Figure 1: The three constructed array coils with covers open: (A) symmetrically distributed equally-sized loops in an overlapped arrangement, (B) a gapped array with symmetrically distributed equally-sized loops, and (C) a non-uniformly loop density variation.

Figure 5: Coil comparison in a healthy volunteer. Color-coded maps of the SNR using each coil at mid-ventricle (A). Statistical comparison of the SNR for each coil at specific regions in the myocardium (B). Tractogram of the entire myocardium (left and right ventricle), color-coded by the myofiber helix angle (C).

(A) 3D rendering of the novel 128-ch brain array with a 24-ch parallel Transmit Array setup. The pTx array is organized in 3 rows of 8. (B) 3D rendering of the whole 128-channel brain receive coil setup at 7 Tesla in SolidWorks (Dassault Systems, France). The table spoon extension (blue) with the Rx and Tx coils is connected to an interface box (brown) that slides into the MRI bore (C) Side and back view of the 128-ch coil array without populated elements. The transmit (Tx) coils (1-ch birdcage or 24-ch pTx) slide over the receive (Rx) array.

(A) Simulated unaccelerated SNR maps of a 32-, 64-, and 128-channel array in a transverse, coronal and sagittal slice. For simulation in MARIE a uniform head model (average brain: epsilon=52, sigma=0.55 S/m), which closely matches the average brain tissue, was used. (B) Simulated accelerated SNR maps for R=5 and R=3x3 of the 32-, 64-, and 128-channel array. (C) SNR profiles through a central axial slice for the 128-channel (red line), 64-channel (green line), and 32-channel (blue line).

Figure 1: Receive array former (helmet) shown from several angles: (B) Front, (C) Side, (D) Back, and (E) Top. Front end electronics (preamps and PIN diode bias tees) are shown mounted to the top plate. The top plate mounts to the helmet via a 3D printed central pillar and four 12 mm diameter fiberglass rods.

Figure 4: Off-coil (top-left) and on-coil (bottom-left) preamp PCBs are shown in the photo along with corresponding circuit schematics (right) for the loop feed circuit and preamp board.

META Head Coil: a) The coil on the 7T MR system including the ³¹P bore coil (embedded in bore) and the ²³Na clamp; b) The eight-channel dipole array as RF transceivers for ¹H and ¹⁹F; c) The fifteen-channel receiver array for ³¹P, ²³Na and ¹³C; d) The entire coil; e) The digital interface platform under the front cover of d.

a) and b): Scan result from a brain ³¹P CSI FID, TR=71.1ms, FA=13deg, voxel size =20*20*20mm³, Hamming-weighted NSA=70, duration=10min: a) a slice from the 3D dataset showing the spectral map at full brain coverage, the background image is overlaid manually for illustration; b) ³¹P spectrum of the marked voxel of Figure a; c) and d): Signals from a close-fit head birdcage^[2] and the META head coil, at the center of a phantom. The SNR comparison was assessed by scaling the data to obtain the same standard deviation in the spectral noise then taking the integral of the signal in spectral domain.

Figure 3 – Experimental SNR comparison between the 64-channel 10.5T receive array (top row) and the 32-channel 10.5T receive array (bottom row) at 7 axial slices.

Figure 1 – As-built photograph of the 64 channel 10.5T receive array for human brain imaging.

Figure 2. Simulated central combined 2D $$${B_1^+}$$$-field distribution in central axial, sagittal, and coronal planes within Duke and Ella voxel models computed for the optimal phase vectors (PV1 and PV2 for Design1, and PV3 and PV4 for Design2). Both cardiac arrays were fed with and all coil elements were fed with equal amplitudes. Values of RSD and mean $$$Tx_{eff}$$$ ($$$\mu{T}\sqrt{kW}$$$) were computed in the selected 2D ROI and written on each image.

