Simulating tDCS-induced electric fields in stroke patients: realistic-lesion head models are needed

Introduction: Transcranial direct current stimulation (tDCS) is tested as tool for post-stroke rehabilitation in aphasia, and individualized simulations of tDCS-induced electric fields (E-fields) can guide its application. However, the accuracy of simulations is challenged by complex and variable tissue properties of stroke lesions. Here, we assessed the impact of stroke lesions on tDCS-induced E-fields realistically in terms of lesion size, shape, and conductivity. Methods: Structural and diffusion MRI datasets of stroke patients with aphasia (n=13, six females, age=38-70 years) and age-matched healthy controls (n=13, eight females, age=24-76 years) from a previous study were analyzed. Simulated E-fields were first compared between healthy head models with and without artificial lesions homogenously filled with cerebrospinal fluid. Then, the effects of lesion heterogeneity were tested by comparing E-fields for models of stroke patients with homogenous versus inhomogeneous (realistic) lesion conductivity informed by diffusion-to-conductivity mapping. Results: Adding artificial lesions to healthy head models altered the magnitude of E-fields (|E|) near the target region (ROI) by up to 47%. Diffusion-to-conductivity mapping revealed substantial variability in lesion conductivities within and across patients. Modifying homogenous to realistic lesion models showed small to moderate |E| differences within the ROI depending on montage type, lesion size, and lesion-to-target distance. Conclusion: Stroke lesions affect tDCS-induced E-fields with substantial variability across montages and individuals. These findings support the use of head models that include realistic representations of the shape, size and conductivity of the lesions to improve the accuracy of individualized tDCS simulations and guide personalized stimulation protocols in stroke rehabilitation.