Figure 3: a, b) In-vivo $$$B_1^+$$$ efficiency (normalized by the input power) show a 1.7-fold improvement in the inferior regions of the brain (outlined in red). c, d) In-vivo SNR maps for a subject show that placing the RF sheet enhances the SNR in the cerebellum and brainstem. A 2.0-fold SNR improvement was obtained for the region, outlined in red. The reference voltage was kept constant in with/without modes.

Figure 4: Low flip angle T1- weighted GRE MR images (at two different axial slices) of the lower brain obtained with (right column) and without (left column) the RF sheet show signal enhancement in the inferior regions of the brain improving visibility. Low flip angle sequence was used to avoid RF over-flipping in the close vicinity of the RF sheet.

Figure 2. CST simulation models of the folded-end (A) and unfolded (B) dipole arrays without decoupling. CST simulation models and results (B₁⁺, B₁⁺ ratio) for 1x8 bent folded-end dipole arrays decoupled using the common parallel dipoles (C) and modified perpendicular dipoles (D). All arrays are loaded by the head and shoulder (HS) phantom. Transmit dipoles and passive dipoles are shown in black and red, respectively. B₁⁺ ratio is calculated by dividing corresponding B₁⁺ maps by one obtained using the array without decoupling (Fig.2A).

Figure 1. A) Current and voltage distributions along the length of the full-length dipole vs a “short” dipole. B) Current and voltage distributions along the length of a folded-end dipole. C) Two modes of a pair of coupled dipole antennas. Dashed curve shows the resonance line of a single dipole. D) General representation of the decoupling effect produced by an addition of the common straight parallel (top) and modified perpendicular (bottom) passive dipoles both shown in red. To increase the electrical length, both passive dipoles have a lumped-element inductor connected in series.

Fig. 1. Comparison of the intraoral coils. (A) Reference loop coil model with the line profile tracks and levels marked. Planar dipole models with (B) ribbon arms, (C) wire arms, (D) meander wire arms, (E) 3-wire arms, (F) 3-wire arms with shorted ends, and (G) 7-wire dipole with high-εr cap. Field maps are shown in the same columns. (H)-(N) Sensitivity within the tongue region is lower in dipoles compared to the loop coil. (O)-(U) Ribbon and multi-wire conductors have better sensitivity across the dipole arms. High-permittivity cap increases transmit efficiency and homogeneity.

Fig. 4. A photo of the in vivo measurement setup. SNR maps at z = 10mm away from the coil planes for intraoral B) loop, C) dipole, D) extraoral SLR only, and SLR combined with E) loop and F) dipole in Tx mode. The image intensities are scaled to the same range to allow a visual SNR comparison. SLR combined with the intraoral dipole provides higher SNR and less signal from the central region of the tongue.

Figure 1. Photographs of the flexible 6-channel array with the protective cover removed in order to view the electronics: (a) unfolded; each coil is loosely secured to the elastic shell with integrated buttons (white circles), (b) zoomed view of elastic pockets that maintain approximately critical geometric overlap between neighbors while allowing flexion, (c) applied to the knee in natural posture, and (d) applied to the flexed knee. Inset S₂₁ values in (c) and (d) show excellent decoupling between neighbor coils in both postures.

Figure 2. SNR maps acquired with the proposed flexible coil (top row) and commercial coil (bottom). The proposed coil provided similar SNR (within 3%) as the commercial coil in the central articular cartilage while enabling approximately 40° knee flexion (right column). The maps were normalized to the SNR value in the articular cartilage during knee flexion.

Fig. 3. Optogenetically evoked BOLD responses in bilateral enhanced FP-S1 regions. a) Time course and spatiotemporal map of bilateral BOLD responses induced directly from optogenetic stimulation in the right FP-S1 (right) and projected left FP-S1 (left, n = 3 rats, 49 trials). b) BOLD-change in each voxel along cortical depth on both hemispheres (n = 3 rats, 49 trials). c) Different laminar-specific responses of both hemispheres. d) Significantly higher tSNR with (blue) than without (red) implanted inductive coils (paired-sample t-test, ***P <0.001, n = 3 rats, mean ± SEM).