Figure 5. (a) Ultra-high resolution T2* weighted images of the SA and LA heart views acquired with array Design2. (b) T2* weighted images of the LA view acquired with the spatial resolution 0.35×0.35×4mm³ at different GRAPPA acceleration factors to check parallel imaging capability of the array Design2.

(A,B) Anterior and posterior housings segment of the 16ch^Tx /64ch^Rx head-neck coil maintaining a constant offset of the transmit to the receiver loops. (C,D) populated coil array with cut-outs for eyes and mouth facilitate visual stimulation and free breathing, respectively. Each Tx loop consists of 7 -13 capacitors, a PIN diode circuitry for active tuning, and a symmetrical drive port. (D) Final enclosed coil plugged into the 7T scanner’s patient bed illustrating the goal of making a form factor matching what is clinically used at 1.5T and 3T.

Sagittal 7T GRE in vivo image acquired with the constructed 16chTx /64chRx head-neck coil to demonstrate the extended coil coverage (TR/TE/α:40ms/5ms/20, matrix: 344x344, resolution: 0.3 x 0.3 x 5mm, BW: 320Px/Hz).

Figure 2: A – Picture of the completed receive array. An input board on the helmet surface consists of the matching and active detuning circuit. A feed-board with preamplifier is on a holder above the helmet; B – Picture of the final setup; C – View from the service end highlighting the visual field.

Figure 1: A – EM simulation setup. Tuning, matching and decoupling was adjusted by loading the coil with a head & shoulder phantom; B – A view of the transmit array model is shown to visualise the arrangement. Channels 1 to 8 are on the top row and 9 to 16 are on the bottom row, arranged one below the other.

Figure 5: Maximum intensity projections of SAR in axial, coronal and sagittal views normalized for 1 μT excitation within the volume outlined in the transmit efficiency maps (Figure 4). The location of the maximum SAR for both is indicated at the right arm, which is the body part nearest to the current-carrying conductors.

Figure 2: Side-views of HFSS simulation models for (left) birdcage coil used as a standard for comparisons of MR transmission metrics and (right) metamaterial liner with human body model shown inside. Locations of transmission ports and dimensions are labelled.

FIGURE 4: ³¹P MRSI results obtained from the external receiver coil in two subjects - also here PME is more intense, particularly in the first subject.

FIGURE 3: ³¹P MRSI results obtained from the saddle and loop coils for the same subject. The observed SNR is highest with the loop coil, albeit that the signal may originate from the jaw muscle. For the saddle coil, in contrast to typically reported ³¹P spectra from muscle where PDE is higher than PME.

Figure 1 Coil outer dimensions. (a)Side view: Legacy 12-Rx Knee Coil on UIH 3T system; (b, c) Side and front view: New 2-Tx 24-RX Knee Coil developed on UIH 5T system.

Figure 5 Clinical scan for human knee.

Figure 3: Susceptibility weighted image (left) and zoomed in images (right) with 0.35*0.35*1.5 mm³ at 5T.

Figure 2: Magnetic resonance angiography images acquired by TOF sequence with 0.6*0.6*0.6 mm³ resolution. (a) and (b) were acquired at 5T and 3T respectively using the same coils as in Figure 1.

Figure1: Two different coils are simulated; (a) A 2ch birdcage coil structure; (b) An 8ch loop coil structure.

Figure2: RF simulation model of the 8ch loop coil used in CST-MWS loaded with an realistic human model (Gustav).

Figure 4: MP2RAGE images of the head (resolution=1.0×1.0×1.0mm, 188 sagittal slices, SENSE=1.7(AP) × 1.7(RL), FA=5°, TR=6.2 ms, TE=2.0 ms, BW=162 Hz/pixel, and TA=6min41sec)

Figure 2: S-parameter matrix of the 8ch array in simulation (a, b) and experiment (c)

Figure 3. |B₁⁺| field and SAR maps of the axial, coronal, and sagittal slices of the head depending on the coil configuration. Note that 3 μT and 8.0 W/kg was normalized to 0 dB for |B₁⁺| field and SAR maps, respectively. The total input power of each configuration was scaled such that the peak of local SAR on the whole head was 8.0 W/kg.

