How Much Does Motion Matter? Evaluating the Motion Robustness of pTx Pulses at 7 T

To investigate the impact of head motion on the performance of subject-specific dynamic parallel transmit (pTx) pulses and the resulting image quality of structural neuroimaging sequences at 7 T.

B₀ and multichannel B₁ ⁺ maps were acquired in 11 healthy subjects across 15 head positions. For each head position, small-tip-angle k_T-spoke pulses and scalable nonparametrized pulses were optimized with different field maps following two experimental conditions: (1) No recalibration (NR): maps acquired at the first head position; (2) Full recalibration (FR): maps acquired at the new head position. Flip angle (FA) maps were then simulated from the optimized pulses. Excitation homogeneity was assessed using the mean relative absolute error (MRAE) compared to the target FA. In five subjects, intensity profiles and gray-to-white matter intensity ratio were also compared in MPRAGE and SPACE images acquired at three different positions-with the same pTx pulse optimized on the field maps acquired at the first position.

The FR condition yielded significantly lower median MRAE than the NR condition (p < 10^-25) across all subjects for both the k_T-spoke (3.55% ± 0.33% vs. 5.22% ± 1.27%) and nonparametrized pulses (6.08% ± 0.93% vs. 7.08% ± 1.47%). A strong correlation (coefficients between 0.88 and 0.96 across the whole brain) was found between intensities in the structural images acquired at the same position as for the pulse optimization (similar to the FR condition) and at another position (the NR condition).

Head motion has a negative impact on pTx pulse homogeneity, albeit with a relatively limited effect size and minimal impact on image quality.