We discuss the justification of Bickle's "ruthless" reductionism. Bickle intends to show that we know enough about neurons to draw conclusions about the "whole" brain and about the mind. However, his reductionism does not take into account the complexity of the nervous system and the fact that new properties emerge at each significant level of integration from the coupled functioning of elementary components. From a methodological point of view, we argue that neuronal and cognitive models have to exert a mutual constraint(MC) on each other. This approach would refuse to award any priority of cognitive approaches over neuroscience, and reciprocally, to refuse any priority of neuroscience over cognitive approaches. MC thus argues against radicalreductionism at the methodological level.