Neuromodulation techniques enable non-invasive perturbations of brain physiology and function that may have therapeutic effects in brain disorders as well as other benefits for enhancing behavioral performance. Transcranial Direct Current Stimulation (tDCS) is one of such neuromodulation techniques that applies a small current (1-2mA) using scalp electrodes. Though tDCS has been shown to improve cognition as well as clinical symptoms [1], the neurophysiological effects of tDCS are still unclear and form an active area of research. In this study, we aimed at evaluating the neurophysiological effects of tDCS through Cerebral Blood Flow (CBF) measurement using arterial spin labeling (ASL).

Concurrent tDCS MRI resting state data were acquired from six subjects. For tDCS, 2 gel wetted sponge electrodes (5x7cm) were placed over the motor cortices (C3/4 on 10-20 system). The stimulation paradigm consisted of applying 0, 0.5, 1.0 and 1.5 mA currents in a pseudo-random block design with each block lasting 2.5 minutes and a ramp time of 30 seconds (Fig1). Data were recorded with a pCASL sequence (FOV=220mm, 24 axial slices, 5mm slice thickness, matrix size = 64×64, TE = 11ms, TR=4s, rate-2 GRAPPA, 7/8 partial k-space, 90 label/control pairs with a scan time of 12min). The tagging plane was positioned 90mm inferior to the center of the imaging slab with a labeling duration of 1500ms and PLD of 1000ms on a Siemens 3T TIM Trio scanner using a 12 channel coil. The scan was performed three times with different orders of current intensities, resulting in a combined 36 min resting state data. A high-resolution structural MPRAGE scan was also acquired.

Preprocessing of ASL data included realignment to the first volume separately for label and control images, quantification of CBF timeseries based on a single-compartment model, coregisteration to the structural scan and subsequent normalization to the MNI152 template.

The CBF time series from three scans were concatenated and a voxel-wise GLM analysis was performed using a predictor for each applied current intensity. At the group level, current effects were tested against 0mA condition by using t-tests to identify consistent across-subject effects. For ROIs showing consistent effects, CBF values were extracted and repeated-measures ANOVA was used to confirm effects across current intensities.

We found significant (p<0.05, cluster>50) increases as well as decreases in CBF for all 3 current intensities as compared to the 0mA baseline condition (Fig2). Conjunction analysis to identify areas with consistent significant effects across conditions only found areas displaying decreases in CBF. Accordingly, ROIs for positive effects were defined by the group-effects of 1.5mA (Fig2), while for negative effects the conjunction map across all current intensities was used (Fig3). Repeated-measures ANOVA revealed significant effects in almost all analyzed ROIs (Table1). Notably, we found a significant positive effect beneath the anode in the motor cortex (ROI 15). Furthermore, decreased CBF was observed in brain areas overlapping the default mode network.In this study, we showed significant CBF changes with applied tDCS currents. CBF increases were observed in a wide-spread network including the motor cortex under the anode. However, due to subthreshold effects in the 0.5 mA condition, no consistent across-condition regions could be detected. Nevertheless, ROIs constructed from significant regions in the strongest current condition indicated a dose dependent CBF increase (Fig4A), especially under the anode. This is consistent with previously reported work [2] as well as the hypothesis that anodal tDCS increases cortical excitability [3]. We also observed decreases in CBF in regions overlapping the default mode network (DMN) (Fig4B). Interactions between motor activity and suppression of the DMN have previously been observed in ASL [4] and causal interactions have been demonstrated between cortical networks and the DMN using neuromodulation and fMRI BOLD [5].In summary, our preliminary results suggest that CBF can be used as a sensitive marker for acute changes in neuronal activity associated with tDCS. Moreover, in contrast to BOLD, ASL provides absolute quantitative CBF that can be used to evaluate the dose-dependency of tDCS applied current.