Now you see it, now you don't: optimal parameters for interslice stimulation in concurrent TMS-fMRI

Abstract

The powerful combination of transcranial magnetic stimulation (TMS) concurrent with functional magnetic resonance imaging (fMRI) provides rare insights into the causal relationships between brain activity and behaviour. Despite a recent resurgence in popularity, TMS-fMRI remains technically challenging. Here we examined the feasibility of applying TMS during short gaps between fMRI slices to avoid incurring artefacts in the fMRI data. We quantified signal dropout and changes in temporal signal-to-noise ratio (tSNR) for TMS pulses presented at timepoints from 100ms before to 100ms after slice onset. Up to 3 pulses were delivered per volume using MagVentures MR-compatible TMS coil. We used a spherical phantom, two 7-channel TMS-dedicated surface coils, and a multiband (MB) sequence (factor=2) with interslice gaps of 100ms and 40ms, on a Siemens 3T Prisma-fit scanner. For comparison we repeated a subset of parameters with a more standard single-channel TxRx (birdcage) coil, and with a human participant and surface coil set up. We found that, even at 100% stimulator output, pulses applied at least - 40ms/+50ms from the onset of slice readout avoid incurring artifacts. This was the case for all three setups. Thus, an interslice protocol can be achieved with a frequency of up to ~10 Hz, using a standard EPI sequence (slice acquisition time: 62.5ms, interslice gap: 40ms). Faster stimulation frequencies would require shorter slice acquisition times, for example using in-plane acceleration. Interslice TMS-fMRI protocols provide a promising avenue for retaining flexible timing of stimulus delivery without incurring TMS artifacts.