Association between tDCS induced GABA change and estimated electric field in the cortex

Transcranial direct current stimulation (tDCS) has wide ranging applications in neuro- behavioural and physiological research, and in neurological rehabilitation. However, it is currently limited by inter-subject variability in responses, which may be explained, at least in part, by anatomical differences that lead to variability in the actual electric field in the cortex. Our aim was to examine whether the variability in electric fields, estimated using computational simulations, explains the variability in tDCS induced GABA changes measured using magnetic resonance spectroscopy (MRS). Data from five studies (total N = 56 complete cases) were combined. The anode and cathode were placed over the left M1 (3 studies, N = 24) or right temporal cortex (2 studies, N = 32), and contralateral supraorbital ridge respectively. GABA to total Creatine ratios were measured and estimated, before and after tDCS application. sLASER MRS data were analysed using LCModel, and MEGA-PRESS using FID-A and Gannet. The electric fields were simulated in a finite element model of the head, based on individual MPRAGE images, using SimNIBS. Twelve linear mixed effects models were run, one for each E-field variable (mean and 95th percentile of magnitude, normal and tangential components), and separately for the M1 and temporal data. We found that in M1, E-field in the MRS voxel is related to the GABA drop, adding to the accumulating evidence that supports individualised dosing of tDCS. We also found an interaction with grey matter volume within the MRS voxel, emphasising the need to appropriately choose and evaluate any outcome measures which we expect to be related to E-field. While we did not find a similar association in the temporal region, given the challenges of modelling the E-field in this region and possible homeostatic metaplastic effects, such an association cannot be ruled out.