromagnetic device, uncontrolled epilepsy, pacemaker, pregnancy, or breastfeeding³⁵.
2. Acquire a high-resolution anatomical brain image that includes both ears and cranial vertex (without folds or deformation from MRI headphones) to allow accurate brain reconstruction by the neuronavigation system.
   1. Use the following recommendation for the MRI sequence:
      3D T1-weighted (T1w) anatomical gradient-echo
      1-mm isotropic voxels (or less)
      ≥1.5-Tesla MRI system (3 T preferred).
   2. Alternatively, use these acceptable sequences:
      3D-FLAIR
      Contrast-enhanced 3D T1w
3. Acquire diffusion-weighted imaging (DWI) before contrast injection for subsequent Diffusion Tensor Imaging (DTI)- based tractography³⁶.
   1. Use the following minimum acquisition parameters³⁷:
      Isotropic 2-mm voxels
      Diffusion-encoding directions: ≥ 25
      b-value: ≈ 800 s/mm²
      ​Non-diffusion-weighted images: ≥ 3 b0 volumes (b = 0 s/mm²)
   2. Use the following recommended parameters (for improved tensor estimation and tractography):
      Diffusion-encoding directions: ≥ 64
      b-value: 1000 s/mm²
      Higher spatial resolution (≤ 2 mm isotropic)

2. Prepare the subject

1. Import the subject's anatomical MRI image into the neuronavigation system to generate a 3D brain reconstruction.
2. Mark the key anatomical points on the MRI within the neuronavigation software (nasion, right ear, left ear).
   1. Use the root of the crus helicis for higher precision.
   2. Alternatively, use the tragus, but its larger surface can increase co-registration mismatch.
      NOTE: To shorten the motor mapping, these preparatory steps can be performed before installing the subject in the room.
3. Position the subject on a comfortable armchair, with a slight recline (20-30°) to reduce back tension³⁸. Adjust the headrest to support the head and neck at the inion.
4. Check for metallic objects in the head and neck areas (e.g., earrings, hairpins, piercings) and remove them before starting the procedure.
5. Prepare the skin on the forehead for head-tracker placement.
   1. Clean the skin using alcohol pads or mild abrasive gel.
   2. Ensure the skin is fully dry before placing the tracker.
6. Place the head tracker on the forehead so it remains stable throughout the stimulation session.
   1. Position it above the eyebrows and below the hairline.
   2. Place it either in the middle or slightly lateral.
   3. Fix the tracker using its adhesive surface or with an elastic band.
7. Co-register the key anatomical points on the patient with the imported image in the neuronavigation software (see Figure 1).
   1. Use the digitizing pen to mark the anatomical landmarks.
   2. Ensure that earlobes are free from the headrest to avoid any displacement of the ear landmarks³⁹.
   3. If the ear anatomy appears distorted on MRI (e.g., folded ear pinna), redefine the corresponding point on the image before digitizing.
8. Once completed, the software validates the three fiducial points if the mismatch error is below 3 mm. If the mismatch error is too large, try the following steps in order:
   1. Digitize the key anatomical points on the patient a second time.
   2. Redefine the left and right ear anatomical points on the MRI.
   3. Digitize while gently pressing the helix of the earlobe, since MRI headphones may have displaced the ear by a few millimeters.
9. Refine the registration by digitizing additional scalp points (scalp-surface matching).
10. Validate the co-registration, with a co-registration error below 3 mm (2 mm preferred). If the mismatch exceeds 3 mm, repeat steps 2.7-2.9.

Figure 1: Co-registration of the patient's head with the anatomical MRI. Left side: Landmark-based registration. Upper panels: Identification of the anatomical landmarks on the MRI (left ear, nasion, right ear) within the neuronavigation software. Lower panels: Digitization of the landmarks on the patient using the digitizing pen.Right side: Surface matching refinement using additional scalp points. Please click here to view a larger version of this figure.