Fig. 2. BOLD responses detected using a line-scanning method in an enhanced region. a) Experimental setup. b) Enhanced focal intensity in the right FP-S1. c) Significantly higher SNR in ROI 1 with the inductive coil. d) BOLD activation map. e) The procedure for line-scanning method. f) BOLD change and spatiotemporal map in the cortex along the cortical depth (20 voxels, 2 mm, n = 3 rats). g) Resting-state hemodynamic fluctuation from right cortex and spatiotemporal map along the cortex. Right, significantly higher tSNR with the inductive coil (blue) than previous results (red).

Figure 2: Numerical simulation results of the B₁⁺-field for three cases: (a) the reference case with voxel model only; (b) the dielectric pad attached on the top of the abdomen; (c) the metasurface attached on the top of the abdomen.

Figure 4: Experimentally measured MRI scans, with Body Matrix Coil used as receive coil, of a volunteer for three cases: (a) the reference case without pads or metasurface; (b) the dielectric pad is attached on the top of the abdomen; (c) the metasurface is attached on the top of the abdomen. The imaging parameters for T2-weighted HASTE sequence: flip angle = 90⁰, TR/TE = 2000/97 ms, acquisition matrix = 256×179, field-of-view = 306×350 mm².

Figure 1: A) Models of the (left) I-MARS Meander, (center) I-MARS Paddle, (right) Fractionated dipole. The I-MARS Paddle had a narrower center compared to the ends to achieve a more focused field distribution.; B) Coronal cut-away views for I-MARS Meander and I-MARS Paddle models; C) Sagittal cut-away view for the I-MARS Paddle, showing the location of the tuning capacitors and balun; D) Axial slice of Duke and a coil element to compare the stabilities of the different coils; E) Axial slice of Duke and eight coil arrays to compare RF shimming performance, B₁ and SAR efficiency.

Figure 3: A) Axial GRE slices of a phantom, with the I-MARS coil arrays in prostate configuration. The reference voltage (V_ref) required to calibrate the B₁ in the ROI (red square) was compared, as well as the SNR in the ROI. B) we-DESS images of the hip acquired with the I-MARS Meander and I-MARS Paddle arrays, comparing the V_ref required to calibrate the B₁ over the femoral head. C) we-DESS images of the shoulder acquired with the I-MARS Meander array.

Figure 1. (a) Conceptual drawing of the proposed RF coil element. (b) Inductance and capacitance changes with stretch.

Figure 4. Photographs of the proposed coil in (a) unstretched and (c) 20% stretched positions and (b), (d) the corresponding sensitivity profiles. (e) Normalized SNR of the relaxed (blue) and 20% stretched (orange) coil as measured transversally at the depth of 3cm inside the load. The color bars illustrate approximate width of the relaxed (blue) and stretched (orange) coil. (f) First in vivo MR image of a wrist made with the proposed coil.

Figure 2. Design drawing of eight-channel TEM transceiver head coil with key components labeled.

Figure 1. Head-only, 1.5T MRI system. Features include YBCO HTSC cryo plate cooled asymmetric magnet, multi-coil field encoding/shimming, 8-channel FPGA controller, non-uniform field imaging, and our eight-channel TR+STAR coil featured here.

Figure 1. Proposed LWA design: (a) top view; (b) side view

Figure 4. Simulated B₁ for the four-element LWA array configuration (a) and a fractionated dipole array (b) in the transverse slice through the prostate of the human body Duke anatomical model for 1 W of accepted power. The corresponding SAR for the LWA array (c) and fractionated dipole array (d) in a transverse slice through the maximum of local SAR.

Results of measurements performed on a phantom with a single antenna. (a,b) saggital slices of B₁ maps (method: DREAM¹⁶ FA/steFA/TE/TR 10/60/1.4/4 ms). (c) profile of B₁ magnitude along red lines in figures a and b. (d,e) Temperature maps at the final timepoint, based on proton resonance frequency shift¹⁷. (FA/TE/TR 110/10/15 ms, heating with 20 W average power, duty cycle 10%, 100 kHz off resonance block pulses) (f) maximum heating as a function of z-position. (g) maximum heating in the whole volume, as a function of time.

Photographs of the constructed antenna. Cable type: Huber Suhner RG223u (a) Overview of the antenna, with various components as indicated and a ruler for scale. (b). Close-up of one of the ends. Inductors at the end were hand-wound using 4 mm long sections of annealed copper wire. The correct inductance value was determined by measuring the reflection at the port and choosing the inductance value that results in an admittance such that Re(Y) = 1/50 S. (c) Matching Circuit.

