Thirty neurologically and mentally healthy, right handed (validated by the Edinburgh Handedness Inventory (OIdfleld 1971) participants were recruited (mean age = 25.5, SD = 5.14; 15 male). Participants were pre-screened for TMS, tDCS and magnetic resonance imaging (MRI) exclusion criteria.
2.2 Experimental procedure
An overview of the experimental procedure and the durations is shown in Figure 1. In all participants, a high-resolution anatomical MRI scan (see TMS and MRI sections below) was measured and the individual resting motor threshold (rMT) was determined using a standardised protocol (Rossi et al. 2009; Rossini et al. 2015). Afterwards the first of three identical rsfMRI measurement was collected for each participant, lasting about 10 min. During the scan participants were shown a small black fixation spot in the middle of a grey background. They were instructed to fixate the dot at all times, to relax, don’t fall asleep, lie as still as possible, and to try not to think of anything in particular. For the following application of brain stimulation by use of tDCS and TMS the participant was brought outside the scanner room but was lying supine on the mobile scanner bed for the entire experiment. After registration with individual anatomical MRI data (see MRI section below for scanning parameters) tDCS was applied over the lDLPFC using neuronavigation (see section ‘Transcranial direct current stimulation’ for more details) for 10 min with either anodal (10 participants), cathodal (10 participants) polarity, or sham tDCS with either anodal or cathodal polarity (10 participants). Blinding for the tDCS protocol was assessed after all experimental procedures by asking the participants whether they knew they had received real or control stimulation and whether they had experienced any unpleasant sensations. After tDCS, participants underwent the second rsfMRI measurement, before they received iTBS lasting 190 secs, applied over the lDLPFC using neuronavigiation (see section ‘Transcranial magnetic stimulation’ for more details) at 90% of individual rMT. After a 7-min-break the participants were rolled into the scanner again, lying in an unchanged position on the scanner bed and the last 10-min-post TMS rsfMRI measurement was conducted. This 7-min interval was kept constant between all participants and was implemented to allow for arousal and discomfort ratings and MRI preparations. Immediately after each stimulation, arousal was assessed using a 9-level self-assessment manikin (SAM) scale with level 1 corresponds to 'very calm and relaxed” and level 9 corresponds to ‘very excited, stimulated, furious, excited, aroused’ (Bradley and Lang 1994). After TMS, an additional 5-level Likert scale was used to measure discomfort during iTBS stimulation (1 = none, 5 = strong). To test whether arousal or discomfort was significantly different between the tDCS groups or measurement points, non-parametric tests were conducted.
2.3 Transcranial direct current stimulation
TDCS was administered either as anodal (1mA), cathodal (- 1mA) or sham (0mA, either anodal or cathodal electrode placement) stimulation using a MRI-compatible stimulator (neuroConn GmbH, Ilmenau, Germany). The fade-in/out period was 10 s. The placement of the active electrode sized 5 by 7 cm was thereby determined by transforming the individual anatomical images into the MNI system using the neuronavigation system (LOCALITE Biomedical Visualization Systems GmbH, Sankt Augustin, Germany) and marking the MNI coordinates (x, y, z) = -50, 30, 36 with the neuronavigation pointer as stimulation target. These coordinates were suggested by Rusjan et al. (2010) for an optimal location of the lDLPFC by neuronavigation as compared to conventional distance-based localization methods. The tDCS reference electrode sized 10 by 10 cm was placed contralateral and supraorbital. The stimulation was applied for 10 min. As during rsfMRI, participants were instructed to relax, don’t fall asleep, lie as still as possible, and to try not to think of anything in particular, but they could close their eyes.
2.4 Transcranial magnetic stimulation
Both, single pulse TMS for determination of the individual rMT and the experimental iTBS was applied using a figure-of-eight coil (C-B60) connected to a MagPro X100 stimulator (MagVenture, Farum, Denmark) guided by neuronavigation. In preparation of rMT determination, the presumed hand area was identified visually through anatomical landmarks in the left motor cortex. Participants were placed in a comfortable chair or lying down on the MRI scanner bed outside the scanner room for registration with their individual anatomical MRI data. Three pre-gelled disposable surface electrodes were fitted to the participant’s right hand (first dorsal interosseous muscle, index finger, inner wrist) to derive MEPs which were monitored on the MEP Monitor (MagVenture, Farum, Denmark) connected to the MagPro X100 stimulator. Biphasic single pulses were applied over the presumed hand area starting at 30% of stimulator output but was increased until clear MEPs and hand muscle contraction could be observed. Intensity was then reduced stepwise to find the lowest intensity that induces supra-threshold (> 50 µV) MEPs above chance, i.e., we used the common rule that the rMT corresponds to the minimum stimulation intensity at which MEPs of at least 50 μV are elicited in at least 5 of 10 consecutive trials (50%) in the resting target muscle (Rossini et al. 2015).
