To the Editor.
—Schwartz et al1 provided an excellent discussion of the utility of positron emission tomography in initial efforts to describe subtypes of depressive disorders. They present as a focus for their discussion the unipolar/ bipolar dichotomy. This is, after all, the best-defined and possibly most important one clinically. Schwartz and colleagues report that the18F-fluorodeoxyglucose positron emission tomographic scans of bipolar and unipolar subjects differ fundamentally. The authors graphically illustrate that bipolar depressed and bipolar mixed subjects exhibit significantly lower metabolic rates of glucose consumption (in milligrams of glucose per 100 g of tissue per minute) compared with unipolar depressed and normal subjects. Further, bipolar depressives exhibit an increase in global glucose consumption on clinical recovery (ie, on passing from the depressed to the euthymic state). In contrast, patients with unipolar depression do not display this dramatic change. Indeed, unipolar subjects do not differ from normal