The relationship between neuroscience and psychology

There are a certain amount of heated arguments within the behavioral sciences about the most appropriate way to study the question of behavior. Cellular and systems neuroscientists tend to have no use for psychological methods, finding them inefficient and messy (when I interviewed for graduate school, one monkey physiologist told me that my research interests were a waste of time that would lead to nothing. Monkey physiology, on the other hand...). People who do more cognitive work often feel that while neuroscience is more exact and perhaps makes more concrete progress, it's progress in the wrong direction.

I recently came across an excellent and succinct explanation of why both methods are necessary:

Experimental psychology on both human subjects and animals is an essential part of the enterprise, for the obvious reason that accurate characterizations of psychological phenomena are necessary to guide the search for explanations and mechanisms. Trying to find a mechanism when the phenomenon is misdescribed or underdescribed is likely to be quixotic. Neurology is an essential part of the enterprise because it provides both important behavioral data on human subjects and hypothesizes connections between specific brain structures and behavior. Neuroscience is essential both to discover the functional capacities of neural components and because reverse engineering is an important strategy for figuring out how a novel device works.

Churchland & Sejnowski. (1991) "Perspectives on cognitive neuroscience" in Lister & Weingarter, Perspectives on Cognitive Neuroscience, pp. 3-23.