Diffusion MRI is a sensitive tool to detect tissue microstructural organization in the brain under normal and pathological conditions. With oscillating gradient diffusion MRI, it is now possible to sample a wide range of the temporal diffusion spectrum to gather a wealth of information on tissue microstructural organization^1,2,3. In particular, the dispersion profile of the temporal diffusion spectrum⁴, which can be approximated with the rate of frequency dependent change for the low to moderate frequency range (e.g., < 300 Hz), has been linked to cell size and surface volume ratio^5,6. In this study, we investigated the dispersion profiles along different directions in the mouse brain and explore the potential relationships between the observed directional anisotropy with underlying microstructural organization.Neonatal C57BL/6 mice (n=6) at postnatal day 10 were subjected to unilateral HI using the Vannucci model⁷, and sacrificed at 24hrs after injury and perfusion fixed with 4% PFA. Ex vivo diffusion MRI experiments were performed on an 11.7 T spectrometer with a 10 mm transceiver volume coil. Pulsed gradient spin-echo (PGSE) data were acquired with δ/Δ=2.6/20ms, and oscillating gradient spin-echo (OGSE) data were acquired at frequencies (f) of 50 Hz, 100 Hz, and 200 Hz. All data were acquired with TE/TR = 60/1000 ms, NA=2, 30 diffusion directions, b = 1000 s/mm², and resolution of 0.125 mm isotropic. Diffusion tensors were estimated from diffusion MRI data at each frequency and the datasets were co-registered with affined registration based on the b₀ images. Along each diffusion-encoding direction, the rate of apparent diffusion coefficient (ADC) increase with respect to the square root of frequency was estimated, and the results along 30 directions were fitted to a tensor model, which called the temporal diffusion dispersion (TDD) tensor here. The scalar indices such as apparent diffusion coefficient (ADC), fractional anisotropy (FA), and colormap were calculated from the TDD tensor in the same way as conventional diffusion tensor.Fig. 1 showed the rates of ADC change with the square root of frequency measured in several gray and white matter structures in the contralateral side of the mouse brain where the tissues were intact (n=6), and the change rates demonstrated directional anisotropy along x/y/z over the 0 (PGSE) to 200 Hz range (Fig. 1D). The trace map of TDD tensor highlighted regions with high density of neurons in the mouse brain, as reported before⁸. The FA map of TDD tensor highlighted several regions with distinct microstructural organization (Fig. 2). For example, several large white matter structures, e.g., the cerebral peduncle, had high diffusion anisotropy in both conventional PGSE/OGSE as well as in TDD (Table1), with the primary eigen-vector of the TDD tensor mostly parallel to that of PGSE tensor. Interesting, the splenium of the corpus callosum showed a relatively low FA, whereas as the region below the corpus callosum—the dorsal hippocampal commissure (dhc) showed high FA values from the TDD tensor compared to the PGSE FA values (Table 1). Fiber orientation density (FOD) reconstructed using constraint spherical deconvolution⁹ based on the PGSE data showed that the dhc contained crossing fibers with two fiber group along left-right and anterior-posterior directions (orange arrows in Fig. 2). While the PGSE-colormap showed an AP-oriented eigen-vector (blue) in the dhc, the TDD tensor indicated a LR-oriented eigen-vector (red) in this region. In the hippocampus, a layer (yellow arrows in Fig. 2) with low PGSE FA values showed high FA values from TDD tensor (Table1). This layer is approximately in the location or stratum oriens, in which several groups of axons and dendrites with different directions mixed together. In the ipsilateral side of the hypoxia-ischemia injured mouse brains, the well-organized hippocampal layers were disrupted (Fig. 3). The hippocampal injury was marked by reduction of trace and FA from the TDD tensor.It is known that temporal diffusion spectrum provides information on the spatial scale of tissue microstructural organization, such as cell size and surface-to-volume ratio, etc^5,6. It is also well known that, in several brain regions, the spatial organizations of microstructures are different along different directions, which may lead to the high anisotropy in the TDD tensor. When these unique microstructural organizations were disrupted by injuries, e.g., hypoxic ischemic injury in this study, both the conventional tensor and TTD tensor showed changes in anisotropy. In summary, the new method reported here revealed an interesting new contrast, which potentially offers additional information about the microstructural organization.