OBJECTIVES: Electric field modeling utilizes structural brain magnetic resonance images (MRI) to model the electric field induced by non-invasive transcranial direct current stimulation (tDCS) in a given individual. Electric field modeling is being integrated with clinical outcomes to improve understanding of inter-individual variability in tDCS effects and to optimize tDCS parameters, thereby enhancing the predictability of clinical effects. The successful integration of modeling in clinical use will primarily be driven by choice of tools and procedures implemented in computational modeling. Thus, the electric field predictions from different modeling pipelines need to be investigated to ensure the validity and reproducibility of tDCS modeling results across clinical or translational studies. METHODS: We used T1w structural MRI from 32 healthy volunteer subjects and modeled the electric field distribution for a fronto-temporal tDCS montage. For five different computational modeling pipelines, we quantitatively compared brain tissue segmentation and electric field predicted in whole-brain, brain tissues and target brain regions between the modeling pipelines. RESULTS: Our comparisons at various levels did not reveal any systematic trend with regards to similarity or dissimilarity of electric field predicted in brain tissues and target brain regions. The inconsistent trends in the predicted electric field indicate variation in the procedures, routines and algorithms used within and across the modeling pipelines. CONCLUSION: Our results suggest that studies integrating electric field modeling and clinical outcomes of tDCS will highly depend upon the choice of the modeling pipelines and procedures. We propose that using these pipelines for further research and clinical applications should be subject to careful consideration, and indicate general recommendations.
OBJECTIVES: Electric field modeling utilizes structural brain magnetic resonance images (MRI) to model the electric field induced by non-invasive transcranial direct current stimulation (tDCS) in a given individual. Electric field modeling is being integrated with clinical outcomes to improve understanding of inter-individual variability in tDCS effects and to optimize tDCS parameters, thereby enhancing the predictability of clinical effects. The successful integration of modeling in clinical use will primarily be driven by choice of tools and procedures implemented in computational modeling. Thus, the electric field predictions from different modeling pipelines need to be investigated to ensure the validity and reproducibility of tDCS modeling results across clinical or translational studies. METHODS: We used T1w structural MRI from 32 healthy volunteer subjects and modeled the electric field distribution for a fronto-temporal tDCS montage. For five different computational modeling pipelines, we quantitatively compared brain tissue segmentation and electric field predicted in whole-brain, brain tissues and target brain regions between the modeling pipelines. RESULTS: Our comparisons at various levels did not reveal any systematic trend with regards to similarity or dissimilarity of electric field predicted in brain tissues and target brain regions. The inconsistent trends in the predicted electric field indicate variation in the procedures, routines and algorithms used within and across the modeling pipelines. CONCLUSION: Our results suggest that studies integrating electric field modeling and clinical outcomes of tDCS will highly depend upon the choice of the modeling pipelines and procedures. We propose that using these pipelines for further research and clinical applications should be subject to careful consideration, and indicate general recommendations.