Since the early 1970s, acupuncture has been the subject of multiple animal experiments and randomized clinical trials. Our understanding of acupuncture from both the clinical and mechanistic perspectives has, as a result, grown tremendously. Yet the final word on acupuncture as a therapy remains mixed, largely due to the contradictory nature of the evidence. With some exception, what clinical conditions would benefit and how acupuncture physiologically operates remains unclear. The impediment to progress is found in three disjunctions in acupuncture research: (1) the biomedical need to standardize treatments creates uncertainty about whether we are studying acupuncture appropriately; (2) the variability in acupuncture styles creates ambiguity about whether we are studying the right style; and (3) the discrepancy between animal and human studies creates questions about whether we truly understand the underlying mechanism responsible for acupuncture's therapeutic effect. We propose that these disjunctions are best addressed with the use of "manualized" protocols in clinical trials that are linked with mechanistic studies. Through this approach, we can create a healthy dialogue between the medical and acupuncture communities and recognize the unique physiologic properties that may be found in each acupuncture style. To illustrate how this proposal may fundamentally change acupuncture research, we present diabetic neuropathy as a particularly interesting model because of its complex heterogeneous pathophysiology.