The term “diffusion tensor imaging” (DTI) is used on many occasions to informally refer to anything related to diffusion-weighted imaging (DWI). From a formal point of view, however, DTI is the practice of fitting a tensor model to the DWI data. It is one of the simplest ways to model the DWI data that accounts, up to some extent, for the anisotropy in this kind of data. Exploiting this anisotropy is key to obtaining the characteristic directionally encoded color (DEC) maps and tractograms that are typically associated to the practice of DWI in general. Hence, it is not surprising many people use the term “DTI” in very different contexts. In this chapter, we aim to give the reader a feeling for what is really under the hood of the true art of DTI: obtaining these so-called diffusion tensors. What are they actually modeling? And, in this context, what is a tensor anyway? There’s a short and clear answer to this: the diffusion tensor describes the apparent diffusion coefficient (ADC), in function of direction. Hmm… “ADC” you say…?