3. Preparation of mapped muscles

1. Give earplugs to the subject and wear protective earmuffs during stimulation.
2. Prepare the skin over the target muscle by gently scraping the skin with alcohol pads and/or cotton pads with mild abrasive gel.
3. Place surface electrodes on the muscles of interest in a belly-tendon montage, as for routine clinical MEPs. Up to six different muscles can be mapped simultaneously.
4. Place the ground electrode on a neutral site, such as the shoulder's stump, the dorsal hand surface, or the medial tibia surface.
5. Connect all electrodes to the EMG amplifier.
6. Start EMG acquisition to display continuous EMG of all channels and verify that the muscles are at rest.
7. Check that EMG channels are free from excessive 50/60 Hz noise (< 50 µV). If the electric hum is excessive, try the following steps in order:
   1. Verify that the electrodes are firmly attached to the skin, without any detachment.
   2. Reposition the electrode cables inside the chair to avoid contact with metallic parts or with the floor.
   3. Move the distal part of the electrode leads away from the neuronavigation system and from AC power sources.
   4. Replace the electrodes and reapply them with a different cable orientation (see steps 3.7.2 and 3.7.3).
   5. Disconnect the chair from its power supply.
   6. Place the ground electrode on the same limb as the mapped muscles.
   7. Repeat the steps in order until the noise is reduced below the threshold.
8. Once the 50/60 Hz noise is minimized, restart EMG recording to reset the baseline.
9. Once these preparation steps are cleared, proceed with coarse mapping of the selected muscles.
   ​NOTE: A standard mapping session should include at least one muscle per upper-limb segment and two lower-limb muscles. Table 1 lists the commonly mapped muscles, which should be adapted according to the lesion location and the patient's clinical presentation³⁴.

Limb	Muscle	Alternative(s)
Hand	First Interosseus Dorsal (FDI)	Abductor Pollicis Brevis (APB)
	Abductor Digiti Minimi (ADM)
Forearm	Flexor Carpi Radialis (FCR)	Extensor Carpi Radialis (ECR)
Arm / Shoulder	Biceps	-
	Deltoid
Leg	Tibialis Anterior (TA)	Soleus (SOL)
Foot	Abductor hallucis (AH)	Medial Plantar (MP)
Face	Orbicularis Oris	Nasalis

Table 1: Suggested muscles for motor mapping.

4. Coarse mapping to identify the hotspot and determine the Resting Motor Threshold (RMT)

1. On the rendered brain volume in the software, adjust the peeling depth between 15-25 mm deep to the scalp, case-by-case, to best reveal the cortical anatomy. The goal is to visualize the precentral and postcentral gyri, the central sulcus, and the superior and inferior frontal sulci.
   NOTE: Identification of the precentral gyrus is easier when the subject presents an "omega-shaped" hand knob^40,41. However, this landmark is inconstant^42,43. In such cases, several methods are recommended to identify the precentral gyrus^43,44,45.
2. Start the stimulator unit.
3. Position the stimulation (figure-of-eight) coil tangential to the scalp (see Figure 2).
   1. Stabilize the coil with one hand on the handle and the other on the coil to keep stable contact with the scalp during repositioning.
   2. Use the neuronavigation assistance (coil-angle, coil-to-head distance, tilt indicators) to ensure accurate coil positioning over each stimulation site.
   3. Maintain a stable induced Electric Field (EF, V/m) by avoiding coil tilting.
   4. Adopt a comfortable posture as the coil can be heavy. Use a cable-holding arm to reduce cable tension while keeping the coil freely movable.
4. Stimulate at an intensity adjusted to elicit responses within the 100-500 µV (peak-to-peak) amplitude range⁴⁶.
   NOTE: This is usually achieved between 35% and 45% of the Maximal Stimulator Output (MSO) for upper limbs and between 50% and 80% of the MSO for lower limbs. However, this range of values applies to healthy subjects and may be higher when the tumor infiltrates motor regions.
5. Note that the coil orientation for the coarse mapping (as well as the fine mapping) depends on the limb mapped (see Figure 3):
   1. For the upper limb and the face: keep a coil orientation perpendicular to the central sulcus (sulcus-aligned), to maintain an induced electrical current in a posterior-to-anterior direction⁴⁷.
      1. For the upper limb: start stimulating over the upper part (shoulder) or the middle part (forearm and hand muscles) of the posterior wall of the hand-knob, facing the superior frontal sulcus.
      2. For the face: start stimulating over the posterior wall of the precentral gyrus facing the inferior frontal sulcus. Check the response latencies to ensure they originate from the corticobulbar pathways. Facial MEPs have a latency of 7-13 ms, whereas direct muscle response (jaw jerk) induced by nTMS has a latency of about 3-4 ms.
   2. For the lower limb: keep a coil orientation perpendicular to the sagittal midline, with an induced electrical current in a mid-to-lateral direction³⁴. Alternative coil orientations include parallel to the sagittal midline^48,49,50 and/or perpendicular to the folds of the paracentral lobule and the precentral gyrus.
6. Perform stimulations over the precentral gyrus.
   1. Space stimulation points 1-2 mm apart, either visually or using a stimulation grid.
   2. When performed visually, sample three parallel lines across the gyrus. This is usually sufficient.
   3. Space each stimulation by at least 1.5 s, preferably with a randomized interstimulus interval.
7. If no responses are obtained, increase the stimulus intensity by 10% relative to the starting value and repeat as before.
8. Stop the coarse mapping once 20-30 responses per muscle are recorded.
9. Review all MEPs to exclude contaminated recordings.
10. Identify the "hotspot" for each muscle. The "hotspot" is the stimulating point that elicits the MEP of the largest amplitude. To ensure reliable hotspot definition⁵¹
    1. Display the recordings of each muscle using a normalized color scale.
    2. Locate the area containing MEPs of the highest amplitude.
    3. Sort MEPs by amplitude, from highest to lowest.
    4. Select the highest-amplitude MEP within this area, avoiding abnormally high single responses (usually the first 2 MEPs).
11. For each muscle, select the hotspot to determine the resting motor threshold (RMT). This will save the coil position and orientation throughout the RMT determination process, ensuring reliable measurement⁵².
12. Determine the RMT for each muscle separately, either using a threshold-hunting technique⁵³ or by identifying the lowest stimulus intensity (% MSO) that elicits MEPs ≥ 50µV in 5 out of 10 consecutive trials (Rossini-Rothwell method)⁵⁴. Use the RMT of each muscle as a reference for setting the stimulus intensity during the fine mapping.

