Abstract
Estradiol (E2) positively influences memory facilitation effects in older women and rodent models by targeting key memory-related brain regions. However, the impacts of E2 on emotional memory processes in younger women are less clear. As women are twice as likely as men to develop trauma related disorders such as posttraumatic stress disorder (PTSD), it is important to understand how hormones like E2 might impact threat memory mechanisms. Using a randomized, double-blinded, cross-over design, we administered transdermal E2 or placebo to 45 naturally-cycling, Black women (18-35 years) with a range of trauma-related symptoms during the early luteal (low endogenous E2) phase of their cycles. The following day, participants underwent a categorical threat conditioning paradigm during fMRI recording and completed a post-scan recognition test of images seen during the scanning session. The next month, participants repeated experimental procedures under the opposite patch condition. Blood samples taken day of scan showed a mean 80 pg/mL increase in serum E2 levels under E2 supplementation. While all participants showed an enhancement of threat on memory, such that threat-associated (CS+) images were later recognized better than neutral (CS-) pictures, neither E2 patch nor PTSD symptom severity predicted recognition performance. However, under placebo, greater bilateral entorhinal cortex (ERC) response during threat vs safety learning (CS+>CS-) was associated with greater post-scan recognition for CS+ compared to the CS- category, indicating greater ERC facilitation of episodic encoding and threat bias in the low E2 condition. We also found that the combination of high E2 and progesterone (P4) was associated with reduced ERC CS+ >CS- activity, potentially explaining why E2 supplementation did not facilitate CS+ >CS- recognition and suggesting an antagonistic role for P4 with E2 in memory facilitation. E2 produced an increase in ERC functional connectivity to the superior temporal gyrus during CS- encoding, which may suggest a shift in ERC engagement away from episodic encoding. These findings indicate that the post-ovulation drop in E2 and potential interactions with P4 facilitate the episodic encoding of safety and threat cues in women, given that exogenous E2 blocked these effects. This study provides novel causal evidence on the role of cyclical fluctuation in E2 in determining episodic components of memory for learned threat and safety.