Reassessment of peripheral nerve stimulation thresholds for the Impulse model-optimized asymmetric head gradient coil

Peripheral nerve stimulation (PNS) remains a physiologic limitation to boosting spatiotemporal resolution with more powerful gradients. We investigate discrepancies in previous measurements and model predictions from PNS experienced by volunteers scanned with the investigational "Impulse" gradient coil on the NexGen 7T scanner.

Twenty-nine volunteers (18 males, mean ± standard deviation age 52.2 ± 17.1 years) underwent PNS characterizations in the scanner. The process was repeated after the subject positions were moved by 2 and 4 cm toward the feet, away from isocenter. These new data were compared with prior experimental data acquired at the factory (32 volunteers, 16 males, mean ± standard deviation age 58.3 ± 13.5 years) and to modeling results initially used to guide the gradient winding pattern.

The PNS threshold for the x-axis (left-right) was significantly below the threshold level predicted by the model used to optimize the wiring pattern and thresholds measured in the factory, whereas there was closer agreement for the y-axis (anterior-posterior) and z-axis (superior-inferior). The x-axis threshold increased as the subject was moved in the Z-direction toward the foot end of the magnet, at the expense of gradient nonlinearity distortions. Sensitivity of the threshold for the x-axis was measured as 20 mT/m per centimeter Z-offset.

The PNS threshold of the x-axis measured in the scanner was much lower than predicted by the optimization model and as measured at the factory. Our measurements verified that PNS thresholds of asymmetric head gradient coils were sensitive to head position, subject variability, and age. The discrepancy of the PNS prediction model remains to be elucidated.