It is well known that in the case of a sufficiently small coil, thermal noise of the system is dominated by coil noise and that cooling of receiver coils, made either of Cu or superconducting materials can provide very significant improvement of SNR [1-3]. Recently, we have further developed [3-5] our 7 T MRI ¹H (300.3 MHz) receiver probe with the tuning/maching and decoupling circuitry. The new probe was made to meet specific operating requirements such as low-loss electronics, stable operating frequency at cryogenic temperatures over long periods of time, very short coil-body distance, short preparation time, and user-friendly operation. Our further objective was to explore the practical parameters of the probe equipped with either copper or superconducting coil and its use for ex-and in-vivo imaging of rat brain.A single-stage GM type pulse-tube cryo-cooler, CryMech model HPT10 [6], was designed to operate horizontally. In order to meet certain cooling requirements and high-magnetic field limitations, non-magnetic titanium was used in the cold head design and the coil was separated by 8" distance from the metal cold-head (Fig. 1a). A rotary valve was also separated from a pulse tube expander (2 m stainless steel flexible helium lines). The HPT10 provides 16W at 77K and the lowest achievable temperature of the coils was measured as 55 K. The coils vacuum shroud was designed to be compatible with 116 mm Bruker transmit coils (Fig. 1bc). Both superconducting (YBCO) and Cu coils were similar in design as previously shown [5]. The matching/ tuning and detuning circuit was integrated with the coil inside the cryostat and it was designed to be the same for either superconducting or normal metal coils [6]. Measured unloaded and loaded (ex-vivo case) Qs were 280 (Cu 295 K), 750 (Cu 77 K), and 50,000 (HTS at 77 K) and 200, 460, and 890, respectively. A 7T/30 USR MRI scanner (Bruker BioSpin, Karlsruhe, Germany and a homogenous “rat-equivalent” phantom was used with the SNR measured by using the macro Auto_snr that is part of the Bruker scanner software (Fig. 1d). This method yields SNR per unit volume that is independent of the acquisition parameters and allows for comparison across different scanners and different acquisition parameters [7]. Comparison of axial ex-vivo rat brain images acquired at 65 K and at 295 K for 11 hours 22 minutes (256 slices) is shown in Fig. 3. The 50 µm isotropic resolution was achieved (3D-RARE, RARE factor 6, TR 500 ms, TE effective 38 ms, NA 8). A twofold increase of SNR was measured (Fig. 1d and 3). In vivo axial rat brain images were recorded also at 65 K and 295 K (Fig. 4). The upper left displays 2D-RARE (factor 4) images of the Splenium, upper right Bregma -3 mm, second row at left Bregma -2 mm and at right Bregma -1 mm. NA=4, acquisition time 33.5 min, TR 10010 ms, TE effective 22 ms, in plane resolution 0.1 mm, 100 slices, slice thickness 0.2 mm. For the same location as in Fig. 4, axial RGB-FA maps were recorded with an EPI-DTI protocol. The colors intensity reflects the FA-value. Splenium is coded red, the Fimbria magenta, the Fornix green and Internal Capsule blue and green. Acquisition parameters were: time 59.5 min, TR 350ms, TE 26ms, SW 227 kHz, NA=1, 2 segments (aka shots), 42 icosahedral orientated directions, 9 bzero images, b-value 800 s/mm^2, resolution 0.24x0.20 mm, slice-thickness 0.20 mm. Fat suppression and saturation bands were activated.The targeted twofold SNR gain has been realized by 20 mm diameter Cu coil for ex-vivo scans (Fig. 1d and 3). In-vivo rat brain scans for 20 mm coil showed rat size dependent gain of SNR in the range of 40% -60% (Fig. 4). It was found that optimization of the coil size for twofold gain for in-vivo measurements requires the diameter reduction from 20 mm to 18-19 mm for copper coil and to 17 mm for superconducting coil. In the latter case additional gain over copper coil of the same size (17 mm) was obtained for both ex-vivo and in-vivo cases. Our technology allows to acquire high resolution in vivo scans and high SNR of structures deep in the brain, which were only accessible earlier in scans performed ex vivo. The ex vivo scans with a slice thickness of 50 μm are close to the thickness of histological sections. We believe that our existing setup is capable to sample acceptable images with thinner slices, possibly 20-30 μm. This facilitates highly improved comparisons between MRI and histological sections.