Improved Coil Combination for Homodyne-Corrected Phased Array Images

INTRODUCTION: Homodyne detection is a partial Fourier reconstruction method that is implemented on many commercial scanners [1]. It is commonly used to reduce the number of high-frequency phase encodes or to correct for off-centered, partial echoes to achieve shorter effective TEs. It uses low frequency k-space data to estimate the phase modulation of each image. Missing k-space lines are filled-in using Hermitian symmetry properties, after which the resulting image is demodulated [1]. Homodyne detection has many other interesting applications such as preserving negative polarity in Inversion Recovery sequences, and it has been used to separate water and fat into quadrature through specialized pulse sequences. It is also particularly useful for low-SNR images because it can be used to avoid magnitude bias [1]. Phased array coils are often used in combination with Homodyne detection for partial Fourier applications. Commercial scanners typically combine these images using the Root-Sum-of-Squares (RSS) method [2], though this practice negates many benefits of homodyne detection by throwing away the phase information. If the underlying image is real-valued (which is true for many applications) the relevant signal in each coil is aligned to the real channel after homodyne processing. It should be possible to combine the real and imaginary channels independently with appropriate weighting factors. This would preserve the complex nature of homodyne images enabling many of the above-mentioned techniques to be carried out in a multichannel experiment.