Transcranial Direct Current Stimulation is a neuromodulation technique that has shown to enhance cognition, while simultaneously gaining importance as an augmentative therapy ^(1,2). However, the underlying mechanism of action of tDCS remains elusive, which makes numerous positive benefits difficult to interpret. Therefore mapping the current density in the brain will provide significant information about the path of electric current and the brain regions stimulated by the current, which will in turn aid in understanding the technique and its consequences. In this study, we acquired the first human brain MREIT data and successfully quantified the current-induced magnetic field which allows the calculation of the actual current density inside the head.

Imaging: MRI of a healthy participant was performed using a 32 channel head coil in a 3T Philips magnet at the McKnight Brain Institute, University of Florida. A high resolution T1 weighted 3D acquisition with a matrix size of 240x240x160,1 mm isotropic resolution and high angular resolution diffusion imaging (HARDI) data with b values of 100 (6 directions) and 1000 (64 directions) at a matrix size of 112x112x70 with 2 mm isotropic resolution, were collected. The low b-value acquisition was repeated with opposite phase encoding to correct for susceptibility distortions ⁽³⁾. MREIT scans were acquired with a spoiled multiple-gradient-echo pulse sequence (see Figure 1) in a single slice acquisition with 10 echoes (TE1 = 6 ms, ESP = 3 ms). 10 slices were acquired with a thickness of 5 mm covering a total of 50 mm with no gap between the slices. A flip angle of 35° was used in conjunction with a TR of 50 ms and 4 averages were collected. The scan was repeated 3 times to facilitate good SNR by averaging. The total scan time for 2 current injections in one direction was approximately 20 minutes. Raw data was exported and processed offline with software developed in house to generate the magnetic field maps produced by the current injection.

Stimulation: Positive and negative current pulses with amplitude of 1.5 mA at a duty cycle of 63% were injected into the participant using two sets of electrode pairs placed at F3 & Right Supraorbital and F4 – Left Supraorbital. The stimulation was performed with a constant current source (NeuroConn DCMC-MR) modified to generate short current pulses in synchronization with the RF pulses from the MRI scanner (see Figure 2).

The high resolution T1 acquisition, its corresponding diffusion imaging slice and the MREIT scan for the same slice are shown in Figure 3. Figure 4 shows the Bz map computed from the MREIT scans pertaining to same region and a line plot showing the actual field values within the brain. Since the first echo in the acquisition contains high SNR while the last has greatest sensitivity to the current induced magnetic field, the individual echoes from the multi echo data can be combined as described in the paper by Atul et al ⁽⁴⁾ to produce an optimal B_z map. The optimal field map so computed from the first five echoes is shown in Figure5.MREIT data with positive and negative current injection was acquired with multiple averages at 1.5 mA (2 mA has been shown to be safe to use in tDCS literature ⁽⁵⁾ which could improve the quality of the Bz maps). By incorporating a no-current injection acquisition, we could compute a T2* map which we can use to generate a more accurate optimal Bz map as shown in ⁽⁶⁾. Also acquiring averages as repeated single-slice image volumes, as opposed to averaging over slices, would provide robustness to motion and enable a safer approach to data collection. Generation of a computational model for calculation of injected current density inside the head is in progress.Successful imaging and determination of current induced magnetic field inside the human head was performed. Repeated measurements in more volunteers will be carried out to ensure reproducible results. The magnetic field information will be used to compute the current density values and compared with results from computational modeling.