To perform an extensive and near point to point comparison investigation between MR T₁/R₁ imaging in vivo and histological transmission electron microscopy (TEM) of the entire corpus callosum (CC) in normal rat, and to explore the relationship between T₁/R₁ and both myelin density and axon size.There were twelve male Sprague-Dawley rats weighing 338 ± 36 g included in the present study. MR T₁/R₁ imaging were performed using a 9.4T/31cm magnet interfaced with VNMRJ consoles (Varian) and a ¹H surface coil. The rat brain anatomic images were acquired with a fast spin echo (FSE) sequence (TE=10ms; TR=4sec; FOV=3.2×3.2cm; matrix=256×256; thickness=1 mm; 8 echo train length). A modified FSE sequence combined with the global saturation-recovery preparation was used to image the parametric longitudinal relaxation time (T₁) or rate (R₁=1/ T₁) of central sagittal slice of corpus callosum with 10 saturation recovery times ranging from 0.1 to 11.9s, resulting in a voxel size of 0.125×0.125×1 mm³ and a total acquisition time of 15 minutes. This R₁ measurement was repeated 4 to 6 times for each rat. One of the rats was sacrificed and the central medial sagittal of entire CC was cut off, processed and examined with the TEM (Joel 1200 EX II, Japan, histology slice thickness = 65nm). Due to the size limitation, the entire CC was cut into 4 segments for better chemical process and TEM study purpose. Each TEM image (MegAview 3 camera, ResAlta Research Technologies Corp) is equivalent to an actual sample area of 13.1 × 9.8 um². Three representative TEM images (magnification of x10000) were randomly taken within each TEM mesh grid (90×90 um²), resulting in about 800 TEM images in total. Myelin density, axon diameter and myelin thickness of CC were analyzed for each TEM images using NIH ImageJ. These parameters were averaged in each TEM mesh grid, co-registered and compared to R₁ image with similar spatial resolution.Figure 1 shows anatomic sagittal T₂-weighted image of the representative rat that underwent histology analysis, 1-D perpendicular projection of R₁ values to the longitudinal main body of the CC from the rat which underwent the TEM analysis (red) and that of the mean and SD of R₁ values of 12 rats (black). Figure 2 shows significant positive Pearson correlation between the CC myelin density calculated from the TEM images and R₁ measured using the MRI from the same rat in each individual segment, T₁/R₁ images of 12 rats created using the MRI technique and myelin density image calculated from the TEM images of CC in the rat underwent histology analysis overlaying on its T₂-weighted sagittal anatomic image. Figure 3 shows negative correlation between the axon diameter in CC measured from the TEM images and R₁ measured using the MRI from the same rat in each segment. Figure 4 displays representative TEM images, the corresponding processed binary images, and axon size distribution histograms of TEM images in CC. Figure 5 shows linear correlation between the myelin thickness of axons and myelin density, between the axon size and numbers of the axons per TEM image throughout the entire CC.The proton parametric longitudinal relaxation rate (R₁) or time (T₁) is a basic yet crucial parameter reflecting the biophysical property of tissue and is determined by the chemical composition and its exchange with the surrounding environment. It not only plays an important role in modulating the magnetic resonance imaging (MRI) contrast by judiciously optimizing acquisition parameters such as TR and TE; but also serves as a biomarker for varieties of diseases including strokes ^{(3, 4)}, tumors ^(5-7) and multiple sclerosis ^(8-10) etc. There is a significantly positive correlation between R₁ and myelin density of CC in normal rat brain, demonstrating in Figure 2. It is consistent with previous comparison studies of MRI T₁ mapping of postmortem human brain or spinal cord with multiple sclerosis and microscopic histology based myelin content measurement, which have demonstrated that T₁ is highly correlated with myelin content ^(11-14). The present study maybe more challenging in a sense of narrower dynamic range of T₁ ([(T_1-maximum-T_{1- minimum}) /T_1-maximum] <14%) of normal CC than that in the multiple sclerosis patient (>0.66 ⁽¹¹⁾, >0.34 ⁽¹²⁾ and >0.23 ⁽¹³⁾). Nevertheless, the unequivocal statistically significant correlation between R₁ and myelin density indicates that T₁/R₁ images are indeed very sensitive to myelin density and is a good indicator of the density of myelin. A weak and less significant negative correlation between the axon diameter and R₁ which is also coincident with previous reports ^(15-17). The underlying mechanism is complicated but is likely still a myelin-dependent phenomenon ⁽¹⁷⁾. This could be partially explained by Figure 5, numbers of axons decreased per TEM image results in reduced myelin density with increased axon diameter although myelin thickness slightly contributes positively to myelin density.In summary, we have performed an extremely strenuous and challenging TEM and MRI comparison study of entire CC of normal rat brain at high spatial resolution (100×100 um² in-plane resolution). T₁/R₁ images are highly correlated to myelin density and provide a good indicator of the density of myelin in vivo. The parametric T₁/R₁ images should provide useful information of the microstructure property of the tissue.