danshel
commented
9 years ago I am looking into whether Siemens allows us to do a retro reconstruction such that we can specify which phase encode line to drop. If so then we could drop a line near the center of k-space (so we really see the effect of doing so) and apply the TMS pulse only during the acquisition of that PE line. We then compare with no TMS (still dropping that one PE line). If their is no significant difference then we could write an EPI sequence that acquires that line twice and only keep the good one.
A goal is to use theta burst TMS with pulses at 50 Hz (20 ms inter-pulse delays). At present, with TMS pulses applied during crusher gradients either side of the fat saturation RF pulse, the per slice acquisition time limits the TMS pulse frequency to around 15 Hz. We need to find a way to apply TMS pulses every 20 ms or so, i.e. somewhere within the EPI readout echo train.
The biphasic TMS pulses may be sufficiently balanced that they can be applied in "dead periods" such as the crusher gradients, and have minimal residual effect (dephasing) of spins. But we need to test how well balanced these biphasic pulses really are. We might use a spin echo MR spectroscopy sequence, for instance. If we pulsed the TMS during one TE/2 only then we may be able to use the final signal intensity as a gauge of the TMS pulse balance, like a bipolar diffusion-weighting gradient.
Assuming proper balance with minimal residual effects, we also need to determine where and how in the EPI pulse sequence to apply biphasic TMS pulses. Daniel's initial suggestion is to simply corrupt one line of (readout) k-space then remove it in post processing.