The use of neuroimaging tools, especially functional magnetic resonance imaging, in nutritional research has increased substantially over the past 2 decades. Neuroimaging is a research tool with great potential impact on the field of nutrition, but to achieve that potential, appropriate use of techniques and interpretation of neuroimaging results is necessary. In this article, we present guidelines for good methodological practice in functional magnetic resonance imaging studies and flag specific limitations in the hope of helping researchers to make the most of neuroimaging tools and avoid potential pitfalls. We highlight specific considerations for food-related studies, such as how to adjust statistically for common confounders, like, for example, hunger state, menstrual phase, and BMI, as well as how to optimally match different types of food stimuli. Finally, we summarize current research needs and future directions, such as the use of prospective designs and more realistic paradigms for studying eating behavior.