In this paper I trace Husserl's transformation of his notion of phantasy from its strong leanings towards empiricism into a transcendental phenomenology of imagination. Rejecting the view that this account is only more incompatible with contemporary neuroscientific research, I instead claim that the transcendental suspension of naturalistic (or scientific) pretensions precisely enables cooperation between the two distinct realms of phenomenology and science. In particular, a transcendental account of phantasy can disclose the specific accomplishments of imagination without prematurely deciding upon a particular scientific paradigm for its experimental investigation; a decision that is best left to the sciences themselves.