Fig 1.
Physiological and behavioral evidence of stress.
(a) Corticosterone levels at different time points, before and after 30 min of restraint stress. Corticosterone levels increased following restraint stress (ANOVA F = 28.27, p = 6.4 × 10−14). Values represent mean ± SE. (b) Top: Corticosterone levels before, during 7 days of restraint stress, and 2 days after restraint stress was stopped. The measurements were pooled from different time points (30 to 120 min after restraint stress). Corticosterone levels increased during the week of repeated stress compared to baseline (ANOVA F = 7.7, p = 6.3 × 10−14). Bottom: Slight increase in pre-restraint corticosterone levels (before the mouse entered the 50 ml tube) as the stressor persists over time (ANOVA F = 3.04, p = 0.028). Values represent mean ± SE. (c) Movement velocity during 20-min open-field sessions in control (marked as C) and mice undergoing repeated stress (marked as S). Mice were tested on both the initial and seventh day of stress, while control mice followed the same protocol but without experiencing restraint stress (N = 5 and 4, accordingly). The mice exhibited reduced activity levels under repeated stress conditions. Notably, there was no discernible distinction in their behavior between the first and seventh day of testing, indicating a lack of adaptation over time (2-way ANOVA, condition F = 157.7, p = 2.1 × 10−08, session number F = 0.26, p = 0.61, interaction F = 1.49, p = 0.24). Values represent mean ± SE. (d) Left: Schematics of two-photon imaging during baseline and repetitive stress conditions. In repetitive stress sessions, the mice were placed in a 50 ml tube for 30 min to achieve mild stress. The imaging session started directly after the restraint. Individual cells were tracked over imaging days. Shown are examples of 2 imaging planes on day 1 and day 9 (scale bar, 50 μm) and the noise-evoked responses of 3 exemplar cells (mean ± SE). (e) Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 2.
Repetitive stress induces a decrease in sound-evoked activity.
(a) Left: noise-evoked activity rates at different noise intensities for chronically tracked PPys cells in baseline and repeated stress conditions (N = 5 mice, n = 285 neurons, mean ± SE). Activity rates decreased during repeated stress compared to baseline (2-way ANOVA, condition F = 185.6, p = 4.8 × 10−42, condition: intensity interaction F = 10.37, p = 9.3 × 10−21, nested ANOVA (mouse nested within session), condition F = 174, p = 1.5 × 10−39, condition: intensity interaction F = 12.7, p = 2 × 10−26, post hoc for each level baseline versus repetitive stress p < 0.01 for all levels above 50 dB, all Bonferroni corrected). Right: noise-evoked activity rates in 2 exemplar mice (mean ± SE). (b) Example of ΔF/F traces of 2 PPys tracked cells, recorded in response to noise presented at sound intensities ranging from 15 to 75 dB SPL in 2 baseline and 3 repeated stress sessions. Marked at time = 0 is the onset of the 100-ms white noise. (c) Top: Mean activity change, mean stress activity minus mean baseline activity, calculated per cell at different noise intensities. The reduction in activity was particularly striking at moderate sound intensities (mean ± SE, 1-way ANOVA, F = 14.24, p = 1.0 × 10−13 and t test for each level: 20 dB p = 8.4 × 10−05, 30 dB p = 1, 40 dB p = 0.350 dB p = 4.9 × 10−12, 60 dB p = 2.4 × 10−15, 70 dB p = 1.5 × 10−05 all corrected for multiple comparisons. Bottom: 50 dB noise evoked activity of 6 exemplar PPys cells during baseline and stressful states. (d) Noise-evoked activity rates at different noise intensities for chronically tracked PV cells in baseline and repeated stress conditions (N = 5 mice, n = 31 neurons, mean ± SE). Activity rates decreased during repeated stress compared to baseline (2-way ANOVA, condition F = 49.6, p = 2.6 × 10−12, condition: intensity interaction F = 1.94, p = 0.02, nested ANOVA (mouse nested within session), F = 56.5, p = 8.8 × 10−14, condition: intensity interaction F = 3.5, p = 3.5 × 10−05). (e) Example of ΔF/F traces of one tracked PV cell, recorded in response to noise presented at sound intensities ranging from 15 to 75 dB SPL in 2 baseline and 2 repeated stress sessions. Marked at time = 0 is the onset of the 100-ms white noise. (f) A second group of mice followed the same procedures but without experiencing daily restraint stress (N = 3 mice, n = 123 neurons). These mice exhibited a minimal change in noise-evoked PPy activity when comparing the first and second week of imaging (2-way ANOVA, F = 1.79, p = 0.11, post hoc baseline w1 50 dB: baseline w2 50 dB p = 1 Bonferroni corrected, nested ANOVA (mouse nested within session) F = 1.76, p = 0.12, mean ± SE). Right: mean PPy activity of 2 representative mice. Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 3.