The experimental excitatory iTBS (Huang et al. 2005) protocol consisted of 600 pulses spaced-out over 3 min and 20 seconds. It was comprised of 20 trains and 10 theta-bursts. Between each of the 2 sec-long trains (50 Hz) there was an 8 sec long pause. The lDLPFC stimulation site was determined the same as the tDCS target. Actual individual stimulation sites were recorded during the TMS procedure and used as subject specific seed regions. Participants received a stimulation at an intensity of 90% of their individual rMT. The mean rMT was 45.63 % (SD = 6.82) of the maximum stimulator output and the mean stimulation applied was 41.01 % (SD = 5.68) of the maximum stimulator output. The stimulation threshold of 90% of the individual resting motor threshold was chosen based on related experiments by our group that found a strong frontostriatal modulation effect at that threshold (Alkhasli et al. 2019).
Simultaneously with each TMS pulse, stimulation markers, including the information of exact position of the coil hotspot and its perpendicular projection onto the brain surface, were recorded by the neuro-navigation system. For each participant, we exported one of the first stimulation marker as volume of interest into the NIfTI file format for further image analyses.
2.5 Magnetic resonance imaging
MRI scans were measured on a Magnetom Prisma 3.0 T whole-body scanner (Siemens Medical Solutions, Erlangen, Germany). Anatomical data was acquired using a three-dimensional magnetization-prepared, rapid acquisition gradient-echo sequence (MP-RAGE) with the following parameter: 300 repetitions, TR = 2300 ms, TE = 2.98 ms, 9º flip angle, FOV = 256 mm, 176 sagittal slices, slice thickness = 1 mm and in-plane resolution = 1 🞩 1 🞩 1mm.
RsfMRI data were measured with a gradient echo planar imaging (EPI) sequence with the following parameters: TR = 2000 ms, TE = 28 ms, 77º flip angle, FOV = 192 mm, 34 axial slices (interleaved acquisition), 3 mm slice thickness, echo planar imaging volumes and in-plane resolution = 3 🞩 3 🞩 3mm. Both sequences lasted about 10 min.
2.6 Pre-processing of resting state
MRI-data was analysed using the Statistical Parametric Mapping software SPM12 (Welcome Department of Cognitive Neurosciences, London, UK) and CONN Functional Connectivity version (18.b, Whitfield-Gabrieli and Nieto-Castanon 2012 ) toolboxes running under Matlab R2012b (MathWorks Inc., Natick, MA, USA). Pre-processing of the rsfMRI data included: removal of first 5 volumes to discard saturation effects, slice time correction, realignment, segmentation, nuisance covariates regression with white matter and cerebrospinal fluid as regressors, head motion correction, head motion scrubbing as regressor, band pass filtering of the frequencies 0.01-0.08 Hz and linear detrending. The root-mean-square of the head motion translation parameters [displacement = square root (x2 + y2 + z2)] across all participants and sessions was 0.13 mm (Max = 0.25 mm, SD = 0.05 mm), with a mean subject-wise maximum of 0.97 mm (Max = 3.07 mm, SD = 0.83 mm).
2.7 Functional connectivity
To explore the whole brain stimulation effect, seed-to-voxel correlations were calculated. For each subject the activity of the individual stimulation site (lDLPFC, a sphere with a diameter of 1cm) was extracted as an unweighted mean BOLD signal change time series. Three-dimensional stimulation seed masks where obtained from each participant’s individual T1-anatomy and then co-registered with the corresponding functional data set. Functional connectivity was then calculated as a Fisher-Z-transformed correlation coefficient, between the stimulation site (lDLPFC, seed) signal and all individual voxel signals, separately. All correlation values are Fisher-Z-transformed. Alpha values inflation caused by multiple comparison was corrected on a cluster-size level.
To explore connectivity patterns specifically between the stimulation site and specific ROIs, ROI-to-ROI-analyses were calculated. Functional connectivity was additionally calculated using a three clusters extracted from the seed-to-voxel analysis, as well as a bilateral striatum masks (Harvard-Oxford atlas, consisting of caudate, putamen and nucleus accumbens). Striatal signal time series were thus extracted from MNI normalized functional data. In total there were four different seed masks: lDLPFC (stimulation site), three extracted clusters and the striatum.
Calculating a 3 🞩 3 mixed effect analysis of variance (ANOVA) of the functional connectivity between the stimulation site seed and each of the additional four ROIs (three clusters and striatum) resulted in four separate ANOVAs. For each ANOVA the within-subjects effect was the repeated measure (baseline vs. post-tDCS vs. post-TMS) and the between-subjects effect was the tDCS group (anodal vs. cathodal vs. sham tDCS).