Figure 1. Diagram of the power splitter model. The final ratio between P₁ and P₂ are adjusted by varying the input port’s position (change a and b).

Figure 4. a) Model with daughterboard positioned to have power splitting ratios indicated below. The red dotted line indicates the splitting position. Predicted (b) and measured B₁+ amplitude maps (c) with different power split ratios, and the differences between the predicted and measured maps (d).

Figure 1: 4-channel stranded wire coil array (a) and coaxial coil array (b) sewn onto fabric to maintain coil overlap for optimal geometric decoupling. PIN bias cables are connected to coil interfaces in the middle. In (b) a floating cable trap used in both setups is shown. Circuit diagrams of coil interfaces are shown below for SWC (c) and CC array (d), containing tuning, matching, active detuning, balun, phase shifter, fuse and preamplifier. 1-channel coil interfaces did not include balun, phase shifter and fuse.

Figure 5: Noise correlation matrix of the 4 channel receive elements of the stranded wire coil (left) and the coaxial coil (right) ranging from 0 (no correlation) to 1 (highest correlation).

Fig 1: Two different types of designs showing the difference in the structure but similar shape.

Fig 2 The coils showing the Magnetic field at the same range with different coupling effect.

Figure 1: Top and bottom sides of prototype PCB with single transmit relay installed

Figure 2: NMR spectrum of DI water sample (1 ml in 10 mm sample tube) for SNR comparison: 8 averages, 1 ml DI water in 10 mm sample tube, 276 kHz (6.5 mT) Larmor frequency, 90 degree tip, 300 us pulse width/TR 15 seconds, 2048 points zero-filled to 8192 points, 500 Hz SW

Geometry of the models of the coils and of the human body

Plot of the average value of the SNR (red lines) and the transmit efficiency (blue lines) in the heart for different values of the permittivity for the two pads.

a) geometry of the double-folded dipole antenna and the circuit diagram of the LC lattice balun matching network, b) simulation geometry, c) constructed double-folded dipole antenna with LC lattice balun matching network

Comparison of the safety excitation efficiency profiles along the X axis (Z=0cm) between the plain dipole, fractionated dipole, and the double-folded dipole with 20mm, 30mm, 40mm and 50mm gap for the depths Y=0cm, Y=5cm, Y=10cm and Y=15cm

Figure 4. Transmit performance of the two arrays. First row: An axial view of B₁⁺-maps obtained by 3D phase-only shimming over the heart with maximum homogeneity constraint. Second row: A coronal view of the B₁⁺-maps obtained by the same shimming solution. Third row: Anterior and posterior views of the 10g-averaged SAR caused by the same shimming solution. The input power level was adjusted to achieve the average 1µT B₁⁺ over the heart.

Figure 2.Coronal and sagittal views from the formation of 16-channel 1D loop-dipole and 2D NODES_Rx over the anterior and posterior sides of a realistic human body model in the EM simulation environment. These EM simulations were used for evaluation of the transmit and receive performances of the two arrays in cardiac imaging at 7T.

Experiment 1, with coils to the left and right. Figs. a and d show the initial and final setups, respectively. Figs. b and e show the initial and final transmit magnetic fields. Figs. c and f show the difference between the transmit fields and the desired fields.

A visual representation of the coil optimization procedure.

Signal conversion setup. The signal is transmitted by coaxial cable in path A, and by coaxial cable with an EO conversion and an OE conversion in path B. The amplitude of each signal is monitored by a spectrum analyzer.

Images acquired using regular galvanic connexion (left) and OE conversions with fiber optic transmission. The SNR are respectively 162 and 5 (ratio of 32). Inner diameter of tube is 15mm.

Fig 3. Side-view (cross-section) MRI mappings of contours of equal magnetic field magnitude above the planar meander-line surface coils shown in figure 2. Field contours are shown above even meander-line coil with 8 current lines and above odd meander-line coil with 7 current lines and added split-current return path.

