The chronotropic response to atropine is biphasic; low doses cause slowing of the sinus rate and high doses cause acceleration. Although it is accepted that atropine functions as a competitive antagonist at high doses, the mechanism of the negative chronotropic response at low doses is controversial. Specifically, it is unclear whether the effect is mediated centrally or peripherally. Since at the time of cardiac replacement all central nervous system connections to the heart are severed, the transplanted heart is a unique model for separating these effects. Graded doses of atropine sulfate (0.5, 1.0, 2.0, 4.0, 8.0 and 40.0 μg/kg body weight) were administered to 12 human heart transplant recipients to test the hypothesis that the bradycardiac effect of low dose atropine is centrally mediated. The baseline sinus cycle lengths of the decentralized donor and innervated native sinus nodes were 694 ± 20 and 733 ± 27 ms, respectively. At the 0.5 and 1.0 μg/kg doses, the cycle lengths of the native sinus node increased by 29.1 ± 13.5 and 23.1 ± 14.2 ms, respectively. At the 2.0 μg/kg dose the sinus cycle length again shortened to control. At the maximal dose of atropine the sinus cycle length shortened by 138.3 ± 29.7 ms compared with control. In contrast, the decentralized donor sinus node exhibited a flat dose response to atropine. High dose atropine (40 μg/kg) caused no change in the donor heart's atrial effective refractory period, corrected sinus node recovery time, or sinoatrial conduction time measured by either the Strauss or the Narula method. In summary, low dose atropine had no effect on the cycle length of the decentralized donor sinus node, but it lengthened the cycle length of the innervated native sinus node. High dose atropine similarly had no effect on the cycle length of the decentralized donor sinus node, but it shortened the cycle length of the innervated native sinus node. High dose atropine also had no effect on atrial refractoriness, sinus node recovery time or sinoatrial conduction time in the decentralized donor heart. Thus, the results suggest that in humans a central site of action is sufficient to explain the transient bradycardia caused by low dose atropine and that the effect depends on intact central neural connections. The data also confirm that high dose atropine has no direct or indirect effect on the human transplanted heart. Although atropine may have central and peripheral cardiac effects, central effects are probably of greater importance. © 1990.