Testosterone (T) may enhance cognitive performance. However, its mechanisms are not well understood. First, we hypothesized that if T's effects are mediated in part through actions of its 5α-reduced metabolites, dihydrotestosterone (DHT) and/or 3α-androstanediol (3α-diol) in the hippocampus, then T, DHT, and 3α-diol-administration directly to the hippocampus should enhance learning and memory in the inhibitory avoidance task. In order to test this hypothesis, gonadectomized (GDX) male rats were administered T, DHT, or 3α-diol via intrahippocampal inserts immediately following training in the inhibitory avoidance task. We found that T tended to increase, and DHT and 3α-diol significantly increased, performance in the inhibitory avoidance task compared to vehicle-administered GDX rats. Second, we hypothesized that, if androgens' effects are due in part to actions of 3α-diol in the hippocampus, then systemic or intrahippocampal administration of 3α-diol should significantly enhance cognitive performance of GDX male rats. Third, we hypothesized that, if androgen metabolites can have actions at estrogen receptors (ERs) in the hippocampus, then administration of ER antisense oligonucleotides (AS-ODNs) directly to the hippocampus of GDX, 3α-diol replaced, rats would decrease learning in the inhibitory avoidance task. We found that intrahippocampal administration of AS-ODNs for ERβ, but not ERα, significantly decreased learning and memory of 3α-diol replaced rats. Together, these findings suggest that T's effects to enhance learning and memory may take place, in part, through actions of its metabolite, 3α-diol, at ERβ in the dorsal hippocampus.