Miniature array postdetection-encoded MRI B Doron Kwiat 4 , Boris Sorokopud, Yoav Eiges, Elad Sity 2000 Computer College, Medical Imaging Department, 43000 Ra’anana, Israel Received 18 February 2006; accepted 4 April 2006 Abstract A method for performing nuclear magnetic resonance (NMR) measurements simultaneously from more than a single radiofrequency (RF) coil is presented. The method employs the detection of magnetic resonance signals in an array of detectors, where each detector is responsible for detecting a unique frequency bandwidth or a magnetic resonance signal from a unique location in a region in a primary, substantially homogeneous, static magnetic field. The detectors may be separated logically into groups, whereby all the detectors in a given group are essentially RF-decoupled from each other to substantially eliminate cross-talk by switching circuits or by being placed from each other sufficiently remotely. Sampling of detected signals from detectors in this array is done simultaneously over groups of noninteracting detectors. The detected signals from all detectors in a given group are simultaneously transmitted to a single preamplifier, thus increasing significantly the signal-to-noise ratio (SNR) in that preamplifier. Prior to transmitting each detected NMR signal of each detector to the preamplifier, each detected signal is separately and uniquely encoded electronically. This provides a method whereby the signal of each detector is uniquely encoded. Accumulating all these encoded signals, which were simultaneously received in a number of RF detectors into a single amplifier, results in the total signal having a high SNR ratio. This total amplified signal is later decoded into each detector’s original signal by a decoding circuitry. Conventional magnetic resonance imaging (MRI) techniques may be thereafter applied to obtain an image. Or else, conventional NMR techniques may be thereafter applied to obtain an improved SNR from a sample, using a single preamplifier with a multitude of detectors. Applying this method to a large number of miniature and closely packed RF detectors placed in an array-like configuration results in an MRI technique with a very fast acquisition time, an increased SNR and a high spatial resolution equivalent to the number of detectors per unit of length. Deblurring and decoupling algorithms allow for images from layers as deep as 6 mm to be acquired. D 2006 Elsevier Inc. All rights reserved. Keywords: MRI; Array; Frequency encoding; SNR 1. Introduction Surface magnetic resonance imaging (MRI) is the method used to produce images of a curved layer (body tissue or another substance) and is the method by which image is detected by a set of receiver coils that are proximal to the desired area of detection. In MRI, images are created by processing signals gathered from straight slices, which are predetermined by gradients. These signals are added to a homogeneous magnetic field whose field lines are all orthogonal to a fixed plane (xy plane). In surface MRI, magnetic field lines are in parallel with the layer of tissue to be scanned. In conventional MRI, a collection of frequency-encoded and phase-encoded signals from a multitude of voxel elements, which are produced in one straight slice with a predetermined gradient, are collected in a single coil. In surface MRI, all voxels in a curved thin layer transmit a single frequency. The signal from each voxel is detected in its neighboring coil, and each coil sends its detected signal for further processing. Due to the fact that the signal from each single voxel is very weak, all collected signals are summed to create a collective signal that is easily detectable. Before this summing can be done, it is necessary to encode each signal uniquely to allow it to be identified later. This is done by giving a frequency shift to each detected signal prior to summation. The advantage of the surface MRI method is that it produces an image of a curved layer immediately, which cannot be done using conventional three-dimensional methods. If a three-dimensional MRI is used to provide a reconstructed image of a curved layer, 0730-725X/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.mri.2006.04.006 B This work has been partially sponsored by the TNUFA Foundation, The Chief Scientist, Ministry of Industry, Government of Israel (fund no. 34679/2004). 4 Corresponding author. Tel.: +972 9 7421276; fax: +972 9 7602272. E-mail address: doron@2000computerschool.co.il (D. Kwiat). Magnetic Resonance Imaging 24 (2006) 963 – 975