Figure 6.

Withdrawal from CIE produces opposing alterations to the excitation/inhibition balance in the dmPFC and vmPFC BLA pathways. A₁, Relative to AIR controls (N = 6), CIE/WD (N = 9) significantly increased the light-evoked E/I ratio recorded from dmPFC-BLA terminals. *p < 0.05, unpaired t test. A₂, CIE/WD significantly decreased peak IPSC amplitudes relative to AIR controls in responses recorded from dmPFC–BLA synapses (see text; *p < 0.05, Bonferroni post-test following two-way ANOVA, significant interaction). A₃, Response onset latency for EPSCs and IPSCs recorded from dmPFC–BLA inputs. ISPCs had significantly longer latencies than EPSCs regardless of exposure condition (***p < 0.001, main effect of response type, two-way ANOVA). A₄, Representative traces recorded from dmPFC-BLA terminals in the two treatment groups. IPSCs recorded at 0 mV and EPSCs recorded at −70 mV. B₁, Light-evoked E/I ratios recorded from vmPFC-BLA terminals. Note the significant decrease (**p < 0.01, unpaired t test) in the E/I ratio in CIE/WD (N = 13) vs AIR (N = 13). B₂, Peak amplitudes of EPSCs and IPSCs recorded from vmPFC–BLA synapses. Significant interaction between treatment group and response type (see text; two-way ANOVA). Note the significant increase in the IPSC amplitude in CIE/WD neurons compared with AIR controls (***p < 0.001, Bonferroni’s multiple-comparisons post-test). B₃, Response onset latency for EPSCs and IPSCs recorded from vmPFC–BLA inputs. ISPCs had significantly longer latencies than EPSCs regardless of exposure condition (see text; two-way ANOVA, ***p < 0.001, main effect of response type). B₄, Representative traces recorded from vmPFC-BLA terminals. EPSC amplitudes were normalized across treatments to emphasize changes to the IPSC amplitude. Calibration: x = 40 ms, y = 50 pA. Dashed lines denote normalized oEPSC amplitudes across the treatment groups to emphasize the treatment effects on the IPSCs.