Figure 2: Experimental nTMS setup. The subject is seated with slight recline and arm support, with EMG electrodes placed over the target muscles. The operator holds the figure-of-eight coil stabilizing it to maintain tangential scalp contact, while monitoring the induced electric field (arrows: direction, circle: intensity) and the induced MEPs. Please click here to view a larger version of this figure.

Figure 3: Neuronavigation interface during mapping. Real-time feedback on coil position (junction of the blue and red arrows), coil tilt, electric field direction (blue-to-red arrow) and field intensity (colored surrounding ring) ensuring accurate stimulation at each cortical site. Upper panel: Coarse mapping of the upper limb, with the coil oriented perpendicular to the central sulcus. Lower panel: Fine mapping of the Tibialis Anterior, with the coil oriented perpendicular to the sagittal midline. Please click here to view a larger version of this figure.

5. Fine mapping

1. Ensure that the subject is fully relaxed, without involuntary muscle contraction.
2. For each muscle, perform the stimulation at 105-110% of its RMT.
   1. Use the same coil orientation as during coarse mapping (see steps 4.5 and 4.6).
   2. Reduce the spacing between stimulation points (4-6 parallel lines per gyrus).
   3. Maintain an interstimulus interval ≥ 1.5 s, preferably randomized.
3. Delineate functional motor maps as cortical areas where nTMS generates MEPs ≥ 50 µV (peak-to-peak).
   NOTE: For lower limb mapping, an alternative is to start at 110% of the upper limb RMT, and to adjust the EF by steps of ± 10 V/m until consistent MEPs are obtained³⁴.
4. Perform the stimulation until motor maps are bordered by one or two consecutive lines of negative sites that fail to elicit MEPs.
   1. If a clear negative border is not obtained, extend the sampling, maintaining the same spacing, until responses disappear reliably.
   2. If positive responses expand in unusual regions, check and adapt coil-angle, EF, and RMT.
      NOTE: The number of points per muscle can vary (30 to 100 pulses) according to the muscle cortical representation and the degree of tumor-induced brain shift.
5. Avoid coil orientations that generate abnormal MEP locations or amplitudes. In particular, a 45° orientation (relative to the midline) can produce upper-limb MEPs very anteriorly and may not be representative of accurate motor cortical representation⁴⁷.
6. Ensure that the motor maps are elliptic, with a few negative sites inside. For negative stimulation points within the motor map, perform additional stimulations at different moments during the evaluation to control for transient changes in motor cortex excitability.
7. If many negative responses (<50 µV) occur during the mapping, try the following steps in order:
   1. Ask the subject to stay awake, since it often reflects a reduction in vigilance state.
   2. Check that the stimulation intensity has not decreased.
   3. Consider repeating the RMT, as the initial value may have been influenced by a transient hyperexcitability state.
8. If many abnormally high-amplitude MEPs appear (> 1000 µV) and the map expands excessively, try the following steps in order:
   1. Ask the subject to relax the limb, even by showing ongoing muscle activity if necessary (signal feedback).
   2. If muscle activity remains, ask the subject to shake the limb or move it in a more relaxed position. If needed, apply a concentric passive movement on the tested muscle (e.g. with an object for the hand muscles and the abductor hallucis, or with the foot support for the tibialis anterior).
   3. Consider repeating the RMT, as the initial value may have been influenced by a transient state of motor cortex hypoexcitability.