Fig 2. Photographs of two meander-line coils. (A) Coil with an even number (8) of wires. (B) Coil with an odd number (7) of wires and an added split-current return path

Figure 4 UVC fixture operating and disinfecting the inner surfaces of the bore and patient table on a 3T Signa Premier system

Table 1 shows the used UVC dots and corresponding UVC dose measured on various GE MR systems_{(Color representation in the table may not be accurate due to image capture challenges. Refer to the corresponding numerical dose value presented)}

Figure 2. Exemplary MRI results for homogeneous phantom images (bottom) and kiwi fruit images (top) without / with the smart metasurface. The SNR in the indicated ROI is given in the subplots, respectively. For the homogeneous phantom images (TR = 100 ms), the imaged slices are orthogonal to the metasurface while for the kiwi fruit images (TR = 1 s) the slices are parallel to the surface.

Figure 1. a) The manufactured smart metasurface prototype. b) MRI results for scans with different nominal flip angles. The SNR in the ROI for the kiwi fruit scans, see Fig. 2, is shown for imaging with the body coil, the body coil in combination with the smart metasurface, and a local single loop coil (SLC). For these scans, TR = 100 ms was used.

Fig.1. Simulation model showing the surface coil with the placement of the capacitors, feedline, and the dielectric sheet over the Cmode capacitor.

Fig.3. Electric fields produced on the surface of the coil where the dielectric sheet is placed on the C_mode capacitor. The Feedline/Power source used is the function of the simulation software used and not the part of the RF coil in practice. Hence, The unusual increase in the electric field around the feed line can be neglected.

Figure 1. Experimental setup. The metasurface pads (MS) were placed between the receiver array coils and the phantom, one on top and one bellow.

Figure 2. Flip angle maps and profiles in sagittal and transversal orientation. First row shows the reference (Ref) maps acquired without the metasurface pads (MS). Second row: maps acquired using the pads (gray rectangles). The regions of interest are indicated with white arrows. Bottom row: sagittal and transversal profiles along the dashed lines in the upper row.

Figure 1: System Photo

Figure 2: System Block Diagram and Circuit

Topology of the proposed GaN T/R switch. The impedance transformation is shown with lumped elements but could also be done by distributed elements or transmission lines. The control of the gate voltage can be as simple as a resistor between a digital logic output and the MOSFET gate but can be extended to arbitrary waves to smooth transient voltages in the RF circuitry.

Pictures of the implementet T/R switch with a few key parts marked. As can be seen, most of the board space is occupied by RF circuitry.

Figure 2a) Block diagram of the prototype feedboard. The switching circuitry is integrated into the coil’s impedance matching network. The coil tune capacitor is split into C_T1 and C_T2. C_T1 comprises fixed capacitors and switches S₃ and S₄, allowing the tune frequency to be adjusted. b) Photograph of the prototype feedboard.

Figure 3) Measured waveforms at the end of a 200W hard pulse. A) shows the sates of the RF pulse and the two switches. B) Shows the voltage from a pickup loop coupled to the coil. C) is the voltage across S₁, measured with a high voltage differential probe.

Figure 5. Comparison of calculated SFF numbers for each Tx – Rx pair in free space (blue) and within the bore (red). Note that SFF = 1 is a perfect reconfiguration, SFF = 0 means no correlation at all. Those within the bore resulted in lower numbers for all cases, compared to the free space. The SFF at the center pair (A_C) shows the minimum difference between free space and the bore, whereas the left and the right pairs represent a relatively large difference.

Figure 3. Calculated group delay distribution in nanoseconds for each antenna pair on the workbench, in free space, and within the bore. Group delays within the bore show higher dispersion for all cases, compared to the free space. These results are corresponding to the transfer functions in figure 2, noting that group delay is the phase gradient over frequency.

Figure 1: Flexible Tunable Capacitor schematic and PCB layout with detailed features, a) The basic schematic consists of capacitors $$$C_s$$$ and $$$C_p$$$ in parallel. $$$C_s$$$ is generated by a series combination of $$$C_{sa}$$$ and $$$C_{sb}$$$. All three variable capacitors are arrays of parallel plate capacitors connected in parallel. b) The top and bottom layers of the flex PCB. The exposed traces can be cut to reduce the capacitance of $$$C_{sa}$$$, $$$C_{sb}$$$, and $$$C_{p}$$$.

Figure 2: RF Simulation Results, a) Simulation setup, impedance measured across frequency (100 MHz to 400 MHz). b) CAD model used for RF Simulations. c) Simulated Capacitance compared to ideal calculations. The variable capacitance ranges from 0.5 pF to 7 pF. d) Step size variation across total capacitance (RF Simulation). Maximum step size 0.25 pF (0.225 pF in ideal calculations).