Figure 1. (a) Dimensions and components of the single MTL resonator. Simulated eight-channel MTL RF head coil of different configurations. Typical configuration and (b) adjusted configuration (1) with a circular shape, (c) configuration (2) with a elliptical shape, and (d) configuration (3) similar with the head shape.

Fig. 4 Simulation (a and b) and experimental (c and d) iSNR maps of the 8-channel HDC dipole antenna array of the “Normal” and “Flipped” setups.

Fig. 5. Profile of the simulated (Blue, Red) and measured (Green , Purple) iSNR for Flipped and Normal setups. The location of the profiles is indicated as black dotted lines in Fig. 4 .

Figure 5. Slices of the short-axis view stack of the pig heart acquired post-mortem. The right panels show increased image slices marked red on the left panel. The LGE in the post-infarction scar is observed in the septal and anterior wall (arrows). Sufficient $$${B_1^+}$$$ field penetration and homogeneity without destructive interferences are achieved in the whole heart in both anterior-posterior and basal-apical directions.

Figure 2. (a) Schematic of the mono-surface antisymmetric 16-element dipole antenna array with element dimensions, paring channels, and channel numbers with every two neighboring dipoles to form 8Tx-channels for the pTx system (8Tx/16Rx). (b) RF simulation model as simulated in CST-MWS of the new dipole antenna array loaded with a dedicated pig thorax phantom. (c)-(d) Prototypes of the dedicated pig thorax phantom and the antisymmetric dipole antenna array.

Fig.1. Simulation model showing the 19-channel hybrid array system a)Foot/Ankle phantom b)3D view of the phantom with RF coil system c) Front View of the system d) Side view of the system e) Back view of the system f) Top view of the system.

Fig.2. B1+ field distribution along the transverse plane of the phantom. The transverse plane propagates from lower dimensions to higher dimensions. B) B1+ field distribution along the sagittal plane of the phantom at the center.

The 28-channel Tic-Tac-Toe (TTT) RF transmit coil design. In a), the full model with 14 TTT panels shown with the Duke head model; in b), the coupled configuration of the TTT, with 4 channels per panel; in c), the uncoupled configuration, with 2 channels per panel; in d) the assembled TTT panel used in the experimental validation

RF shimming with the 28-channel uncoupled TTT configuration (14 panels, as shown in Figure 1a); in a), several axial slices from below the brain stem to the top of the head and mid-sagittal and coronal slices. The simulations show homogeneous whole-brain coverage. In b), the maximum intensity projection of the SAR in axial, coronal, and sagittal planes respectively; in c), summary of the statistics. These are high values of homogeneity and SAR efficiency at 7T

Figure 1. Different body coil architectures. (a) TEM Coil, (b) birdcage coil, and (c) 2D c-HPL coil. All coils have a shield length of 1 m and a shield diameter of 66 cm. Coils have the same length (L) of 33 cm in longitudinal direction and diameter (D) of 60 cm.

Figure 3. In silico SAR maps of volumetric body coils. Phantom results (top row) were used to calculate the SAR efficiencies and realistic body model results (bottom row) were used to verify the previous results. Color-bars are normalized to 0.1 W/kg.

Fig.1. Side view of the asymmetric ring and asymmetric gap birdcage coil. The front side ring has the larger diameter and narrower gap to the RF shield.

Fig.3. B₁⁺ amplitude line profile for the model (B) and (C) at the coronal plane and X=20 cm along the head-feet direction. Uniformity between Z=±10 cm greatly improved by taking the design of asymmetric ring width in the model (C).

Fig. 4. Transmit performance of the 8-channel dipole-fed rectangular dielectric resonator (RDR) array compared with the 16-channel loop-dipole coupled RDR array (circularly polarized mode). The 16-channel array yielded significantly higher B1+ (+35%) and SAR (+9%) efficiency in the center of the spherical phantom.

