Anisotropic conductivity modeling for tDCS in Parkinson’s disease using multidimensional diffusion MRI

In a modeling study of 29 adults—12 with Parkinson’s disease and 17 controls—multidimensional diffusion MRI changed predicted transcranial direct current stimulation electric-field magnitudes by no more than 6.6% compared with conventional single-shell diffusion tensor imaging. The investigators described their approach as the first multidimensional-diffusion tDCS conductivity model and compared it with DTI and isotropic models across four electrode montages. The results demonstrate the feasibility of incorporating multidimensional diffusion MRI into tDCS simulations, but they do not show that it produces more accurate stimulation dosimetry.

All three models used the same head mesh, electrodes and electric-field solver, allowing the researchers to compare how different representations of brain conductivity affected the predictions. White-matter conductivity was conventionally estimated from single-shell DTI, while q-space trajectory imaging supplied a mean tensor expected to have less kurtosis bias. The principal difference between the two anisotropic models was tensor orientation, which varied by 18 to 25 degrees across white matter. Despite those directional differences, their predicted electric-field magnitudes differed by 6.6% or less under the tested conditions.

An exploratory comparison found no significant differences in predicted electric fields between participants with Parkinson’s disease and controls in any brain region or across the four montages. The authors characterized this group comparison as underpowered, so the finding should not be interpreted as evidence that the two populations receive equivalent stimulation exposure. The small sample also did not provide ground-truth measurements of brain conductivity or tDCS electric fields. The study therefore characterized the sensitivity of model predictions rather than establishing improved dosimetric accuracy.

Whole-brain predicted electric field correlated with brain stiffness measured by magnetic resonance elastography, with a partial correlation of 0.58. After cerebrospinal fluid morphology was included in the statistical model, this relationship was no longer significant, and the partial correlations fell to 0.06–0.09. The result shows that the association weakened after statistical adjustment for cerebrospinal fluid morphology; it does not by itself establish a causal influence of that anatomy. Based on the findings, the authors concluded that Parkinson’s disease modeling efforts should emphasize cerebrospinal fluid- and atrophy-aware head models and dose normalization rather than automatically adopting a more complex diffusion tensor. Overall, the choice of diffusion tensor was a second-order influence on predicted electric fields compared with individual anatomy under the conditions examined.

Subject of Research: Comparison of multidimensional diffusion MRI and DTI models for predicting tDCS electric fields in Parkinson’s disease

Article Title: Anisotropic conductivity modeling for tDCS in Parkinson’s disease using multidimensional diffusion MRI

Article References: Osorio Jurado, S., Skorpil, M., Svenningsson, P., Moreno, R., & Olsson, C. (2026). Anisotropic conductivity modeling for tDCS in Parkinson’s disease using multidimensional diffusion MRI. Frontiers in Human Neuroscience, 20, Article 1972676. https://doi.org/10.3389/fnhum.2026.1972676

Image Credits: AI Generated

DOI: 10.3389/fnhum.2026.1972676

Keywords: tDCS, Parkinson’s disease, multidimensional diffusion MRI, q-space trajectory imaging, diffusion tensor imaging, electric-field modeling, brain conductivity, MRI elastography, cerebrospinal fluid, white-matter anisotropy, Anisotropic, conductivity