Harmonizing the stimulation dose of focal transcranial direct current stimulation across target sites

Focal transcranial direct current stimulation (tDCS) using center-surround electrode montages enables region-specific cortical targeting, and holds promise for both cognitive neuroscience and clinical interventions. However, systematic examinations of dose-response relationships and their regional differences are lacking, hampering informed selections of suited stimulation parameters. In this preparatory methodological study, we present a modeling-based framework to support harmonized empirical dose-response studies of focal tDCS across different target areas. It covers three steps: Determining the approximate electric field strength that had led to behavioral and physiological effects in related prior tDCS studies. In our case, this led to a field strength of 0.2 V/m on average across magnetic resonance images (MRIs) from 43 participants and eight target areas related to different cognitive and motor functions. Second, optimizing the radii of center-surround montages for each target area to - on average across participants - achieve the intended field strength while maximizing focality. An additional test of cross-sample generalization in an independent sample confirms that the intended target field strength is achieved on average for new participants. Third, the pre-determined montage radii and a method for the individualized positioning of the center-surround electrode montages are provided for prospective planning in empirical dose-response studies. By harmonizing the electric field strength between different target regions at the group level, but preserving inter-individual variability, our framework will enable systematic analyses to relate the field strength to behavioral and neuroimaging outcomes, and to assess differences of these relations across regions. The described computational tools are open-source, allowing other researchers to tailor our framework to their specific research questions; and are currently used in a multi-center study involving approximately 1000 datasets.