Fig. 2. Phantom measurements: MR images obtained for a single-element loop- and loop-dipole-coupled rectangular dielectric resonator antenna. Three different profiles were selected and signal intensity as a function of distance was shown below.

Figure 3: Distribution of B1+ fields over a spherical phantom, shown as flip angle (FA) per 500V input: a) simulation of standard TTT panel, b) simulation of microstrip TTT panel, c) experimental B1+ map of microstrip TTT panel. Dashed lines in (a, b) show central slices used in Figure 4.

Figure 1: a) standard Tic-Tac-Toe (TTT) panel, b) microstrip TTT panel, both 4.25in x 4.25in

Figure 2 Long and short-axis view of the heart acquired at different flip-angles with both arrays. A pronounced appearance of blood-flow induced artifacts is observed with the new array, probably due to lower B₁-gradients in the anterior-posterior direction.

Figure 5 Result from optimized B₁-vectors calculations. Panels (a) and (b) show the B₁-field on the "Duke" model in the transversal and sagittal slices respectively. A sufficient penetration depth with better coverage of the posterior heart (lung and heart contours are shown) is predicted by these simulations for future measurements.

Figure 1: Multi-channel RFPA for UHF MRI

Figure 2: Pre-distortion algorithm

FIGURE 1: Numerical setups: the voxel model placed inside the body birdcage coil with M-coil (a), H-coil (b), and MH-coil (c) placed around the breast.

FIGURE 3: Numerically calculated B₁⁺ (a–c) and SAR_av.10g (d–f) maps for the voxel model placed inside the birdcage coil without (a,d) and with the MH-coil. The root mean square |B₁⁺| value was calculated for the breast volume. (d-f) SAR_av.10g distributions are plotted through the local SAR maximum plane. The white circles indicate local SAR_av.10g maxima.

Figure 1. a) Photo of the prototype of the volumetric harvesting coil and b) its circuit diagram.

Figure 3. Influence of the pulse sequence type on the harvested power at the same Vadj =180 V, with default FA and load resistor 500 Ohm.

Fig. 3. Estimated B₁⁺ maps of Slice 28 (marked with red boxes in Fig. 2A, 2B) in CP mode (A) and B₁-shimmed mode (B). The averaged B₁⁺ fields of all voxels in Slice 28 with the HDC-MTL transceiver array (right panels) were 2.1 times and 1.4 times higher than those with the control array (left panels). 1D profiles of selected voxels (red lines in 3A and 3B) of Slice 28 in CP mode (C) and in B₁-shimmed mode (D) showed significant B1⁺ improvement and enhanced penetration depth after B₁ shimming.

Fig. 4. Estimated relative B₁^- maps of 36 imaging slices covering > 10 cm in the middle of the head-shaped phantom (A); and Slice 28 (marked with red boxes in Fig. 4A) B₁^- map (B) with the HDC-MTL transceiver array and the control array in the B₁-shimmed mode. The 1D profiles of selected voxels (red lines in 4B) showed significant improvement of the B₁^- field with the HDC-MTL transceiver array, especially in the periphery areas.

Figure 2. A selection of mode distributions of the proposed 4 × 8 c-HPL coil, where each mode occurs at a different frequency due to mutual coupling. Here, mode (1,4) occurs at the lowest frequency with a higher spatial variation among others. Mode (1,1) is tuned to 298 MHz and showed the best B1+ uniformity.

Figure 1. (a) A 16-leg high-pass birdcage coil, (b) a 2D cylindrical high-pass ladder (2D c-HPL) coil array (4 × 8). Both coils are loaded with a spherical phantom at isocenter and driven with an accepted RF power of 1W.

Figure 1 Different transmit setups. 32-ch receive array with the 16 elements chosen as transceivers shadowed (a); external 16-ch transmit array with high-permittivity helmet (b); 32-ch receive array with HPM with the 16 elements chosen as transceivers shadowed (c). Note that in (b) the 32-ch array between the external transmit array and the HPM helmet is not shown for clarity. Simulations were repeated also for the transmit setup in (b) after removing the HPM helmet.

