Results showed that fear on the retrieval test was significantly lower in mice treated with any of the three doses of AM3506, relative to vehicle (ANOVA effect of treatment: F3,39 = 3.76, P<0.05, followed by post hoc tests, n = 10–11) (Figure 1c). This is consistent with significant facilitation of extinction by AM3506 treatment. It is notable that fear levels at the beginning of extinction were not affected by AM3506 treatment. This in line with prior studies showing that CB1R gene knockout or antagonism impairs fear extinction and adaptation without affecting fear expression.3,49 Interestingly, we also found that the long-term extinction-facilitating effects of AM3506 were independent of any demonstrable within-session extinction (Supplementary Figure S2a). This dissociation echoes the finding that pre-extinction treatment of S1 mice with the clinically efficacious anxiolytic, fluoxetine, also produces marked reductions in fear during retrieval without concomitant within-session reductions.25 It is also in line with the earlier observation that manipulations which facilitate long-term extinction in CB1R knockout mice (for example, spaced training) do not produce within-session extinction.50 On the other hand, there is evidence that nonspecific
