MR functional imaging, due to the improvement in ultra-speed imaging technology such as echo-planar imaging, has become a very powerful technique since the beginning of the nineties. This imaging technique is divided into diffusion imaging, perfusion imaging and cerebral activation. Diffusion imaging probes the mobility of water molecules characterized by a diffusion coefficient called the apparent diffusion coefficient (ADC) for biological tissues. Perfusion imaging gives hemodynamic information due to the regional cerebral blood volume by the use of contrast agents such as chelates of gadolinium carrying strong magnetic susceptibility. Both imaging techniques can provide information in a wide nosological range : cerebral ischemia, in the acute phase and in case of intracranial tumors, contributing to tumoral grading, localizing the site of biopsy, and assessing response to therapy (after radiotherapy for example). Nevertheless, a wide range of domains remains incompletely studied, for example cerebral white matter diseases and neurodegenerative diseases. For clinical applications, a precise knowledge of the potentials of both techniques and their limitations is needed. Limitations result from the large number of often patient-related parameters, imaging technique (perfusion) and data analysis. Powerful software has been developed in the workstation environment. Thus this imaging technique requires up-to-date equipment and close collaboration between clinical and research teams for optimal efficiency.