Stress-induced vagal activity influences anxiety-relevant prefrontal and amygdala neuronal oscillations in male mice

Abstract

The vagus nerve crucially affects emotions and psychiatric disorders. However, the detailed neurophysiological dynamics of the vagus nerve in response to emotions and its associated pathological changes remain unclear. In this study, we demonstrated that the spike rates of the cervical vagus nerve change depending on anxiety behavior in an elevated plus maze test, and these changes were eradicated in stress-susceptible male mice. Furthermore, instantaneous spike rates of the vagus nerve were negatively and positively correlated with the power of 2-4 Hz and 20-30 Hz oscillations, respectively, in the prefrontal cortex and amygdala. The oscillations also underwent dynamic changes depending on the behavioral state in the elevated plus maze, and these changes were no longer observed in stress-susceptible and vagotomized mice. Chronic vagus nerve stimulation restored behavior-relevant neuronal oscillations with the recovery of altered behavioral states in stress-susceptible mice. These results suggested that physiological vagal-brain communication underlies anxiety and mood disorders.

Fig. 1. SD stress attenuates VN spikes.

a Experimental timeline. b SI ratios (n = 36 mice) and representative trajectories in a SI test. c SI ratios at two time points (n = 15 mice). R = 0.66, P = 0.011, two-sided Pearson’s Correlation. d, e Schematic illustration of VN recordings with a cuff-shaped electrode. Similar electrodes used for 45 mice. f A filtered VN trace, its power, and EMG power during a rest period. The orange region is magnified (bottom; dots represent spikes). g VN spike power and rates during quiescent periods (n = 9, 12, and 12 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. power: naïve vs susceptible, *P = 0.012, resilient vs susceptible, *P = 0.0044; rate: naïve vs susceptible, *P = 3.9 × 10⁻⁴, resilient vs susceptible, *P = 0.0040, two-sided Tukey’s test. h The percentages of open arms (n = 18, 16, and 20 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. i (Left) Moving speed. (Right) Same as h but for the percentage of move states in closed arms. naïve vs susceptible, *P = 0.0012, resilient vs susceptible, *P = 0.0029, two-sided Tukey’s test. j VN spike power in the arms (n = 11, 12, and 14 mice). Data are presented as mean ± SEM. naïve: open vs closed move, t₁₀ = 3.10, *P = 0.034; open vs closed stop, t₁₀ = 4.16, *P = 0.0058; closed move vs closed stop, t₁₀ = 2.98, *P = 0.041; resilient: open vs closed move, t₁₁ = 3.85, *P = 0.0081; open vs closed stop, t₁₁ = 5.74, *P = 3.9 × 10⁻⁴; closed move vs closed stop, t₁₁ = 3.87, *P = 0.0078, two-sided paired t-test followed by Bonferroni correction. k Across-group comparisons. *P = 0.035, two-sided Tukey’s test. Source data are provided as a Source Data file.

Fig. 2. VN activity and PFC-AMY oscillations.

a Images of c-Fos expressions in the NTS, PFC, and AMY, after VNS (left) and the percentages of c-Fos-positive neurons (right; n = 7 and 7 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. NTS: t₁₂ = 14.86, *P = 4.3 × 10⁻⁹; PFC: t₁₂ = 3.53, *P = 0.0042; AMY: t₁₂ = 0.98, P = 0.35, two-sided Student’s t-test from bootstrapped datasets. b Histological confirmation of electrodes. Similar recordings were obtained from the PFC and AMY in 63 and 51 mice, respectively. c A PFC LFP trace, its wavelet spectrum, an AMY LFP trace, and a VN trace. d Correlations between VN power and PFC or AMY power against each frequency band in naïve mice (n = 9 and 9 mice). Shaded areas represent SEM. Black bars represent 2–4 Hz and 20–30 Hz bands. e Changes in VN power and PFC power. f, g PFC neuronal firing with 2–4 Hz and 20–30 Hz oscillations. h MVLs of PFC neurons for 2–4 Hz and 20–30 Hz oscillations against their firing rates (n = 47 neurons from 6 mice). Orange dots show significance. i PFC 2–4 Hz and 20–30 Hz power on the EPM (n = 16 mice). Data are presented as mean ± SEM. 2–4 Hz: open vs closed move, t₁₅ = 2.87, *P = 0.035; open vs closed stop, t₁₅ = 3.56, *P = 0.0086; 20–30 Hz: open vs closed move, t₁₅ = 3.89, *P = 0.0044; open vs closed stop, t₁₅ = 6.24, *P = 4.7 × 10⁻⁵; closed move vs closed stop, t₁₅ = 6.07, *P = 6.4 × 10⁻⁵; two-sided paired t-test followed by Bonferroni correction. j Same as i but for AMY (n = 9 mice). open vs closed move, t₈ = 3.73, *P = 0.017; open vs closed stop, t₈ = 6.32, *P = 6.8 × 10⁻⁴; closed move vs closed stop, t₈ = 4.13, *P = 0.0099; two-sided paired t-test followed by Bonferroni correction. Source data are provided as a Source Data file.

