This paper aims to investigate whether and how cognitive neuroscience can provide theories, methods and useful data to psychiatry. In the first section the project of cognitive neuropsychiatry is discussed in the context of the changes occurred in the relationship between psychological and neurological sciences in the last thirty years. In the second section the multilevel explanatory style of cognitive neuropsychiatry is examined, arguing that it carves a middle way in between the reductionist excessive insistence on ontological parsimony and unification of science, on the one hand, and the antireductionist claim for the strong autonomy of special sciences, on the other. The third section is dedicated to the analysis of an exemplar of mechanistic explanation in cognitive neuropsychiatry: the attempt to develop a neurocognitive account of Capgras delusion.
Keywords: Cognitive neuroscience, neuropsychiatry, mechanism, explanatory levels, delusion