Transcranial Magnetic Stimulation (TMS) Coil Placement

All repetitive transcranial magnetic stimulation (rTMS), for investigational or clinically-approved treatment (Cohen et al., 2022), requires a method to position a coil on the scalp. rTMS coils are positioned according to application-specific cortical targets, such as the left dorsolateral prefrontal cortex (DLPFC) for depression treatment. Since anatomy varies, coil placement relies on varied approaches:

1. Relative to TMS evoked motor responses

A common approach. First, the M1 site may be identified based on motor evoked potentials (MEP). Then, approaches like the “5-cm rule” combine a functional evoked response (e.g. TMS-MEP) with a geodesic rule (e.g. “5 cm”) (O’Reardon et al., 2007).
Advantage: Simple to use and validated in clinical trials.
Limitation: MEP-based targeting does not fully account for anatomical variability at the therapeutic area (which is not M1).
Resources: How to find TMS motor threshold.

2. Scalp Measurements Only

Due to the cost and complexity of navigation systems, localization based on scalp measurements using external landmarks are common. The Beam-F3 method is the most widely used for DLPFC targeting (Beam et al., 2009; McClintock et al., 2017). An “Updated Scalp Heuristic” provides a formula to obtain geodesic measurements for four potential TMS scalp targets, including one left DLPFC site (Mir-Moghtadaei et al., 2022). A scalp mapping system, known as Tetra codes, has also been proposed to enhance precision of scalp-based TMS targeting (Li et al., 2024).
Advantage: Simple and validated in clinical trials.
Limitation: No based on individual brain shape.
Online tools: TMS Targets and ClincalResearcher.org

3. MRI-Guided Hardware Neuronavigation

The gold-standard approach uses individual magnetic resonance imaging (MRI) and stereotactic neuronavigation hardware to co-register the coil with the subject’s MRI in real-time (Comeau, 2014; Neggers et al., 2004; Ruohonen et al., 2010; Valter et al., 2024).
Advantage: Individualized, precise, and applicable to any brain target.
Limitation: Requires an MRI and navigation hardware.
Resources: Explanatory Video

4. Template-Based Hardware Neuronavigation

When individualized MRI data is unavailable, navigation hardware can still be used to map anatomical landmarks in 3D space in a "generic" head.. Algorithms then deform a standard head model, such as the Montreal Neurological Institute template (MNI; Mazziotta et al., 2001) to subject-specific head size and shape (Valiulis et al., 2018), yielding a subject-specific head model for target localization.
Advantage: Semi-individualized and applicable to any brain target.
Limitation: Requires navigation hardware.

5. TMS targeting using MRI-free head modeling

This approach aims to combine 1) the simplicity of scalp based methods such as Beam-F3 that do not require a MRI or neuronavigation hardware; with 2) a method to account for individual brain anatomy and targeting of any brain region. For example the Valter-MNI (Valter 2025) approach relies on the same three head measurements at Beam-F3, but then warms a head-model to the individual subject creating a targeting system for DLPFC or any brain region.
Advantage: Semi-individualized and applicable to any brain target.
Limitation: Not as precise as MRI and hard-ware based navigation.
Online tools: TMS Targets.