Inter-Channel Correlation-Based EMI Noise Removal (ICER) for Shielding-Free Low-Field MRI.

Abstract

For low-field magnetic resonance imaging (MRI) in unshielded environment, existing methods have been proposed to eliminate electromagnetic interference (EMI) noise in each single radio-frequency (RF) receive coil. In the present study, we propose to use the EMI information from multiple MRI receive coils collectively in EMI denoising. The proposed method leverages the information of inter-channel correlation, including EMI detectors and RF receive coils to remove EMI noise. Calibration signals from both EMI detectors and RF receivers are concatenated to determine a de-correlation matrix, which is then used to denoise MRI signals. Saline phantom and in vivo experiments demonstrated the efficacy of the proposed method in EMI elimination, showing that the proposed method outperformed advanced EMI elimination methods with up to 16.34% improvement in EMI noise removal percentage (NRP) and 1.58dB improvement in signal-to-noise ratio (SNR), along with reduced computational time. The proposed method effectively removes EMI noise and shows improved performance by using information from all receive coils. This method allows for the design of multi-receive coils for low-field MRI such as phased-arrays, which have the potential to enhance the performance in noise removal and improve the SNR in MRI signal acquisition.