6. Post-processing analysis of MEP data and export

1. Review and adjust MEPs for each muscle.
   1. Open the MEP review panel or signal viewer in the neuronavigation software.
   2. Inspect each recorded MEPs to correct amplitude and latency and adjust markers if needed.
2. Exclude artefactual or abnormal stimulation points.
   1. Open the stimulation list or mapping workspace in the software.
   2. Remove stimulation trials containing artefacts or incorrect coil positions (see Figure 4).
3. Display the motor map for each muscle in a binary format (positive/negative; above/below 50 µV).
4. Export the positive stimulation points at 15, 20, and 25 mm depth inbinarized DICOM format. Use these files for fiber-tracking to reconstruct the CST, using the positive stimulation points as seeds for the tractography.
5. To measure other cortical maps parameters (center of gravity, map density, motor map size), export the data at the stimulation peeling depth or at 20 mm (standard peeling depth)^{25,55,56,57,58}.

Figure 4: Post-processing analysis of MEP data. MEP traces are reviewed to correct amplitude and latency markers and exclude artefactual trials (right panel: example of a trial contaminated by ongoing EMG activity). The two stimulations (red circles) illustrate "abnormal responses" occurring in the negative area, likely related to coil-orientation effects. Please click here to view a larger version of this figure.

7. Post-processing analysis of motor mapping

1. Import the DICOM of the motor maps into an image analysis software suitable for neurosurgical neuronavigation for brain tumor removal.
2. Register the anatomical image (T1w) with the motor maps DICOMs and the DWI. Import and register additional images if necessary (e.g., FLAIRw, SWI, T1w-gadolinum enhanced).
3. Generate objects from the motor map DICOMs and enlarge them by 1-2 mm to improve sensibililty⁵⁹.
4. Crop the motor maps to remove ears and nasion to prevent abnormal fiber reconstruction during fiber-tracking.
5. Draw an ending ROI manually at the inferior pontine level, ipsilaterally to the mapped hemisphere.
6. Perform the fiber-tracking, using the motor map ROIs as seeds, and the pontine ROI as the endpoint. Commonly used tractography algorithms include deterministic streamline tracking or probabilistic tractography depending on the clinical question and the fiber-tracking outcomes.
   NOTE: When using open-source diffusion software, several preprocessing steps are required before tractography (denoising, Gibbs artifact correction, movement and distortion correction, B1 bias-field correction, tensor fitting, and FA map generation).
7. Adjust the parameters of the fiber-tracking on a case-by-case analysis. Recommended parameters are a minimum of 110-120 mm length, a maximum angulation of 30°, and a FA set at 75% the FA Threshold (FAT, corresponding to the FA at which the first CST fibers become visible)^60,61.
8. Segment the brain tumor on other images (e.g. FLAIR, gadolinium T1w) and create a corresponding object.
9. Display the CST either for each limb's part (in different colors) or for the entire motor mapping.
10. Integrate all the data (cortical seeds, CST, brain tumor object) into the OR navigation software for neurosurgery.

Results

We present representative steps and the results of motor mapping obtained in different healthy subjects and in patients who underwent motor mapping in a clinical environment, using our neuronavigated TMS system. CST reconstruction was performed using an imaging-processing software suitable for neurosurgery planning, capable of multimodal image registration and DTI-based tractography. The neuronavigation system integrates a navigated figure-of-eight coil, a stereotaxic camera, an EMG amplifier, and provides real-time visualization of the induced Electrical Field on the 3D brain reconstruction using an individualized multi-sphere head model.

