Potential source of MRI signal change during transcranial direct current stimulation
Guoxiang Liu1,2, Takashi Ueguchi1,2, Ikuhiro Kida1,2, Ken-ichi Okada1,2, and Yasushi Kobayashi1,2

1National Institute of Information and Communications Technology, Suita-shi,Osaka, Japan, 2Osaka University, Suita-shi, Osaka, Japan


In this work, we implemented tDCS experiments on a monkey brain and a phantom at a 7T human MRI scanner to investigate the possibility to measure current flow during tDCS. Our results showed that imaging distortions caused by current in lead wires but not in brain is a possible source of BOLD-like MRI signal changes.


In recent years, transcranial direct current stimulation (tDCS) is often used in functional magnetic resonance imaging (fMRI) studies as a noninvasive brain stimulation technique, because tDCS has been shown to alter cortical excitability and activity via weak direct currents [1]. However, Antal et al. [2] have reported that there is potential confounding interference from tDCS-induced current flow in fMRI studies. They also mentioned that such kind of signal change, in turn, attracts the possibility to measure current flow in the brain during stimulation. In this work, we implemented tDCS experiments on a monkey brain and a phantom on a 7T human MRI scanner using a similar paradigm in ref. [2] to investigate the possibility of current flow measurement in the brain during tDCS. Our results showed that imaging distortion caused by current flow in lead wires but not in the brain is a possible source of BOLD (blood oxygenation level-dependent)-like MRI signal changes.


We performed tDCS experiments with a Japanese monkey (Macacca fusucata, female, 7kg) and a phantom on a 7T whole-body scanner (Siemens Healthcare, Erlangen, Germany) using a 32-channel phased array head coil (Nova Medical, MA, USA). In order to suppress head motion, the monkey was anesthetized using ketamine (10mg/kg) prior to the image acquisition. The monkey was placed in supine position with its head and neck fixed to the coil using a thermoplastic immobilizer. Direct current with 10 mA intensity was generated by an electrical simulator (Nihon Koden, Tokyo, Japan) outside the magnet room. Two pair of rounded electrodes (1 cm in diameter) were connected to the simulator via cable running through a radio frequency filter tube in the cabin wall. Each pair of electrodes were placed over the left/right SM1 and over their contralateral orbits. The lead wires of one pair of electrodes were placed along the longitudinal direction (z direction in x-y-z coordinate system), and the other pair was placed on the surface of the monkey’s head in x-y plane. Gradient-echo echo-planar imaging (EPI) sequence was used to obtain oblique-axial fMRI data at nominal spatial resolution of 1.5×1.5×1.5 mm3. Acquisition parameters were as follows: TR = 1000 ms, TE = 20 ms, flip angle = 41°, GRAPPA = 2. Similar tDCS validation experiment was performed using a brain tissue-mimicking phantom made of carrageenan gel. The stimulation paradigm was implemented as a block design with stimulation ON and OFF for blocks of 10 s each, and repeated 20 times. Stimulation-induced fluctuation in local magnetic field was detected as phase oscillation in gradient echo images (TR/TE = 300/20 ms) acquired at the same slice position. For comparison with ref. [2], “BrainVoyagerQX” (Version, Brain Innovation, Maastricht, The Netherlands) was used to analyze EPI time series without any preprocessing. Generalized linear model (GLM) analysis was applied for each experiment with stimulation ON/OFF as a binary regression variable.

Results and Discussion

Under the stimulation through the lead wires placed in the x-y plane, which was also used in ref. [2], BOLD-like MRI signal change was detected at the superficial areas (Fig. 1, top row). Similar pseudo-activation areas with inverted signal polarity were found when reversing the direction of the current flow. In contrast, stimulation through the wires placed in the z direction presented almost no pseudo-activation areas (Fig. 1, bottom row). These results suggest that EPI is insensitive to the current flow in the brain caused by tDCS, but rather sensitive to the magnetic field change caused by the current in lead wires. The results from the phantom experiment support this explanation. In the gradient echo scans, the phase shift caused by current-induced local magnetic field change can be calculated by comparing the data with and without tDCS (Fig. 2, top row). This kind of phase shift was observed when stimulation was given through the x-y plane, but was not detected under stimulation through the z direction. There was another evidence that the polarity of stimulation-induced MRI signal change in EPI time series of the phantom was highly dependent on the “polarity” of phase encoding direction (i.e., “AP” and “PA”; Fig. 2, bottom row). In conclusion, MRI signal change observed during tDCS is more likely due to EPI distortion stemming from current flow in lead wires.


This study was supported in part by Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research, "KAKENHI" (No. 26282223 and No. 26350471).


[1] Antal, A., Polania, R., Schmidt-Samoa, C., Dechent, P., Paulus, W., 2011b. Transcranial direct current stimulation over the primary motor cortex during fMRI. Neuroimage 55, 590–596.

[2] Antal, A., Bikson, M., Datta, A., Lafon, B., Dechent, P., Parra, L. C., et al. (2014). Imaging artifacts induced by electrical stimulation during conventional fMRI of the brain. Neuroimage 85(Pt. 3), 1040–1047.


MRI signal change during tDCS on monkey brain. Top row: lead wires in x-y plane. Bottom row: lead wires in z direction.

Phase change during different current.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)