Transcranial magnetic stimulation patterns in heterogeneous brain tissue: focality, reproducibility and true sham stimulation

Transcranial magnetic stimulation (TMS) has been gaining status as an innovative research and possibly therapeutic tool for non-invasive stimulation of brain tissue. However, optimal values of key stimulation parameters such as coil size, shape and orientation and current pulse characteristics remain largely undetermined. Further, the influence of brain size and shape as well as of the strong heterogeneity of brain tissue on stimulation efficacy is unclear. In order to calculate the current density distribution induced in brain tissue by TMS, a complete set of MRI images of the human brain is discretised over a cubic mesh, constructing a spatial map for tissue electrical conductivity. The magnetic induction field is calculated as a function of coil geometry and position as well as incoming current pulse characteristics, after which Maxwell's equations in integral form are discretised and solved over the mesh by a successive overrelaxation procedure. Findings are validated by recording motor evoked potentials from the right abductor pollicis brevis muscle from a healthy subject in a stereotaxic framework. The inhomogeneous electrical characteristics of brain tissue are seen to significantly influence stimulation patterns, as a symmetrical primary electric field results in a highly asymmetrical current density distribution. Also, several commonly used sham stimulation configurations elicit conductive patterns which achieve up to 40% of the strength of real stimulation. Further, variations in coil position of the order of a 7 degree tilt, which are expected to occur in non-stereotaxic stimulation, can alter the stimulation intensity by up to 38%. Knowledge of coil specifications alone is clearly not sufficient to control stimulation conditions. In accordance with our findings, several clinical studies observe measurable effects during sham stimulation, and the sensitivity of stimulation intensity to tiny coil rotations afford a partial explanation for the poor reproducibility and partial disagreements observed across TMS studies. Our findings bear direct relevance to any application of TMS, both investigative and therapeutic.