We have reported that physostigmine, a reversible cholinesterase inhibitor, enhances verbal memory in patients with Alzheimer disease (AD). To elucidate the mechanism of cognition enhancement, plasma hormones were measured during high-dose acute and low-dose chronic steady-state intravenous infusions of physostigmine in nine subjects with AD. High-dose hormone responses were measured during and for 24 h after the infusion of physostigmine 1-1.5 mg over 45-60 min. Chronic responses were measured during continuous intravenous infusions of physostigmine at doses (0.5-25 mg/day) that escalated over 2 weeks, and then during 1 week infusion of the dose that optimized cognition (2-12 mg/day) or placebo administered in a randomized, double-blind, cross-over design. A replicable improvement in verbal memory was found in five subjects. High-dose physostigmine infusion that produced noxious side effects resulted in significant elevation above baseline in plasma levels of adrenocorticotrophic hormone (ACTH) (p = 0.0001), cortisol (p = 0.0001), and beta-endorphin (p = 0.0001). Chronic physostigmine administration, in the absence of adverse effects, produced no significant elevation in ACTH (p = 0.08), cortisol (p = 0.70), or beta-endorphin (p = 0.82). These results indicate that high-dose physostigmine activates the hypothalamic-pituitary-adrenal (HPA) axis, likely representing a "stress response." In contrast, cognition-enhancing doses do not produce a peripheral corticosteroid response. Thus, physostigmine-induced memory improvement is independent of the activation of the HPA axis.