(A) Timeline for injections of Poly I:C, treatment with PLX5622 (PLX; depletes microglia) or control chow and behavior approach. (B) Timeline for injections of 4-hydroxytamoxifen (4-OHT) or vehicle and Poly I:C, treatment with PLX, and behavior procedures. (C) Male and female WT/Poly I:C mice after PLX (n = 6; P < 0.0001, t = 14.66), WT/Poly I:C mice after control chow (n = 8; P < 0.0001, t = 9.04), and Cx3cr1Cre/Tsc2Flox (Tsc2+/−) Poly I:C mice after control chow (n = 6; P < 0.0001, t = 17.66), but not Cx3cr1Cre/Tsc2Flox (Tsc2+/−) Poly I:C mice after PLX (n = 7; P = 0.19, t = 1.37), show normal social memory. Two-way ANOVA analyses using time spent exploring the novel mouse as the dependent variable revealed a significant effect of treatment (F(1,23) = 9.63, P = 0.005), and Sidak post hoc test revealed a significant difference between Cx3cr1Cre/Tsc2Flox (Tsc2+/−)/Poly I:C mice after PLX and Cx3cr1Cre/Tsc2Flox (Tsc2+/−)/Poly I:C mice after control chow (P < 0.01, t = 3.42). (D) Male and female Cx3cr1CreER/Tsc2Flox/vehicle mice injected with Poly I:C after PLX (n = 9; P < 0.0001, t = 10.38), but not Cx3cr1CreER/Tsc2Flox/4-OHT mice injected with Poly I:C after PLX (n = 8; P = 0.40, t = 0.84), show normal social memory. Data represent means ± SEM as well as values for individual mice. As indicated, in (C), Tsc2+/− represents Cx3cr1Cre-Tsc2Flox. In (D), Tsc2+/− represents Cx3cr1CreER/Tsc2Flox.