< Back to previous page

Publication

Using a 3D-MRI system for coil positioning in repetitive transcranial stimulation

Book Contribution - Book Chapter Conference Contribution

Abstract: In the therapy of depression, repetitive transcranial magnetic stimulation (rTMS) of the left dorsolateral prefrontal cortex (DLPFC) has been introduced. A crucial problem in rTMS-studies remains the exact localization of the coil with respect to the DLPFC. In most studies, the coil position is determined by identifying the subject's motor cortex and moving the coil 5 cm more rostrally from this point. This standard coil positioning method is based on the Talairach atlas and does not take in account individual anatomy. In our study we introduced an individualized method based on a 3D MRI of the head, the "3D-MRI system". The coil position was derived as the projection of the DLPFC on the skin, defined on a 3D Surface rendering of the brain relative to the nose and the right ear. To evaluate the importance of introducing individual anatomical information during coil positioning, we compared both localization techniques on 36 female volunteers. We found a mean shift between both positions of 27?14 mm. The individual positions found with the 3D-MRI system, show a variation of 18?8mm with respect to the mean coil position. Our results show the need for an rTMS-coil positioning method which incorporates anatomical information. The combination with MRI-imaging techniques can be an option.

Introduction: Several studies have shown the therapeutic value of repetitive transcranial magnetic stimulation (rTMS) in depression and other psychiatric disorders (2). Although the best stimulation region for rTMS-treatment is still an issue of debate, most neuroimaging findings suggest the left dorsolateral prefrontal cortex (DLPFC) as target site (1). Crucial in rTMS-treatment is the accurate positioning of the rTMS-coil over the stimulation area. Standard methods for coil positioning are derived from the Talairach atlas (3) and do not take in account any individual anatomical information. The aim of our study was to show the importance of incorporating individual anatomical information when determining the coil position by comparing a standard method, the "5-cm rule", with an rTMS-MRI combined method, the "3D-MRI system".

Materials and methods: We included 36 healthy female subjects (mean age 25?6 years) in the study. Exclusion criteria conformed the current guidelines for rTMS and MRI research. The local medical ethics committee approved of the study and all subjects gave informed consent.
The standard method for coil positioning is the "5-cm rule": the DLPFC is defined as 5 cm rostral to the motor cortex (Fig 1.A). In this case the motor cortex is identified by stimulating the brain with an rTMS-stimulator (Magstim Company, Wales, UK) by looking for the maximum response recorded with EMG over the right musculus abductor pollicis brevis.
To introduce individual anatomical information, all subjects underwent a T1-weighted MRI (3D-TFE, voxel size 1x1x1 mm) of the brain using a 1.5T Intera

Figure 1 - (A) Schematic representation of the 5-cm rule: the DLPFC is defined 5 cm rostrally as the motor cortex (MC). (B) Schematic representation of the 3D-MRI method: the DLPFC is defined by the distances from right ear to top (|R-T|), from nose to top (|N-T|) and from top to DLPFC (|T-D|). An example is shown in (D). (C) Comparison of both methods: the distance and angle between both positions are calculated as defined in the figure.

(Philips, Best, the Netherlands) MR scanner. All post
processing was done on a ViewForum (Philips, Best, the Netherlands) console. We located the DLPFC visually on the 3D surface rendering of the brain based on the known gyral morphology. The corresponding coil position was found by making the projection on the skin. On a 3D-reconstruction of the head we marked 4 reference points: right ear, left ear, vertex and nose. A fifth reference point, top, is found as the crossing between two axes: one from


right ear to left ear through the coil position and one from nose to vertex. For an accurate determination of the coil position on a patient's head, we measured the distance from nose to top, from right ear to top and from top to the coil position (Fig 1.B and 1.D).
To compare both positioning methods, the subjects were wearing a tight fitting bathing cap on which we drew two reference axes, one from right ear tot left ear and one from nose to atlas. We measured the coordinates of both positions relative to the vertex, the crossing of both reference axes. We calculated the distance between both positions and the angle between the axis passing both positions and an axis perpendicular to the reference axis from nose to atlas (Fig 1.C). We compared the positions found with the standard method to the respective positions found with the 3D-MRI system and we compared the individual positions to the mean position found with the 3D-MRI technique.

Results: Comparison between the 5-cm rule and the 3D-MRI system, shows a mean shift between both positions of 27?14 mm and an angle of -11?66° (Fig 2). In most subjects, the standard method positioned the DLPFC too rostrally. To measure the dispersal of the individual positions, we compared these individual positions with the mean position of the DLPFC as determined with the 3D-MRI system. We found a mean shift of 18?8 mm and a mean angle of 15?101°. These results demonstrate that the exact position of the DLPFC is highly individual.

Discussion: A previous comparison of the standard method to a neuronavigational method showed that the current positioning method used in rTMS-studies easily fail to target the DLPFC (5). Our study confirms this result. Where Uwe focused more on whether his method positioned the DLPFC more dorsally or rostrally in comparison to the standard method, we focused on the dispersal of the individual positions. This individual heterogeneity can be a reason for the heterogeneous outcome of rTMS-treatment effect studies. If we would like to improve the efficacy of this promising technique, we should use coil positioning techniques witch take the individual anatomy in account. Making the combination with imaging techniques like MRI can be a solution.

Conclusions: To improve the coil positioning in rTMS-therapy, new methods like the 10-20 EEG-system (4) and the neuronavigational system (6) are introduced. In the 10-20 EEG-system, the positioning of the DLPFC is relative to the electrode positions used in EEG-recordings. This method has still the drawback that it is determined from the Talairach brain and does not really account for anatomical differences between individuals. In the neuronavigational system, recordings of the head with a 3D-camera are coregisterd to MRI-images. This method is accurate and does take in account individual anatomy but it is very complex to perform. In our 3D-MRI system we made a combination of common software tools used on an MRI-site and tools used in psychiatry. The new positioning method is as easy to perform as the standard method but takes individual anatomy in account.
Book: BHPA
Publication year:2008
Keywords:fMRI, TMS
  • ORCID: /0000-0003-3647-4446/work/104771311
  • ORCID: /0000-0002-3601-3212/work/91494516