Fig. 4: CSDS impacts neural oscillation during social interaction.
A Example PSD curves from the Cg1 of Control (left) and CSDS (right) mice during social interaction test. Red lines indicate With CD1. Black lines indicate No CD1. B Mean PSD ratio curves from the Cg1 of Control and CSDS mice during social interaction. Data are mean ± s.e.m.. Orange lines indicate control group. Blue lines indicate CSDS group. C The comparison diagram of mean PSD ratio within Control and CSDS groups in diverse frequency bands. Two-Way repeated measures ANOVA with Šídák’s multiple comparisons test. The detailed statistic results were shown in Supplemental Table 1. D Pearson correlation analysis between mean PSD ratio and social index in high theta frequency band from Cg1. Pearson correlation analysis. r = −0.5318, *** P < 0.0001. E Pearson correlation analysis between mean PSD ratio and social index for eight brain regions in all frequency bands. The detailed correlation statistic results were shown in Supplemental Table 2. Black dotted line represents the significance level: 0.05. F Binary classification accuracy based on the oscillation features from individual brain region for Real and Shuffle data. Unpaired t-test. BLA: 72.62% ± 0.20%, t = 22.89, P < 0.001; NAc: 71.73% ± 0.18%, t = 21.33, P < 0.001; Cg1: 71.71% ± 0.17%, t = 21.81, P < 0.001; PrL: 70.28% ± 0.17%, t = 18.34, P < 0.001; LHb: 68.75% ± 0.21%, t = 20.28, P < 0.001; vHPC: 68.64% ± 0.19%, t = 22.65, P < 0.001; VTA: 68.59% ± 0.25%, t = 19.05, P < 0.001; IL: 65.59% ± 0.33%, t = 20.06, P < 0.001. G Binary classification accuracy based on the oscillation features from any two brain regions. Colors in bottom left indicate accuracies and some of them are significant different from their random. ***P < 0.001. H Binary classification accuracy based on the oscillation features from multiple brain regions (start from BLA, other brain regions were successively added as participants in order of individual brain area classification accuracy). An ordinary one-way ANOVA with Tukey’s multiple comparisons test. F(7,792) = 48.06, P < 0.001; BLA: 72.62% ± 0.20%, indicators number 2: 74.52% ± 0.35%, BLA vs 2: P = 0.0032; indicators number 3: 76.57% ± 0.36%, BLA vs 3: P < 0.001; indicators number 4: 77.03% ± 0.34%, BLA vs 4: P < 0.001; indicators number 5: 78.57% ± 0.35%, BLA vs 5: P < 0.001; indicators number 6: 78.64% ± 0.35%, BLA vs 6: P < 0.001; indicators number 7: 78.41% ± 0.39%, BLA vs 7: P < 0.001; indicators number 8: 80.18% ± 0.32%, BLA vs 8: P < 0.001. I The weight of individual brain region in Control/CSDS classification model. An ordinary one-way ANOVA with Tukey’s multiple comparisons test. F (7,792) = 12.93, P < 0.001. SBLA: −5.47% ± 0.17%, SVTA: −4.06% ± 0.31%, SBLA vs SLHb: P = 0.0096; SPrL: −3.43% ± 0.27%, SBLA vs SPrL: P < 0.001; SCg1: −3.38% ± 0.30%, SBLA vs SCg1: P < 0.001; SIL: −2.41% ± 0.28%, SBLA vs SIL: P < 0.001. C, E–I: All data were shown as mean±s.e.m. For BLA, Cg1, IL, LHb, NAc, PrL in Control: n = 30 mice; For vHPC, VTA in Control: n = 28 mice; For BLA, Cg1, IL, LHb, NAc, PrL in CSDS: n = 31 mice; For vHPC, VTA in CSDS: n = 28 mice. *P < 0.05, **P < 0.01, ***P < 0.001.