Figure 2. B₁⁺ divided by the square root of absorbed power for the four setups for three orthogonal sections at the center of the head. The 16 transmit elements were combined in CP mode. The HPM had a negligible effect for the case of the external transmit array.

Figure 1. Complete breast/torso coil setup: (a) Front and (b) back mechanical assembly, (c-d) Coil assembly

Figure 3. SSFSE, Left: acceleration factor of 3, Right: acceleration factor of 4. Improved sharpness is observed with higher acceleration due to reduced T2 blurring resulting from a shortened echo train. Arrow indicates region of fibroglandular tissue.

Figure 1. (a) 3D design of the coil former. The upper part (top) is detachable from the lower part for size adjustment. The Top and base are 3D printed using rigid PLA filaments. The sidewalls are 3D-printed using semi-flexible filaments TPU so the sidewalls can be pulled inwards-outwards to shrink/extend the size of the coil. (b) Schematic circuit design of individual coil loop. Ct=15-40pF, Cm=33pF, Cd=24pF, Cc=8.2pF.

Figure 2. (a) The coil adjusted to 180mmx 220mm and loaded with a sphere phantom in a diameter of 180mm; (b) The coil adjusted to 220mm x 260mm and loaded with the same phantom. (c) A axial-plane image acquired with coil setting (a); (c) A axial-plane image acquired with coil setting (b); (e) SNR distributions along the horizontal center line of image (c) (red) and image (d) (blue).

Fig.3. 0.6 mm isotropic modulated flip angle technique in refocused imaging with extended echo train (MATRIX) images were acquired using the commercial head coil (32ch; top row) and the quadrature birdcage/47Rx coil array (47ch; bottom row) at R = 7-fold acceleration. The acceleration direction was on phase encoding and slice phase encoding. H = head; P = posterior; R = right; A = anterior.

Fig.2. Inverse g-factor maps. To display coil acceleration ability of each direction, each orientation was accelerated on different directions. Transverse orientation: Left-to-Right acceleration direction. Sagittal orientation: Anterior-to-Posterior acceleration direction. Coronal orientation: Head-to-Feet acceleration direction. Regions of interest (ROIs) is marked as black dashed line. The mean g-factor value of each ROI is written beside the related image. L: left, R: right, A: anterior, P: posterior, H: head, F: feet.

Figure 5: A, (Finger tapping > Rest) BOLD fMRI significant clusters (z>3.1) for 2 sample participants along axial slices at the height of the cerebellum, crossing the hand region in cerebellar lobule V (left) and cerebellar lobule VIII (right). B, Group Z-stat distribution for each coil. C, Barplots of active voxels at group level in relation to the distance from the skull.

Figure 1: A, view of the cerebellar coil. Red, transmit elements. Blue, receive elements. B, The transmit array (housing partially removed). C, The receive array (open covers). D, Distribution of voxels in the cerebellum with respect to distance from the skull (median distance=2.58cm (dotted line)). E-G, 3D-EPI slices. E, Head coil without B₁ shimming. F, Head coil after B1 shimming. G, Cerebellar coil. Note the reduced deeper sensitivity deeper due to the small total surface area.

Figure 1. The pictures of the coils and their schematics. a) U-shaped coil , b) 3-turn solenoid coil, c) 8-turn solenoid coil. d,e) Schematics of the U-shaped coil and a solenoid coil. e) is citation from reference[4]. C_m and C_t enable fine adjustment of the coil impedance matching and resonance frequency tuning

Figure 4. The images of a 40 μm-thick rat brain slice. a) U-shaped coil, b) 3-turn solenoid coil, c) 8-turn solenoid coil. The signal value was measured in the red square region, and the noise value was measured in the yellow rectangle. The color scale was adjusted for each image.