Left: Illustration of the RF topology of the high-power T/R switch. The antiparallel PIN diode pairs are driven in a cascaded way. Towards the RX port, PIN diodes with shorter carrier lifetime are deployed.

Right: Implementation of such an RF topology. In addition, auxiliary circuits and low noise amplifier (LNA) can be seen.

Top row: transient switch behavior when switching from the TX to the RX state. The reverse bias is built up in a cascade, which reduces the transient voltage peaks.

Bottom row: Snippet out of the passive self-triggered PIN diode driver. As long as a PIN diode is still in its low impedance state, it has a forward voltage present, despite bulling a reverse current out of it. When the PIN diode changes to its high impedance state, current is drawn out of the next stage.

Figure 3. Circuit-level simulation results of the preamp decoupling abilities in different types of coil. This ability was evaluated by the resonate response difference between the termination of 50Ω and the termination of the preamp. The real part of the preamp (Re[Z_in]) was set with different values from 0Ω to 10Ω. In practice, Re[Z_in]< 5 Ω is recognized as an acceptable low-input impedance. For LIC, C_m=59.9 pF and C_t= 33.1pF. For HIC, L_m= 483.8 nH and C_ms= 7.1 pF. For small-C_m LIC coil, C_m = 5.5 pF, C_t = 1000 pF and C_ms = 7.1 pF.

Figure 5. SNR and noise correlation simulation results. A: Calculated SNR maps in the central transverse slice. B: Average noise correlation of neighbor coils (i.e., left-middle and right-middle coils). C: Calculated SNR values in an 8-cm-deep region that is directly under the middle coil.

Fig. 1A) Surface current density plots for the coil setup A1 (100 mm diameter) evaluated at frequencies for 3T (top) and 7T (middle) MR as well as at its f₀ (109 MHz, bottom) on the three substructures iC, oCi and oCo. B) shows the corresponding surface current plots for all three frequencies (3T, 7T and f₀ in green, blue and black, respectively) on the aforementioned substructures.

Fig. 2 Simulation results for all setups. Ratio of the mean surface current at the operating (Larmor) frequency over at self resonance plotted against the deviation from the self-resonance f_L/f₀. The size of each data point is proportional to the electrical stub length over the wavelength in the cable.

Figure 1: Representation of simulation design for a) the current 16 channels Tic Tac Toe (TTT) coil; b) the conformal 16 channels TTT coil; c) the new conformal 32 channels TTT coil with 2 channels per panel. A single panel of the 32 channels conformal design is represented in d) and the photo of the actual built panel is represented in e).

Figure 4: Comparison between the magnitude of experimental B1+ maps of a conformal (a) and a planar (b) 4.25in Tic Tac Toe panels, acquired using the same spherical phantom at approximately the same distance between the panel and the phantom. The B1+ maps shown are the center slices for each orientation (axial, sagittal and coronal). Intensities are normalized to the same scale.

Figure 1: Numerical model of the 2-channel, 32-rung high pass birdcage coil (a) and the numerical pregnant body models (b) used in this study. Only the skin, uterus and fetus are shown in the body models for simplicity. GA: gestational age, w: weeks, BMI: body mass index.

Figure 3: Overall (maternal and fetal) peak local SAR, fetal peak local SAR and fetal average SAR for different RF shim settings where relative amplitude and phase of the two channels is varied from 0 to 2 (vertical axis) and from -90° to 270° (horizontal axis) respectively. All shim settings are normalized to maternal whole-body average SAR of 2 W/kg. CP mode: circularly polarized birdcage mode, improved imaging performance: both average B₁⁺ and B₁⁺ variation is improved compared to CP mode, no SAR increase: maternal or fetal SAR does not increase compared to CP mode.

Figure 1:

Two images to the left: Examples of B₁⁺ maps after 2D and 3D RF gain optimization from the same subject. The 2D gain optimization for this subject caused B₁⁺ of 160% in the center of the brain. The 3D optimization caused B₁⁺ of 140% (yellow arrows). The green arrows indicate lateral area typically presenting B₁-induced inhomogeneity whose severity vary particularly with head size and placement of dielectric pads.

Two right images: Example of a bias field corrected 3D MPRAGE image and the computed bias field. The red arrows demonstrates 40% MPRAGE magnitude variations.