Figure 5 displays the RMT determination at the hotspot determined from the coarse mapping. The coil position and orientation are maintained at the exact same location throughout the entire procedure with the help of the neuronavigation target. Figure6 displays a motor mapping of a healthy subject. The left lower limb (thigh, leg, foot), upper limb (shoulder, forearm, hand), and face were mapped. Positive stimulation sites (color-coded by MEP amplitude) and negative sites (gray) delineate the motor cortical representation. Figure7 displays the motor mapping and CST reconstruction in a patient with lung cancer metastasis involving the premotor region and revealed by an upper limb motor deficit.

Figure 5: Coarse mapping and RMT determination at the hotspot (First Interosseus Dorsalis) in a healthy subject, using neuronavigated TMS. The hotspot, identified using coarse mapping (lower left panel), is selected as the target for the RMT determination. The coil position and orientation are maintained at the exact same location throughout the entire procedure, with the help of the neuronavigation target (lower right panel). Motor-evoked potentials (MEPs) are acquired with continuous EMG traces and epoch responses. Please click here to view a larger version of this figure.

Figure 6: Motor cortex mapping of lower limb, upper limb, and face muscles using neuronavigated TMS. Muscles recorded in the lower limbs: quadriceps femoris (green), tibialis anterior (orange), abductor hallucis (yellow). Muscles recorded in the upper limbs: abductor digiti minimi (green), flexor carpi radialis (orange), deltoid (yellow).Muscles recorded in the face: Nasalis (blue), Triangularis (purple). Please click here to view a larger version of this figure.

Figure 7: Motor cortical mapping and CST reconstruction for neurosurgical planning. nTMS-motor mapping (left panel) and nTMS-guided reconstruction of the corticospinal tracts (right panel) in a patient with brain metastasis (white) from lung cancer. Muscles recorded: abductor hallucis (purple), tibialis anterior (blue), deltoid (yellow), flexor carpi radialis (red), first interosseus dorsalis (green), orbicularis (cyan). Please click here to view a larger version of this figure.

Discussion

In this paper, we present a standardized and reproducible protocol for functional motor cortical mapping with nTMS directly applicable to preoperative surgical planning. By combining neuronavigation with the subject's anatomical brain reconstruction, this standardized protocol makes it possible to identify and delineate motor-eloquent cortical regions during an exam lasting less than 90 min, depending on the number of muscles studied. This approach is particularly relevant in patients with motor-eloquent tumors, where anatomical reconstruction of the CST is often limited by two factors: (i) anatomical displacement due to mass effect and/or edema and (ii) functional reorganization of motor representations. Anatomical seeding tractography based on fixed anatomical landmarks can therefore be misleading in localizing the cortical origin and propagating errors throughout the fiber tracking. Functional motor cortical mapping addresses this issue by using nTMS-positive sites as cortical seeds, thereby anchoring tractography to the patient's current motor map that drives corticospinal output. During post-processing analysis, the cortical ROIs derived from the motor maps should be enlarged by 2-3 mm to mitigate fusion-associated mismatch and to standardize ROI volume (0.9 ± 0.1 cm³), reducing operator and between-subject variability and improving CST tractography comparability⁵⁹. Compared with landmark-based tractography, nTMS-seeded tractography yields more plausible and somatotopically consistent CST reconstructions, with fewer aberrant streamlines and lower inter-rater variability^27,61,62. Compared with fMRI-based seeding, nTMS-based tractography also produces more plausible reconstructions and higher interrater consistency in patients with tumors adjacent to the CST²⁵. It also allows the extraction of several metrics from the nTMS-motor mapping and the nTMS-seeded CST, which may serve as a predictive factor of postoperative motor outcome. At the cortical level, the presence of nTMS-responsive sites within the tumor has been associated with an increased risk of motor deficit, with a positive predictive value ranging from 50-90%^30,63,64,65. In contrast, resection of nTMS-negative sites is considered safe, with a high negative predictive value ranging from 90-100%^30,31,65. At the subcortical level, a tumor-to-tract distance <8-12 mm has been identified as a critical threshold associated with an elevated risk of post-operative deficit, as long as the tumor does not invade the precentral gyrus^{66,67,68,69,70,71}. Additionally, microstructural alterations of the nTMS-seeded CST (decreased Fractional Anisotropy with increased Mean Diffusivity) have also been proposed as further risk factors for post-operative deficit⁷⁰. Finally, the use of nTMS-based tractography has been associated with a greater extent of resection and prolonged survival while preserving motor function, supporting its integration into preoperative planning⁷².

