Abstract： Objective To investigate diffuse kurtosis imaging (DKI) in the description of cerebral white and grey matter and application for diagnosis of stroke.Methods The profiles of cerebral magnetic resonance imaging (MRI) in 2 normal controls and 15 patients with stroke (9 with cerebral infarction,including 3 respective cases of acute,subacute and old infarction) were analyzed. Diffusion coefficient (D value) and diffusion kurtosis (K value) of each signal were obtained via nonlinear fitting corresponding to the location of individual pixel in 9 diffusion gradient fields for depiction of ADC and DKI maps.The images of ADC and DKI maps were studied and compared.Results DKI map contained more texture details and abundant tissue characters.Acute and subacute cerebral infarction presented hyperintensity signals in DKI,while old infarction presented hypointensity signals,with mean relative MK value (rMK) of (2.193±0.166),(1.789 ± 0.162) and (0.564± 0.069),respectively.Conclusion DKI offers more information of brain tissue,particularly in terms of grey matter and stroke foci,as compared with ADC that mainly focuses in tissues.Features of variation in rMK value of cerebral infarction can be captured.