Fig. 1. Experimental design, physiological, and behavioral results.

A Participants were tested on two experimental days: day 1, stimulus encoding and pharmacological manipulation of post-encoding noradrenergic activity and day 2, memory recognition. Both encoding and test took place in the MRI scanner. Critically, to investigate time-dependent consolidation processes, the memory test took place either 1d or 28d after encoding. The image of the playground is licensed under Creative Commons License; courtesy of Tomasz Sienicki (https://commons.wikimedia.org/wiki/File:Playground_29_ubt.JPG; image unchanged). B Effective manipulation of noradrenergic arousal after encoding: While groups did not differ at baseline (all p > 0.242, two-tailed Welch’s t-tests) or shortly after encoding (all p > 0.111, two-tailed Welch’s t-tests), participants of the yohimbine (YOH) group had significantly higher systolic (all p < 0.005, two-tailed Welch’s t-tests) and diastolic (all p < 0.005, two-tailed Welch’s t-tests) blood pressure from 85 min after drug intake until the end of experimental day 1 (drug × time: all p < 0.001, mixed ANOVAs). C A generalized linear mixed model (LMM) with the between-factors drug and delay and the within-factor emotion revealed no group-difference in immediate free recall performance on day 1, suggesting that encoding was comparable in the four groups. However, while memory performance significantly decreased from 1d to 28d after encoding (main effect delay: β = −1.12, p < 0.001, LMM), post-encoding noradrenergic arousal reduced this time-dependent memory decline (drug × delay: β = 0.64, p = 0.029, LMM): The YOH group showed a significantly smaller decrease in memory performance from 1d to 28d than the placebo (PLAC) group. All n = 104 participants. Bars represent mean ± SEM. Source data are provided as Source data file. **p < 0.010; ***p < 0.001.