Pre-clinical MRI systems (small-bore animal scanners or spectroscopy systems) have been used to examine small human brain tissue samples post mortem in order to investigate fundamental neuroanatomy questions at the mesoscale ^1-4. These studies take advantage of the high field strengths and increased gradient performance of preclinical systems, but are inherently limited to small human tissue samples smaller than ~20x20x20mm³, thus restricting size of the brain region examined. Extending such studies to larger samples, while maintaining optimal Signal-to Noise Ratio, requires sample-size specific RF-coils used in a large bore system ⁵. Previously, we showed the results of using a single-loop RF-coil to image moderately sized specimens (~40x40x20mm³) at high resolution using a large-bore human 9.4 T MRI system ⁶. Here, we test extending high resolution post mortem MRI to larger tissue samples by designing and building a cylindrical phased-array receive coil (Rx) and a suitable sample holder for use in a large-bore 9.4T system. The aim was to achieve high resolution anatomical images for a large ~80x80x80mm³ human occipital lobe sample, which would be difficult to achieve either in a preclinical system due to space constraints or when using an RF-coil designed for imaging the human brain in vivo.The receive array consists of 16 individual coil elements in a phased array, arranged in a 8x2 matrix around a 9 cm external diameter hollow Polycarbonate cylinder (Figure 1 (L)). The coil layout allows an 80x80x80 mm³ cylindrical imaging volume, ideal for large tissue samples. Each coil element is ~4.2 cm in diameter, and consists of a balun circuit along with a tuning/matching network, in order to ensure optimised sample loading. Each coil is connected to low-noise preamplifiers using λ/4 length coaxial cables and lumped-element networks, which also help in preamplifier decoupling and impedance matching. Cable traps were included to help reduce any common mode current effects, and were fashioned by winding the coaxial cable (connecting the preamplifier to the coil plug) around a nylon screw and soldering a tuning capacitor across its ends. The coils were all tuned and matched to 399.7 MHz, with a loaded reflection coefficient better than -20dB for each coil element at 50 Ohms. Preamplifier decoupling for all Rx elements was better than -25dB. A separate 16 channel Tx-array coil was used as a transmit coil⁷ (Figure 1(M)).The sample holder was designed using SolidWorks and 3D printed with SOMOS Watershed XC11122 material (DSM, Heerlen, NL), which has susceptibility close to that of water, thus helping us reduce susceptibility mismatch effects. The sample was preserved in Formalin and placed inside the container, which was then sealed using a vacuum sealing paste (Dow Corning, Michigan, USA) (Figure 1(R)). The noise correlation for the Rx array ranged between 0.1% and 49%, with an average of 15% (Figure 2). Images were acquired using a Gradient Echo sequence at 100μm isotropic (TR/TE = 19ms/7.5ms, FOV = 83.2mm, flip angle = 25˚) in an investigational 820mm bore human 9.4T magnet equipped with an 80mT/m maximum amplitude gradient set (Siemens MAGNETOM 9.4T with AC84mkII).The high SNR achieved resulted in 100μm isotropic anatomical images (Figure 3), which show good delineation between white and gray matter in the occipital lobe over the entire imaging volume. We notice a slight signal drop-off towards the center of the sample under investigation, which can be corrected with better B1+ shimming or by using kT-points technique for a more homogeneous sample excitation.We show that by using a combination of specialized RF coils specific to a particular post-mortem tissue sample and a high magnetic field, we can achieve very high resolution anatomical imaging of human brain samples. Further work would include higher resolution anatomical imaging and constructing larger receiver arrays to accommodate samples as large as 120x120x120 mm³.