Repetitive stress causes a reduction in neural contrast.
(a) Mean pre- and post-sound activity presented as deconvolved spikes. Pre- and post-sound activity increased during repeated stress (2-way ANOVA condition F = 413, p = 1.4 × 10−263, nested ANOVA (mouse nested within session) F = 668.5, p < 2.2 × 10−16, mean ± SE). (b) Example of deconvolved spike traces of a tracked PPy cell, recorded in response to noise presented at sound 50-55-60-65 dB SPL in baseline and stress sessions. Marked at time = 0 is the onset of the 100-ms white noise. (c) Neural contrast calculates as (post-sound activity—pre-sound activity)/(post-sound activity + pre-sound activity)*100. The neural contrast decreased during repeated stress (2-way ANOVA, condition F = 84.4, p = 4.3 × 10−20, condition: intensity interaction F = 3.7, P = 8.3 × 10−06, nested ANOVA (mouse nested within session), condition F = 50.7, p = 1 × 10−12, condition: intensity interaction F = 7.7, P = 3.1 × 10−105, mean ± SE). (d) Neural contrast for the control group. Control mice followed the same protocol but without experiencing restraint stress. There was no change in neural contrast between the first and second week of imaging (2-way ANOVA, condition F = 2.9, p = 0.08, nested ANOVA (mouse nested within session), condition F = 0.7, p = 0.3, mean ± SE). Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 4.
Alterations in tone and FM sweeps evoked activity during repetitive stress.
(a) Pure tone-evoked activity for chronically tracked cells at different sound intensities during baseline and repeated stress conditions. Activity rates decreased at moderate sound intensities, while activity at high intensities remained comparable to baseline intensities (Fig 2D, 2-way ANOVA, condition F = 16.1, p = 6.02 × 10−05, nested ANOVA (mouse nested within session) F = 16.1, p = 5.9 × 10−05, mean ± SE). Bottom: Mean activity change calculated per cell at different sound intensities. The reduction in activity was particularly striking at moderate sound intensities (t test for 40 dB p = 0.003, 50 dB p = 0.006), whereas activity at higher and lower intensities remained on par with baseline values (t test for 20 dB = 0.2, 30 dB = 0.14, 60 dB = 0.07, 70 dB p = 0.27). Values represent mean ± SE. (b) FM sweep-evoked activity for chronically tracked cells at different sound intensities during baseline and repeated stress conditions. There was a decrease in activity during repetitive stress (mean ± SE, 2-way ANOVA condition F = 13.9, p = 2.03 × 10−04, condition: intensity interaction F = 4.3, p = 0.01, nested ANOVA (mouse nested within session) F = 2.1, p = 0.001 condition: intensity interaction F = 2.1, p = 0.001), especially at moderate sound intensities (post hoc 50 dB p = 2.36 × 10−04, but not at lower 30 dB p = 1 or higher intensities 70 dB p = 1 Bonferroni corrected). Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 5.
Divergent modulation of inhibitory cells during repetitive stress.