Figure 2:

Plots showing variability of CV across the slices in the individual brainmasks for group 1 (black, n=48), group 2 (blue, n=29) and group 3 (red, n=18). The slice-indexing is feet-head, i.e. high numbers corresponds to the top of the head.

Figure 3: The data acquired on subj-02 on the same scanner using NOVA Coil #1 with two different dielectric pad configurations. (A) Schematic of dielectric pads placement (pink is the additional pad used), (B) B1+ maps (C) perfusion-weighted images. Overlaid rectangular outline indicates the slab coverage for perfusion imaging and cyan outline is the low B1+ from the 1-1 configuration. Note the impact of the 2-1 pad configuration, best illustrated in the axial slices.

Figure 4: (A) B1 histograms in left and right hemispheres of subj-02 for 1-1 and 2-1 dielectric pad configuration (inset schematic) (B) Histogram showing the improvement in B1+ in the 2-1 configuration (region-of-interest is the low B1+ outline in Fig 3B top row). Yellow background and dotted line indicate 95% efficiency threshold for the tr-FOCI inversion pulse.

Four proposed structures for exciting two coil elements with two RFPAs. A Conventional pTx excitation with one dedicated amplifier per coil. B Power sharing with a 90° hybrid coupler. C Power sharing with a 90° hybrid coupler and phase shifter. D Power combination with a Wilkinson power combiner.

Top row is a shim well-suited to power sharing and bottom row is a shim poorly-suited. Left column shows RFPA output power for the standard system and system with the power sharing network. The right column shows coil drive power levels for the standard system and power sharing network.

Figure 1 – EM simulation model of a shielded 3T 4-channel TxArray coil loaded with a detailed human head model¹². The shield is slit into four segments evenly distributed along the axial direction, where the adjacent slits are connected via two 3 nF capacitors at positions facing the coil's end-rings.

Figure 4 - Magnitude and phase profiles of the B₁⁺ field for all lumped ports/elements in the central axial plane. For each B₁⁺ profile, the corresponding port was fed by a 1 volt, and other ports were terminated.

Fig. 1 - (a) Sagittal T1-weighted image indicating the pCASL imaging region (orange), TOF coverage (green), labelling plane (yellow), position of coronal slices for ROI definition as depicted in (b) (blue), and control region for field homogeneity (white); (b) TOF coronal slices showing the ICA (red) and VA (blue) ROIs; and (c) corresponding B0 fieldmap space (where the labelling plane is depicted in yellow), from which the ∆B0 values are extracted for the artery ROIs.

Fig. 3 - B0 fieldmap for an illustrative subject, with RF coil shiming “off” (top) and “on” (bottom). The reduction in field inhomogeneity between the shim “off” and “on” conditions is evident in the posterior neck region, as indicated by the arrows.

Fig. 5. In vivo MRI experiments (3D-GRE: TR/TE = 6.5/2.82ms, FOV = 256x240 mm², slice thickness = 1.0 mm, FA = 4º, reference transmit voltage = 100 V) in one human male subject using two blocks: thinner (d/b = 0.25) and thicker one (d/b = 0.75). Three different regions of interest (head, calf, wrist) were investigated. The quality of all of the images was significantly compromised for the larger block (very noisy). The overall quality of the images depended on the level of curvature of the anatomical structure. The acquisition parameters of the RF pulse sequence were used to scan each body part.

Fig. 4. Visualization of dielectric modes: the comparison between the electromagnetic field simulations and MR measurements for two elements: thinner one (d = 0.25b) and thicker one (d = 0.75b). GRE imaging was used (TR/TE=8.6/4.0 ms, FOV=250 x 250 mm², slice thickness = 7.0 mm, FA=15º, reference transmit voltage = 5 V). The simulations are in an excellent agreement with the measurements and show significantly different magnetic field distribution between the blocks. The mode that propagates within the thinner block was interpreted as TE_11δ^z, and within the thicker one as TE_1δδ^y.

Fig. 4. Intrinsic SNR (iSNR) maps of the 32-channel loop array (a) and 32-channel sleeve antenna (b) arrays with phantom in the axial, coronal, and sagittal plane. Ratio maps (c) between (b) and (a) and profiles (d) along the indicated lines.

