Spatial sensitivity of the optimal dose in TES to the skull and scalp conductivity specifications

Abstract

Introduction:Electrical impedance tomography (EIT) is a non-invasivetechnique that can be used to estimate the scalp and skull conductivityvalues. These values are required for modelling transcranial electricalstimulation (TES) to estimate the current density dose on the brain. Wepreviously found individual EIT estimates for four subjects using detailedhead models [1]. The question we address here is: what?s the impact in TESof using individual EIT estimates versus typical literature values?Methods:Usingfinite element simulations, we computed the maximumpossible TES dose for each element of the grey matter (GM) along itsnormal-to-cortex orientation, assuming afixed current budget of 1mA.Optimal current injection patterns where obtained using the reciprocityoptimization method [2]. The same procedure was followed for modelsusing bEIT estimated (Table I in [1]) and literature (scalp: 300mS/m, skull:8mS/m) conductivity values.Results:At each GM location, we computed the normalized differencebetween the maximum doses obtained with both conductivity specifica-tions. It is known that some deep regions can also get stimulated with TESat intensities of the same order as shallow targets due to the brain ge-ometry and the highly-conductive CSF paths (e.g. [2]). Despite this fact, wefound that TES dose on shallow brain regions, i.e. regions closer to theskull, were more sensitive (~70% difference) to scalp and skull conductivityspecifications than deeper regions (<20%).Discussion:Our results suggest that to estimate TES dose on shallowtargets, it is highly recommended to use scalp and skull conductivity es-timates as accurate as possible. On the other hand, deeper targets - nomatter if their doses are large or small - are less sensitive to scalp and skullconductivity mis-specifications.