Practical scan-length considerations for mapping upper limb movements to the somatosensory/motor cortex at 7T

Abstract

The relationship between motor cortex (M1) and upper limb movements has been investigated extensively using functional MRI (fMRI). While most research has focused on applications, very few studies have focused on practical aspects related to developing the fMRI protocol. Thus, the effect of scan length on M1 activations during various upper limb movements remains unclear. Scan length constraints are important for conducting motor experiments within a 60- or 90-min scan session. We targeted this gap by studying 7T fMRI activations in a male participant while performing eight different upper limb movements (of the fingers, wrist, and elbow) across 16 task runs (8 with the left arm, 8 with the right arm, 88 mins total fMRI duration). Standard activation analyses were performed (Z>3.1, p<0.01, cluster thresholded) independently for 14 different cases (2 runs through 8 runs, left and right arm) and compared. We found diminishing returns with higher number of runs (activations gradually plateaued with runs). We observed two clusters of movements, one with generally higher activation (more activated voxels and higher Z-stats) and the other with lower activation. To achieve similar statistical power, movements with lower activation required longer scanning (more runs). Based on these observations, we propose a one size does not fit all practical protocol within a 60-, 75-, or 90-min scan session, wherein different number of runs are assigned for different movements. Our study could benefit researchers who are designing upper limb fMRI experiments.