Fig. 1. Photographs of the (a) 32-channel loop and (b) Sleeve antenna arrays

Fig. 3. A) Electromagnetic field simulations: B₁⁺/B₁^- field distribution in the spherical phantom as a function of distance (5-mm: tuned, 20 and 30 mm: detuned). B) Difference maps showing the expected gain in B₁^- for the dielectrically-shortened dipole antenna and its combination with the loop element (especially for the detuned case).

Fig. 1. Simulation view and photos: A) Diagram of the loop coil B) Diagram of the inductively-shortened dipole antenna (Ind), C) Diagram of the dielectrically-shortened dipole antenna (Diel) D) Comb 1 (Loop+Ind) E) Comb 2 (Loop+Diel).

Figure 5: Mechanical design of the 48-channel self-decoupled Tx-only array along with the anatomy-fitting Rx coil. This design could be used for human brain and spinal cord imaging as well as brain only imaging.

Figure 4: (A) and (B): Diagram of a single self-decoupled loop (Yan et al, 2018, Nat. Commun.) coil and a single self-decoupled dipole antenna folded in figure-of-8 shape (Lu et al, 2020, ISMRM). (C) and (D): Simulation model of the 48-element Tx array and truncated human model, along with the simulated S-parameter plots of 3 loops and 3 folded dipoles in two rows. Isolation between any two of 3 loops and 3 dipoles is better than -15 dB by using self-decoupling technology. Loops in adjacent rows are overlapped to ensure enough B₁ field coverage in the longitudinal direction (z-direction).

A 64-channel high-density receiver array as shown is combined with an 8-channel transmit head-coil (not shown).

High resolution (0.8mm isotropic) MPRAGE with SENSE (left) and Compressed SENSE (right) obtained in 4:35 min.

Fig 3. A. Side and top views of the constructed 30-element coil. B. The Nova 32-channel receive coil inserted to show the excellent fit inside the coil. The receive coil sits on an acrylic part so that the pTx coil can slide easily without moving the patient from the helmet. C. The 30-element coil without the shield is shown in detail. All coils are tuned and matching to 298MHz and have a lattice balun at the feedport to reduce cable currents. The head-shaped phantom (The Monster Makers, Cleveland, OH, USA) was filled with a tissue-mimicking solution to tune and match the coil on the bench.

Fig 4. A. The simulation results from a 2x2 self-decoupled and overlapped array in Ansys HFSS (Canonsburg, PA, USA). This decoupling strategy is used throughout the constructed 30-element coil. B. The plot shown is the S11 measurements for all 30 coils at 298MHz. C. A selected set of coupling bench results are shown for a section of the array (coils 2,3,4,12,13,14,22,23,24) for ease of understanding. These decoupling results carry over for the entire coil. All measurements were taken on a Keysight 4-port VNA E5071C and ports that were not being used were terminated using a 50-ohm load.

Figure 3: a) SNR Images of cylindrical phantom from 3 different knee coils (Rigid, Medium Flex, Large Flex) for the central slice, with masked images for the outer 1/3 and inner 2/3 of the phantom b) SNR Images of 6-tube phantom for the same knee coils at the same slice, with masked images showing the top and bottom tubes c) Plot of the average SNR with standard deviation bars for cylindrical phantom images d) Plot of the average SNR with standard deviation bars for 6-tube phantom images.

Figure 1: Flow chart of the pseudo-multiple replica method for calculating SNR shown in Robson et al.³

Figure 4) SNR maps (AU) of Baseline and with uHDC blocks. The SNR is enhanced significantly (up to 4 times near the surface of the phantom).

Figure 3) Noise correlation matrix of the Baseline and with uHDC blocks. This figure shows that the uHDC blocks reduce considerably the noise correlation between adjacent loops as well as non-adjacent loops.

Fig. 1 ― a) Image acquisition (Exp.) and modelling (Mod.) with a copper (top) and a superconducting (bottom) surface coils. b) Adjusted quality factor Q_mod as a function of B²_1,nom evaluated with the model for both the copper and HTS surface coils.

Figure 1: (a) 4-Channel High Impedance Coil placed on the Robotic Motion Device. (b) Complete view of the 4ch HIC including transmission lines, cable traps and preamplifiers. (c) Medial and (d) lateral close-up view without transmission lines, cable traps and preamplifiers.

Figure 3: Comparison of SNR maps of the HIC (top) and the LIC (bottom) using a foot-shaped agar phantom.