Interactions between head motion and coil sensitivity in accelerated fMRI

Parallel imaging is widely adopted to accelerate functional MRI (fMRI) data acquisition, through various strategies that involve multi-channel receiver coils. However, the non-uniform spatial sensitivity of multi-channel receiver coils may introduce unwanted artifacts when head motion occurs during the few-minute long fMRI scans. Although prospective correction provides a promising solution for alleviating the head motion artifacts in fMRI, the relative position of the fixed multi-channel receiver coils moves in the moving reference frame, potentially resulting in artifactual signal.

We used numerical simulations to investigate this effect on fMRI using two parallel imaging schemes: sensitivity encoding (SENSE) and generalized autocalibrating partially parallel acquisitions (GRAPPA) with acceleration factors 2 and 4, towards characterizing the regime over which parallel-imaging fMRI with prospective motion correction will benefit from updating coil sensitivities to reflect relative positional change between the head and the receiver coil. Moreover, six subjects were scanned with acceleration factors 2 and 4 while performing a simple finger-tapping task with and without overt head motion.

Updating coil sensitivities showed significant positive impact on standard deviation and activation maps in presence of overt head motion compared to that obtained with no overt head motion.

The parallel imaging fMRI with updated coil sensitivity maps were compared to that with the coil sensitivity maps acquired at the reference position.

Head motion in relation to a fixed multi-channel coil can adversely affect the quality of parallel imaging fMRI data; and updating coil sensitivity map can mitigate this effect.