Noise analysis in fast magnetic resonance electrical impedance tomography (MREIT) based on spoiled multi gradient echo (SPMGE) pulse sequence

摘要

Magnetic resonance electrical impedance tomography (MREIT) is a promising non-invasive method to visualize a static cross-sectional conductivity and/or current density image by injecting low frequency currents. MREIT measures one component of the magnetic flux density caused by the injected current using a magnetic resonance (MR) scanner. For practical in vivo implementations of MREIT, especially for soft biological tissues where the MR signal rapidly decays, it is crucial to develop a technique for optimizing the magnetic flux density signal by the injected current while maintaining spatial-resolution and contrast. We design an MREIT pulse sequence by applying a spoiled multi-gradient-echo pulse sequence (SPMGE) to the injected current nonlinear encoding (ICNE), which fully injects the current at the end of the read-out gradient. The applied ICNE-SPMGE pulse sequence maximizes the duration of injected current almost up to a repetition time by measuring multiple magnetic flux density data. We analyze the noise level of measured magnetic flux density with respect to the pulse width of injection current and T2 * relaxation time. In due consideration of the ICNE-SPMGE pulse sequence, using a reference information of T2 * values in a local region of interest by a short pre-scan data, we predict the noise level of magnetic flux density to be measured for arbitrary repetition time TR. Results from phantom experiment demonstrate that the proposed method can predict the noise level of magnetic flux density for an appropriate TR=40ms using a reference scan for TR=75ms. The predicted noise level was compared with the noise level of directly measured magnetic flux density data.