Estrogens are potent neuroprotectants both in vitro and in vivo. In the present study, we compared the potency and efficacy of a non-feminizing estrogen, 2-(1-adamantyl)-4-methylestrone (ZYC-26), with its parent estrogen, estrone, and an expected non-neuroprotective 3-O-methyl analog of (17beta)-2-(1-adamantyl)estradiol (ZYC-23). These estratriene derivatives were tested for their ability to protect in an in vitro lipid peroxidation model, to neuroprotect against oxidative stress in cell culture models, to bind the estrogen receptors (ERalpha and ERbeta), to elicit uterotrophic effects, and to affect brain damage from transient middle cerebral artery occlusion. We observed that in contrast to estrone, neither ZYC-26 nor ZYC-23 bound to either estrogen receptors (ER) and both failed to elicit a uterotrophic response. In vitro, the active estrogen analogue ZYC-26 was more potent that estrogen in its ability to inhibit lipid peroxidation and to protect HT-22 cells from either glutamate or iodoacetic acid (IAA) toxicity. Further, ZYC-26 was as active in preventing brain damage from transient middle cerebral artery occlusion (MCAO) as was estrone. Collectively, these studies suggest that the antioxidant activity, rather than ER binding of non-feminizing estrogens such as ZYC-26, mediates their potent neuroprotective activity. Further, in view of the now known toxicities of chronic feminizing estrogen use in older women, non-feminizing estrogens may be a useful alternative for estrogen-induced brain protection.