Diffusion kurtosis imaging (DKI) was suggested to describe the exact non-Gaussian diffusion in tissue and also provided accurate and abundant information of the microstructure^1-3. This paper is to investigate the critical b-value in the acquisition and reveal the influence on the kurtosis representation of white matter microstructure with diffusion kurtosis imaging (DKI). Nine normal healthy volunteers (4 male, 4 female) aging from 23 to 26 participated and were scanned in a Siemens 3.0 T MRI Scanner System using a 32-channel head coil. A six b-values DWI acquisition including 0, 500, 1000, 1500, 2000, 2500 sec/mm² with 30 diffusion encoded orientations was conducted on each individal. Other acquisition parameters: TR=6800ms, TE=101ms, voxel size=2.0×2.0×2.0mm³, FOV=256×256 mm², metrics=128×128, number of sagittal slices=52 with no gap. And a high resolution 3D T1-weighted MPRAGE was acquired with the following imaging parameters: TR = 1900 ms, TE = 2.52 ms, matrix = 240×256, FOV = 240×256 mm², and slice thickness = 1 mm. After image quality assessment, one participate was excluded due to poor quality of imaging data. After eddy current and head motion correction were finished in FSL sofrware, Diffusional Kurtosis Estimator (DKE) was ultilized to calculated kurtosis indexes radial kurtosis(RK), axial kurtosis(AK) and mean kurtosis(MK) from the three combinations of b-value (BL: 0,1000 sec/mm²,2000 sec/mm²; BH:0,1000sec/mm²,2500sec/mm² and TA: 0, 500, 1000, 1500, 2000, 2500 sec/mm²) DWIs data. Tract-based spatial statistics analysis for whole brain was firstly performed on RK and AK under significance of FDR corrected p<0.05 in FSL software. Then the 3D geometrical shape analysis of kurtosis for corpus callosum were performed, in which the corpus callosum was divided into four parts. For each part, the averaged 3D spatial kurtosis distribution was reconstructed and graphied for detailed microstructure. Visually from Fig.1, the post cingulate white matter and splenium of corpus callosum revealed differences in two kurtosis indexes, and more significant differences were observed in AK map than MK maps, especially in the corpus callosum, posterior cingulate white matter and bilateral posterior thalamic radiation (especially left) . Predominantly, two-paired t-test of the differences between BL and TA contributes more than that BH and TA. Note the MK statistic-analysis, in this case, almost none differences have been factored out. As diffusion tensor had little difference between these protocols, Fig.2 is the 3D surface of the kurtosis tensors’ cross section in CC over all orientations. All the blue lines directed the principal diffusion orientation which represented the fiber direction. Some detailed difference was obvious that although the orientation and radius profile all showed some characteristic, the Fig.2-EFGH was more similar to the Fig.2-IJKL, while the Fig.2-ABCD indicated higher variability compared to the other two lines in Fig.2. In summary, the radial kurtosis and axial kurtosis from different maximum b-value acquisitions lead a significant difference to kurtosis estimation in white matter mainly on corpus callosum and corona radiate. The specific microstructure and the geometric shape of white matter revealed specific sensitivity and variability with the varied b-values.Two diffusion weighted b-values acquisition was popular in DKI application since that one low b-value and one high b-value was demonstrated to be better fitted the diffusion decay⁴. Also 2000 sec/mm² and 2500 sec/mm² were mostly used in DKI research as maximum b-value^5,6. We compared BH and BL to explore the influence of this variation of maximum b-value on the kurtosis estimation. As an control experiment, TA including all b-values was compared with the considered variation of b-values' number and also this six b-value acquisition was popularly used^2,3,6. Based on this consideration, the results above might confirmed that the critical b-value was the maximum. The significant regions were located in the corpus callosum and corona radiata which reflected that this maximum b-value influence was associated with the microstruture. For the detailed graphic in Fig 2, radial profile in the fiber direction and angular profile in the myelin direction were showed difference caused by the different maximum b-value. This was agree with the previous research⁷ that the maximum b-value could not impact the fiber orientaion, but the fiber radius profile and the fiber longitudinal profile which was found in this paper. High b-value was highly sensitive to the non-Gaussion as well as the microstructure^8,9, and the relation between them can obtain a further insight into the non-Gaussion diffusion mechanism and the intrinsic principles of DKI^4-6 .What can be concluded that maximum b-value is a key factor that decides the kurtosis representation of white matter, and the proper value should vary from specific microstructure.