Simultaneous Transcranial Magnetic Stimulation and Functional Magnetic Resonance Imaging: Aspects of Technical Implementation

Abstract

The simultaneous transcranial magnetic stimulation (TMS) and functional magnetic resonance imaging (fMRI) offers a unique opportunity to non-invasively stimulate brain circuits while simultaneously monitoring changes in brain activity. However, to take advantage of this multimodal technique, some technical issues need to be considered/addressed. In this work, we evaluated technical issues associated with the setup and utilization of this multimodal tool, such as the use of a large single-channel radio frequency (rf) coil, and the artifacts induced by TMS when interleaved with the echo-planar imaging (EPI) sequence. We demonstrated that good image quality can be achieved with this rf coil and that the adoption of axial imaging orientation in conjunction with a safe interval of 100 ms, between the TMS pulse and imaging acquisition, is a suitable combination to eliminate potential image artifacts when using the combined TMS-fMRI technique in 3-T MRI scanners.

Keywords: echo-planar imaging; functional magnetic resonance imaging; image artifact; radio frequency coil; temporal signal-to-noise ratio; transcranial magnetic stimulation.

FIGURE 1

(A) A picture of the Tx/Rx and 20-channel coils at the top-left corner. MRI signal intensity (mean), temporal standard deviation (tSD), and temporal signal-to-noise ratio (tSNR) images are shown for echo-planar imaging (EPI) images of the bottle phantom acquired with both rf coils (Tx/Rx and 20-channel); tSNR differences between 20-channel acquisitions (with and without parallel imaging) and Tx/Rx acquisitions (Diff-tSNR) are also shown; NP, no parallel imaging. (B) Schematic design of the transcranial magnetic stimulation (TMS) coil positioned over the left side of the phantom, oblique to the xy-plane. Axial view of the mean EPI images is displayed for the axial, sagittal, and coronal data acquisition of the bottle phantom with and without the TMS coil. The difference images (without - with TMS coil) are also displayed. (C) Brain EPI images for the axial and oblique data acquisition acquired with the TMS coil, positioned over the left dorsolateral prefrontal cortex (DLPFC, MNI = -50,30,36, highlighted in red), in addition to the oblique acquisition without the TMS coil are shown, on coronal and axial views.

FIGURE 2

(A) Schematic illustration of interleaved transcranial magnetic stimulation (TMS) and functional MRI (fMRI) data acquisition. dt is the time interval between TMS pulse (in green) and imaging acquisition block; TR is the repetition time. (B–F) Region of interest (ROI) values of MRI signal (ratio between value for each time point and the value for the first one) in a sphere (6 mm radius, in red) located at the slice 2 (the first slice acquired in an even number of slices acquisition mode) for experiments 1–4 as follows: (B) echo-planar imaging (EPI) acquisition without TMS pulse (experiment 1); (C) EPI acquisition with TMS applied at a rate of 1 pulse per volume with dt varying between 300 and 50 ms at a step of 5 ms (experiment 2); (D) EPI acquisition with TMS applied at a rate of 1 pulse per volume with dt varying between 150 and 50 ms at a step of 2 ms (experiment 3); (E) EPI acquisition with TMS applied in a rate of 1 pulse per volume with dt varying between 110 and 10 ms at a step of 2 ms (experiment 4) showing one image as an example of those unaffected by TMS and another from the last time point strongly affected by TMS pulse; (F) the same as (E) but without the last time point (dt = 10 ms). (G) Signal-to-noise ratio (SNR) values in the ROI (sphere of 12 mm radius highlighted in red displayed between G,H) for each time point in the time series are plotted for each experiment shown in (B–E) [SNR values for NO TMS (B) are plotted at dt = 0]. The red arrows highlight those time points with SNR values lower than the normal range determined by the NO TMS run; (H) mean tSNR values in the same ROI as (G) for NO TMS (experiment 1), dt300 (experiment 5, single-pulse TMS followed by imaging with dt of 300 ms), dt100 (experiment 6, single-pulse TMS followed by imaging with dt of 100 ms), dt50 (experiment 7, single-pulse TMS followed by imaging with dt of 50 ms), dt100-HF [experiment 8, 5 pulses (10 Hz) TMS followed by imaging with dt of 100 ms; HF, high-frequency], and dt50-HF [experiment 9, 5 pulses (10 Hz) TMS followed by imaging with dt of 50 ms]. The bar graph displays the mean tSNR values in the ROI with the percent change as defined in Eq. 3 for each TMS experiment compared with the baseline (NO TMS), the error bar displays the SD of the tSNR values inside the ROI.