Computational Modeling Analysis
E-field simulation was performed using SimNIBS (Version 4.0, Copenhagen, Denmark) (Saturnino, Madsen, and Thielscher 2019), an open-source software that allows the estimation of the tDCS-induced E-field distribution in individual brains. Firstly, head models of each participant were generated using charm (from individual T1- and T2-weighted structural MRI data (Puonti et al., 2020)), which segments the MRI images into nine distinct tissue types. For this analysis, the default conductivity values for each tissue were applied. Subsequently, the software generated a 3D tetrahedral mesh structure for each segmented tissue, allowing the simulation of the tDCS-induced E-field in each participant’s brain. Segmentations were manually verified to investigate any potential errors.
The tetrahedral head meshes resulting from the segmentation procedure were used to simulate the E-field distribution resulting from the two active tDCS protocols (1.5 mA and 3 mA). The sham protocol was not taken into consideration for our analysis as it does not produce an E-field in the head. To simulate the E-field in the brain, the tDCS set-up applied was identical as used in the experimental sessions. The electrodes (5x5cm) were placed over the MNI coordinates retrieved from neuronavigation, pointing towards Cz, with a thickness of 5 mm (electrode + conductivity paste). SimNIBS scripts were executed in MATLAB (version R2022a).
Finally, the E-field magnitude within predefined brain regions of interest were extracted. These regions were the bilateral DLPFC, bilateral ACC, bilateral insula and bilateral amygdala. The DLPFC was extracted from the Sallet atlas (Sallet et al., 2013), whereas the ACC, insula and amygdala were extracted from the Brainnetome atlas (Fan et al., 2016) (Sup. Material - Appendix 1).
Values analyzed in this study were the mean magnitude of the electric field (magn-E component).