During motor mapping, a key parameter that strongly influences the spatial distribution of MEPs and the interpretability of motor maps is the stimulation intensity (SI). Higher SI increases response probability and spatial spread (risking false positive responses), whereas insufficient SI increases the risk of false negative responses. To minimize this bias, the SI should be scaled relative to the RMT and, when possible, adjusted to maintain a stable target EF. In practice, near-threshold SI strikes a balance between sensitivity and specificity and provides conservative maps close to direct electrical stimulation mapping. On the other hand, choosing a supra-threshold SI (e.g., 120% RMT) can be justified when clinical safety prioritizes sensitivity at the map margins, acknowledging that higher SI systematically expands the motor map⁷³. In the context of mapping multiple muscles, the use of a single SI may bias the mapping toward the lowest-threshold muscle, as adjacent muscles could have different excitability profiles. Accordingly, RMT should be estimated for each muscle⁷⁴. On the other hand, significant changes in cortical excitability, reflected by unexpected changes in MEP amplitudes, may occur during a motor mapping session, requiring re-estimation of the RMT and adjustment of the SI.

The use of stimulation grids during motor mapping helps standardize spacing and facilitates map quantification (i.e., by counting active squares). However, grid size directly shapes the results: large squares may overestimate map size, whereas small squares increase the risk of undersampling. Recent evidence suggests that nTMS mapping can be performed without grids, using an anatomy-guided approach with denser stimuli near the anatomical landmarks and map edges⁷⁵.

Several quantitative parameters can be derived from motor mapping, such as the center of gravity (CoG), motor map area, and volume. The CoG is defined as the amplitude-weighted location in coordinates which represents the center of the motor representation⁵⁸. Serial examinations have shown shifts in CoG in brain tumor patients^76,77,78, capturing evidence of functional reorganization over time in the motor cortex. Motor map area and volume represent the spatial extent of the motor representation. Area is commonly derived either by counting the active squares on a stimulation grid or by using spline interpolation in grid-free stimulation, which connects the positive stimulation points with smooth polynomial curves to generate continuous surface or volume⁵⁶. These metrics can be monitored longitudinally (follow-up study or assessment of an intervention) or compared to the contralesional hemisphere to investigate cortical motor plasticity^79,80,81,82. Quantitative motor mapping metrics have the potential to be extended beyond neuro-oncology, providing biomarkers of the motor system integrity and disease-related plasticity in neurological diseases^55,83.

Although nTMS is now well established for preoperative motor mapping, several limitations should be acknowledged. First, the accuracy of co-registration and cortical mapping remains partly operator dependent. Proper training in coil handling, head-tracker stability, and prompt adjustment of stimulation are required to ensure reliability and reproducibility of the technique, although previous studies have shown that nTMS provides reliable motor topography with good inter-operator agreement between expert and novice examiners⁸⁴. A second limitation relates to the influence of perilesional edema and mass effect on tractography. Excessive perilesional edema can reduce the accuracy of nTMS-based CST reconstruction, particularly in voxels adjacent to the lesion⁸⁵. Similarly, discrepancies between preoperative datasets and the real intraoperative anatomy may occur due to intraoperative brain shift^86,87. Because brain shift cannot be fully prevented - especially in tumors with important mass effect - the accuracy of nTMS-derived motor regions (both cortical and subcortical) may decrease during the later stages of resection. Several strategies can mitigate these inaccuracies, including limiting unnecessary cortical exposure, repeatedly checking superficial anatomical landmarks⁸⁸, and using intraoperative imaging such as MRI, ultrasound, or CT, combined with brain deformation correction^89,90,91,92. Finally, regarding safety, nTMS has demonstrated a favorable safety profile in patients with tumor-related epilepsy. In large series, stimulation-induced seizures are rare or absent during preoperative mapping⁹³, supporting the safety of this technique when appropriate precautions are taken.

Overall, nTMS provides clinically useful functional information to surgical planning and opens the path to longitudinal studies of motor-system plasticity in various neurological or psychiatric diseases.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Elements software	BrainLAB AG, Munich, Germany		Image processing software and OR neuronavigation software
Neuronavigation TMS system	Nexstim, Helsinki, Finland	NBS 5.1 system	Navigated TMS system with figure-of-eight coil and EMG amplifier
Surface electrodes for EMG recording	Natus, Middleton, WI, USA	9013L0453	For EMG recording