To the Editor: We appreciate the comments of Drs. de Haan and van Amelsvoort regarding our recent article and are happy to expand on speculation regarding putative mechanisms underlying the observed treatment effects. First, we should highlight that our comments regarding the doses used in our study referred to clinical practice in the management of patients with suboptimal treatment response—not to putative levels of receptor occupancy—and readers should recognize the unique challenges inherent in treating this patient population.

Second, we believe that the contention of Drs. de Haan and van Amelsvoort that higher D₂ receptor occupancy limited cognitive improvement is not supported by our data. Several facts argue against this interpretation. The mean dose levels (milligrams per day) achieved in the second period of the study (last observation carried forward) were 526.6 (SD=140.3) for clozapine, 30.4 (SD=6.6) for olanzapine, 11.6 (SD=3.2) for risperidone, and 25.7 (SD=5.7) for haloperidol, as described elsewhere (1). We thus believe that at follow-up all patients in the olanzapine, risperidone, and haloperidol groups likely had levels of D₂ receptor occupancy well in excess of 80% and within the range at which D₂ occupancy asymptotes (2). Therefore, the possible discrepancies between groups in D₂ occupancy were probably subtler than suggested by Drs. de Haan and van Amelsvoort.

It is still conceivable that even subtle differences in overall D₂ occupancy or other differences, such as rates of D₂ receptor dissociation (2, 3), may help explain the differences in neurocognitive function observed in our study. If the differences between treatments were mediated primarily by differences in D₂ receptor occupancy, changes in motor function (the strongest correlates of D₂ receptor availability in the striatum [Volkow et al., 1998]) should have differed between groups, but this was not the case. Moreover, if neurocognitive improvement had been limited by high D₂ occupancy, the clozapine-treated patients might have been predicted to show the greatest improvement, but they did not (despite good clinical efficacy).

We examined the effects of all antipsychotic doses and anticholinergic treatment effects (using benztropine doses and extrapyramidal symptom ratings) and found that these variables did not substantially influence our findings. In summary, we believe that our data do not support a simple explanation of neurocognitive improvement in terms of lower D₂ receptor occupancy, and thus we highlighted in our Discussion not only the possible D₂ receptor effects but also a broad range of other possible mechanisms that might be responsible, including serotonergic, adrenergic, cholinergic, histaminergic, and non-D₂ dopaminergic receptor effects, that may have differed among treatments.

We believe further that our findings agree well with a larger literature, suggesting that attempts to explain complex clinical effects in terms of receptor binding profiles have been challenging at best (4). We agree that it would be valuable to conduct research comparing the neurocognitive effects of antipsychotic treatments, specifically targeting levels of D₂ receptor occupancy, as an experimental manipulation. It is important to recognize, however, that any strategy of reducing D₂ receptor occupancy for cognitive augmentation might best target patients with a history of treatment response more favorable than our patients showed to minimize the possible clinical risks of relapse.

Reprints are not available; however, Letters to the Editor can be downloaded at http://ajp.psychiatryonline.org.