Fig. 3. Disruption of anxiety-related PFC-AMY LFP patterns in stress-susceptible mice.

a Comparisons of PFC 2–4 Hz and 20–30 Hz power among the open arms, move states in the closed arms, and stop states in the closed arms in stress-resilient mice (n = 10 mice). Each thin line represents each mouse. Data are presented as mean ± SEM. 20–30 Hz: open vs closed stop, t₉ = 5.29, *P = 0.0015; closed move vs closed stop, t₉ = 6.00, *P = 6.0 × 10⁻⁴, two-sided paired t-test followed by Bonferroni correction. b Same as a but for AMY power (n = 11 mice). 20–30 Hz: open vs closed stop, t₁₀ = 4.73, *P = 0.0024; closed move vs closed stop, t₁₀ = 4.95, *P = 0.0017; two-sided paired t-test followed by Bonferroni correction. c, d Same as a, b but for stress-susceptible mice (n = 15 and 10 mice). e Across-group comparisons of PFC power shown in Figs. 2i and 3a, c. 20–30 Hz: open: naïve vs susceptible, *P = 0.048, resilient vs susceptible, *P = 0.011; closed stop: naïve vs susceptible, *P = 0.016, resilient vs susceptible, *P = 0.042, two-sided Tukey’s test. f Across-group comparisons of the datasets of AMY power shown in Figs. 2j and 3b, d. Source data are provided as a Source Data file.

Fig. 4. Vagotomy disrupts anxiety-related PFC-AMY LFP patterns.

a LFP signals were recorded from mice with vagotomy. b The percentage of time spent in the open arms (left; t₂₇ = 1.68, P = 0.10, two-sided Student’s t-test) and the percentage of move states in the closed arms (right; t₂₇ = 3.20, *P = 0.0035, two-sided Student’s t-test) to the total recording time in vagotomized mice (n = 11 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. c Comparisons of PFC 2–4 Hz and 20–30 Hz power among the open arms, move states in the closed arms, and stop states in the closed arms in vagotomized mice (n = 11 mice). Each thin line represents each mouse. Data are presented as mean ± SEM. d Same as c but for AMY power (n = 11 mice). Source data are provided as a Source Data file.

Fig. 5. VNS restores stress-induced PFC-AMY oscillations.

a VNS was applied to stress-susceptible mice. b SI ratios (n = 18 and 10 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. Z = 4.18, *P = 3.0 × 10⁻⁵, two-sided Mann–Whiney U test. c The percentage of the open arms (left; t₃₇ = 0.043, P = 0.97, two-sided Student’s t-test) and the move states in the closed arms (right; t₃₇ = 3.04, *P = 0.0044, two-sided Student’s t-test) (n = 18 mice). Box plots show center line as median, box limits as the 25th percentile and the 75th percentile, whiskers as minimum to maximum values that are not outliers. d VN spike power in the susceptible + VNS group (n = 7 mice). Data are presented as mean ± SEM. open vs closed stop, t₆ = 5.88, *P = 0.0032; closed move vs closed stop, t₆ = 7.72, *P = 7.4 × 10⁻⁴; two-sided paired t-test followed by Bonferroni correction. e Across-group comparisons of VN spike power. open: t₁₉ = 4.19, *P = 5.0 × 10⁻⁴: closed stop: t₁₉ = 2.61, *P = 0.017, two-sided Student’s t-test. f PFC power (n = 11 mice). Each thin line represents each mouse. Data are presented as mean ± SEM. 20–30 Hz: open vs closed move, t₁₀ = 5.12, *P = 0.0013; open vs closed stop, t₁₀ = 6.09, *P = 0.00035; closed move vs closed stop, t₁₀ = 4.76, *P = 0.0023, two-sided paired t-test followed by Bonferroni correction. g Same as f but for AMY power (n = 10 mice). 20–30 Hz: open vs closed stop, t₉ = 7.87, *P = 7.6 × 10⁻⁵; closed move vs closed stop, t₉ = 5.26, *P = 0.0016; two-sided paired t-test followed by Bonferroni correction. h, i Across-group comparisons of PFC and AMY 20–30 Hz power. PFC: open: t₂₄ = 3.47, *P = 0.020: closed stop: t₂₄ = 2.18, *P = 0.039; AMY: open: t₁₈ = 3.08, *P = 0.0065, two-sided Student’s t-test. Source data are provided as a Source Data file.