The difficulty is that identical time-varying field gradients trigger nerve stimulation at different places in different patients. The tingling effect can also kick in at different magnetic-field-gradient switching rates. Theoretical studies have struggled to explain these real-life observations. Now, researchers in Nottingham, UK, have come closer to understanding when PNS occurs following a comprehensive numerical analysis of electric fields induced in the body during MRI (Phys. Med. Biol. 52 2337).

"There is quite a large literature – and a good literature – in this area, but it is sometimes quite hard to see exactly what happens as you move the body within the gradient coils, and what the differences are between the different coil types," said Richard Bowtell, professor of physics at the Sir Peter Mansfield MR Centre, University of Nottingham, who designed the study. "We have hopefully begun a more systematic investigation of this phenomenon."

The team used the commercial software package MAFIA and the 3D HUGO body model to calculate the spatial distribution of electric fields induced in the human body during a typical MRI scan. The magnetic field was varied at 1 kHz to produce a 100 T/m.s rate of change of gradient at the centre of typical x-, y-, and z-gradient coils. Numerical calculations were made with the coils centred on the subject's head, hips and heart.

A smoothly varying magnetic field produced highly heterogeneous maps of induced electric field and current density. Statistical analysis revealed that the induced electric field was largest in areas of high resistivity, and that current density was greatest in areas of high conductivity. The analysis also confirmed that the largest electric fields were in outer regions of the body, where nerve stimulation is unlikely to leave any lasting damage.

"It would be tremendously bad news for MRI if we were to stimulate cardiac nerves and muscle, but this study tells us that the electric fields are extremely low in the heart," said co-author Martin Bencsik, senior lecturer in physics at Nottingham Trent University.

The data showed no clear link between local values of switched magnetic field and induced electric field or current density. This finding quashes the widely held belief that local magnetic-field measurements could be used to predict the site of PNS hot-spots.

"Although the local magnetic field doesn't tell you about the location of nerve stimulation, you can still gather useful information from these simulations about where the electric field is largest for a given arrangement of subject and where they are positioned within the gradient coil," Bencsik said.

The maps also support suggestions that the threshold level of gradient switching at which PNS starts could be raised by shifting the maximum value of applied magnetic field along the body axis. This strategy, which would mean better images without patient discomfort, could be implemented on commercially available MRI scanners by adding new hardware.

Future simulations should examine electric-field patterns in subjects positioned close to – but not inside – gradient coils, Bowtell said. This could provide information on the exposure to induced currents faced by MR staff. A similar modelling approach applied to discrete portions of the body could also yield electromagnetic-field maps with enhanced spatial resolution.

Bencsik added: "PNS is a nuisance in MRI, but there are other clinical applications where it is desirable to stimulate nerves with an applied magnetic field. This again is an area of physiology that is poorly understood, and could be advanced by high-resolution electromagnetic maps of the human body."