(a) Mean noise-evoked activity (deconvolved spikes) for different noise intensities in the baseline (gray) and repetitive stress (blue) for chronically tracked SST cells (N = 4 mice, n = 121 tracked cells). The activity was averaged over a 300 ms period before (soft) and after the sound (dark colors). There was an increase in pre- and post-sound activity during repeated stress (mean ± SE, 3-way ANOVA, condition F = 805, p = 1.5 × 10−171 nested ANOVA (mouse nested within session) F = 853, p = 2.4 × 10−181). (b) Example of deconvolved spike traces of a tracked SST cell, recorded in response to noise presented at sound 50-55-60-65 dB SPL in baseline and stress sessions. Marked at time = 0 is the onset of the 100-ms white noise. (c) Mean neural contrast between the pre-sound and post-sound windows for tracked SST cells in baseline and repetitive stress. The neural contrast was calculated as (post sound activity—pre sound activity)/(post sound activity + pre sound activity)*100 per cell. There was an increase in the neural contrast during repeated stress (mean ± SE, 2-way ANOVA, condition F = 108.4, p = 3.5 × 10−25, nested ANOVA (mouse nested within session) F = 116.3, p = 7 × 10−27). (d) The activity per SST cell was normalized (z-score) before calculating the mean noise-evoked activity. There was an increase in sound-evoked activity during repeated stress (mean ± SE, 2-way ANOVA, condition F = 150.4, p = 3.4 × 10−34, nested ANOVA (mouse nested within session) F = 125.6, p = 7 × 10−29). Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 6.
The effect of stress on sound processing increases with repeated exposure.
(a) Change of noise-evoked activity over time. Comparison of noise-evoked activity after a day, 3 days, and 5 days of stress to baseline. Middle-intensity sound-evoked activity decreased as stress became chronic for PPy cells (2-way ANOVA, condition S1 F = 35.3, p = 2.9 × 10−09, nested ANOVA (mouse nested within session) F = 4.6, p = 6.7 × 10−11, S3 F = 211, p = 1.7 × 10−47, nested ANOVA (mouse nested within session) F = 218.9, p = 5 × 10−49, and S5 F = 111.4, p = 6.4 × 10−26, nested ANOVA (mouse nested within session) F = 99.2, p = 2.8 × 10−23) and PV cell (2-way ANOVA, condition S1 F = 8.4, p = 0.003, nested ANOVA (mouse nested within session) F = 8.2, p = 0.0042, S3 F = 50.9, p = 1.8 × 10−12, nested ANOVA (mouse nested within session) F = 67.1, p = 8.1 × 10−16, and S5 F = 37.6, p = 1.2 × 10−09, nested ANOVA (mouse nested within session) F = 42.6, p = 1 × 10−10). While it increased for SST cells (2-way ANOVA, condition S1 F = 44, p = 3.7 × 10−11, nested ANOVA (mouse nested within session) F = 31.1, p = 2.5 × 10−08, S3 F = 61.2, p = 6.6 × 10−15, nested ANOVA (mouse nested within session) F = 56.6, p = 6.4 × 10−14, and S5 F = 159, p = 9.6 × 10−36, nested ANOVA (mouse nested within session) F = 129.2, p = 1.8 × 10−29). Values represent mean ± SE. (b) Change in activity per tracked cell in response to a 50 dB noise is defined as the slope of a linear fit of the mean activity of the baseline days and activity on days 1, 3, and 5 of stress (examples of the linear fit on the right). We found a negative slope in most PPY and PV cells, indicating a decrease in activity as the stress becomes chronic (t test for PPY cells p = 9.8 × 10−14, PV cells p = 0.006) and a positive slope for SST cells indicating an increase in activity with successive applications of the stressor (t test for SST cells p = 9.5 × 10−06). (c) Comparison of mean neural contrast between the pre-sound and post-sound windows for tracked PPY and SST cells in baseline and different repetitive stress sessions. The neural contrast was calculated as (post sound activity—pre sound activity)/(post sound activity + pre sound activity)*100 per cell. The difference in activity between the pre-sound and post-sound periods decreased for PPy cells (2-way ANOVA, condition S1 F = 18.9, p = 1.3 × 10−05, nested ANOVA (mouse nested within session) F = 15.3, p = 9 × 10−05, S3 F = 71.7, p = 2.8 × 10−17, nested ANOVA (mouse nested within session) F = 52.5, p = 4.6 × 10−13, S5 F = 56.7, p = 5.4 × 10−14, nested ANOVA (mouse nested within session) F = 63.5, p = 1.7 × 10−15) and increased for SST cells as the stressor became chronic (2-way ANOVA, condition S1 F = 42.4, p = 8.2 × 10−11, nested ANOVA (mouse nested within session) F = 4 3.2, p = 5.6 × 10−11, S3 F = 43.7, p = 4.3 × 10−11, nested ANOVA (mouse nested within session) F = 47.68, p = 5.9 × 10−12, S5 F = 127.1, p = 4.9 × 10−29, nested ANOVA (mouse nested within session) F = 126.2, p = 7.9 × 10−29). This was especially apparent for mid-sound intensities. Values represent mean ± SE. Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 7.