Modelled and simulated circular overlapped coil arrays for highly SMS accelerated cardiac images.

Inverse g-Factor maps of a representative sagittal slice through the heart. g-factors were derived from axial SMS accelerations with multiband factors 4, 6, 8, and additionally combined with in-plane accelerations of R=2 and R=3. The heart region was covered by 60 slices of 2 mm each. 15, 10, and 8 collapsed slices were needed to reconstruct the full heart for MB=4, MB=6, and MB=8, respectively.

Fig.5 (animation): 2D concurrent multi-slice accelerated dynamic imaging of swallowing an ~10 ml bolus of pineapple juice. Non-Cartesian spiral under-sampling at ~27 fold acceleration level was combined with a sparse SENSE reconstruction scheme. The transport of the bolus is robustly captured in the three sagittal slices with adequate spatial resolution (2.4 mm²), and temporal resolution (17.1 ms).

Fig.2: (a) Individual coil images from the 16 channel elements and (b) the R=1 SENSE coil combined image using the proposed airway coil. The coil offers high sensitivity in all upper-airway regions of interest (eg. lips, tongue, hard palate, soft palate, epiglottis, glottis).

(A) Geometry of the three used receive coils with measured inductance and resistance using a vector network analyser at 32 MHz. The multi-loop coil is printed on 1.6 mm FR4 (1 oz copper), while loop coils are made from $$$(\oslash2 mm)$$$ copper tubing. (B) The coils were hosted in a Styrofoam box and laid parallel to the box surface. Separation $$$g = 2$$$ cm. The phantom was positioned adjacent to a $$$(\oslash5 cm)$$$ transmit coil, under a plastic arc.

(A) Normalized simulated SNR for thermal and (B) hyperpolarized MRI as a function of field strength and coil temperature, (C) Ratio between equivalent sample and coil resistance. (D) Potential SNR gain relative to 300 K receiver coil at 3T for hyperpolarized MRI. All simulations consider a $$$\oslash$$$ 10 cm coil constructed from $$$\oslash $$$ 2 mm copper wire placed directly on the phantom. Inset dashed line indicating 0.75T.

Figure 4: Measured and simulated B₁⁺/√P maps in the central sagittal and transversal slice (FOV=256x256mm², 0.5x0.5mm² in-plane resolution, 1.5mm slice thickness, V_ref=150V). Different orientations are indicated by black or green dots: their location corresponds to the coil conductor position and their size indicates the oCo current density (inhomogeneous for the 1G, homogeneous for the 2G coil). Thin black lines in transversal measured maps mark the boundaries of regions where the flip angle was higher than the dynamic range of the mapping technique.

Figure 1: (a) Schematic showing gap positioning, (b) photographs of fabricated coils and interface circuit, (c) coaxial cable cross section, (d) phantom photo and investigated coil orientations.

The process has three major steps. First, the design is made in a CAD software. Then the part is 3D printed and finally the conductive polymer is electroplated with copper to increase its conductivity.

In the top row are renderings shown for the intended wrist coil. In the bottom row is the actual 3D printed and electroplated coil displayed. Not only does the fabricated coil look similar as in the renderings, they also closely match the geometrical dimension as specified in the CAD.

Figure 1. A photograph of the proposed, multi-channel J-pole antenna array showing its essential parts and components.

Figure 3. Attenuation maps of the proposed J-pole array (left) and the reference array (right). The centre part of the antennas was measured.

Figure 4: Qualitative comparison between the simulated B1+ maps, for one representative coil element of the array and the middle sagittal slice. The results for the incomplete and full matrix are displayed for three canonical ranks (4,10,16), along with the relative error with respect to the ground truth (left) in logarithmic scale. Voxels outside the scatterer are masked for enhanced visualization.

Figure 3: Overall compression factor (left axis) and relative error (right axis), for the full and incomplete matrices and various canonical ranks.