Alterations in sound processing could have significant implications for perception.
(a) Left: Noise correlations between the normalized activity of pairs of PPy cells, PV cells, and a combination of both in the 500 ms surrounding the noise stimulus presentation at 50 dB SPL across the different sessions (mean ± SE). We found an increase in noise correlations for all pair types as the stressor was repeated (2-way ANOVA cell type:session F = 11.1, p = 7.8 × 10−16). Right: Same for SST cells (1-way ANOVA F = 39.5, p = 4.4 × 10−33). (b) Confusion matrices depict decoding accuracy for sound level based on population activity. Decoding was performed on 1,000 randomly drawn samples of 400 PPy units. Although the population activity in the ACtx could reliably decode sound intensity, this ability was impaired during repeated stress (all stress sessions were included in the analysis). Right, F1 scores, a measure of precision and recall, deteriorate with repeated stress exposure (mean ± SE). F1 scores were calculated as: (2*TP)/((2*TP) + FP + FN). TP—true positives, FP- false positives and FN—false negatives. Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.
Fig 8.
Repetitive stress modulates loudness perception.
(a) Left: The schematic depicts the design of the head-fixed 2AFC loudness perception task. Mice were trained to categorize a white noise stimulus as either “soft” or “loud.” During testing, mice were required to correctly categorize 40–45 dB SPL (“soft”) and 75–80 dB SPL (“loud”) to receive a reward but were allowed to choose either spout for mid-level intensities ranging from 50- to 70-dB SPL. Right: an outline of the training timeline for this task (see Methods for more details). (b) Left: Mean lick loud probability in baseline (gray) and repeated stress (purple). During repeated stress, there was a reduced loudness reporting across moderate intensities (N = 6, mean ± SE, 2-way ANOVA condition F = 121.9, p = 5.9 × 10−25, intensities F = 347.2, p = 2.7 × 10−175, interaction F = 9.11, p = 1.5 × 10−11). There was no significant change in trained intensities 40–45- and 75–80-dB SPL (post hoc, p > 0.05 Bonferroni corrected). Right: exemplary behavior of 2 mice in baseline and repeated stress conditions. (c) Perceptual boundaries in different conditions. We define the perceptual boundary as the intensity where the psychometric fit to the choice function, crosses PLoud = 0.5. Mice exhibited increased perceptual boundaries during repeated stress, indicating a decreased loudness perception (t test, p = 1.2 × 10−05, mean ± SE). (d) Control animals (n = 3) underwent an identical learning process without experiencing stress. There was no change in loudness perception (2-way ANOVA condition F = 0.21, p = 0.64, intensities F = 190.8, p = 6.7 × 10−119, interaction F = 0.08, p = 0.9, mean ± SE). (e) In control animals, there was no change in perceptual boundary (mean ± SE, t test p = 0.6). (f) Left: Schematics depict the design of head-fixed 2AFC tone in noise detection task. Middle: There was no improvement in tone in noise detection during repeated stress (2-way ANOVA condition F = 0.9, p = 0.51, mean ± SE). Right: exemplary behavior of 2 mice. Source data for this figure can be found at: https://www.ebi.ac.uk/biostudies/studies/S-BSST1689754.