Figure 1. Coil arrangements. (a)Basic configuration of loop/CRC hybrid multi channel array (LCA) RF coil.(b)Developed torso coil using LCA(8ch). (c)Developed spine coil using LCA(8ch). (d)Conventional body coil for a vertical field MRI(6ch).

Figure 3. Phantom experiment at 1.2T vertical field MRI. (a)Developed body coil (LCA torso and spine). (b)Conventional body coil(SSCA).

Figure 4 A and B: Simulated receive sensitivity (B₁^-) maps in the central axial slice of a cuboid phantom (15x30x15 cm³), with one coil set to 1W input (active) and the other coil terminated with the preamplifier loading (passive). The B₁^- distortion due to non-perfect decoupling can be easily observed in the residual B₁^- maps calculated by subtracting the baseline B₁^- of a single coil. C and D: Plots of normalized B₁^- values at the 10-cm-deep region (white dotted circle in Figures 4A and B).

Figure 5 Illustration of an example of a 29-element high-density over-overlapped coil array focusing on prostate imaging. A: Planar unfold view. B and C: Three-dimensional view.

Figure 1: RF currents flow on the integrated RF/wireless coil design for simultaneous MR image acquisition (red) and wireless data transfer (orange), while a DC voltage (blue) is applied to the PIN diode from a GPIO pin from the Wi-Fi transceiver module for Q-spoiling.

Figure 3: Gradient-echo EPI mean images, SNR maps of the mean images, and temporal SNR (TSNR) maps of the image time series acquired in a water phantom, showing no degradation in image quality for wireless Q-spoiling using the iRFW coil compared to conventional wired Q-spoiling.

Fig. 1. Universally sized AIR 64-Channel phased array head neck coils for carotid, head/neck and cervical spine MRI (a) Prototype1 16-Channel neck/cervical spine anterior setup coil assembly with 48-Channel head coil (b) Prototype2 16-Channel neck/cervical spine posterior setup coil assembly with 48-Channel head coil

Fig. 5. (a) Sagittal T2, 20 cm FOV, 2mm slice thickness image and (b) MRA images obtained with the protoytpe1 16-Channel neck/cervical spine coil with 48-Channel head coil. Axial reformatted images of the cervical spine from a 3-D CUBE T2-weighted FSE sequence (0.7mm isotropic) demonstrates improved SNR with the (c) protoytpe1 16-Channel neck/cervical spine coil with 48-Channel head coil compared to the (d) conventional coil. Note improved visualization of the intrathecal nerve rootlets (arrows) with the prototype1 coil.

Fig3. Absolute Metabolite concentration map in mMol unit obtained from ERETIC (3^rd column) and IWR (4^th column) methods corresponding to the anatomical MEMPRAGE image (2^nd column) and MEMPRAG image down-sampled to MRSI size (1^st column ).

Fig4. Bland-Altman plots corresponding to the metabolite concentration map shown in Fig.3, which indicates the agreement between IWR and ERETIC methods.

\(Fig.1\). Block diagrams of the simulation methodology: TLE user interface to 3D microcoil model

\(Fig.2\). The complete electric model of 3D NMR microcoil prototype

T2w images with SENSE 1 (left), 2 (center) and 4 (right). Bottom row shows magnification of the prostate. From an acceleration of 4 small artefacts are observed, as well as the expected reduction in SNR.

3D T1w acquisition of the lower body. Left, no SENSE acceleration, right: 3x3 SENSE acceleration. Artefacts are predominately in the peripheral region of the body.

Fig. 1: 4-ch coaxial coil module. The compact cylindrical housing (diameter = 5.2 cm, height 3.2 cm) surrounds the PCBs with all electrical components and four preamplifiers.

Fig. 2: a) Differently arranged 12-ch layouts built from the same modules M1, M2 and M3. b) MR images (GRE2D) acquired on a flat phantom in a slice at a distance of 5 cm to the coil arrays. The distance from the arrays to the phantom is around 5 mm.