Article Figures & Data
Figures
- Figure 1.
Two-photon Ca2+ imaging of GCaMP6s neurons in A1. A, Experimental paradigm. Animals are raised in normal environments until at least P21. Animals then either stay in the normal lighted environment or are DE for 7 d. Cartoon by Zara Kanold-Tso. B, Imaged field with GCaMP6-expressing neurons. Exemplar neurons indicated by white circles. Scale bar = 100 μm. C, Cell-attached patch recordings in vitro from GCaMP6S-expressing neurons. Top row shows current trace, with inset showing magnified action potential. Middle row shows the corresponding Ca2+ rise (ΔF/F) in response to one action potential. Bottom row shows the corresponding two-photon fluorescence images: first image is of cell preceding spike, middle image is of cell at peak of fluorescent response, and third image shows the difference (scale bar = 5 μm). Boxplots shows median and interquartile range of fluorescent-evoked responses to one spike in control and DE mice. DE does not alter the amplitude of spike-induced fluorescence transients (mean ΔF/F ± SEM per spike: NR = 10.25 ± 0.29%, n = 62 spikes; DE = 9.97 ± 0.20%, n = 37 spikes; two-sample Kolmogorov–Smirnov test, p = 0.16). D, Sound evoked fluorescence traces in five exemplar cells (indicated in B). Black lines indicate mean trace for responses that passed the significance criterion (ANOVA p < 0.001), while thin gray traces show individual trials. Colors indicate tone frequency 4–64 kHz. E, Fraction of responsive cells decreases in L2/3 following DE (mean ± SD, NR = 64.2 ± 26.9%, DE = 37.4 ± 28.4%, Wilcoxon rank-sum test, p = 0.0167) with no change in L4 (NR = 72.1 ± 22.0%, DE = 66.9 ± 25.4%, p = 0.35). F, Spontaneous activity, as measured by SD of the baseline in ΔF/F traces, increased in L4 and L2/3 after DE (NR median ± iqr L2/3 = 6.1 ± 4.7, DE L2/3 = 6.2 ± 6.2, p = 0.0018; NR L4 = 6.6 ± 6.7, DE L4 = 9.6 ± 7.9; Wilcoxon rank-sum test, p < 10−28).
- Figure 2.
DE increases the responsiveness and frequency selectivity of neurons in both L4 and L2/3. A, Exemplar tuning curves (mean ± 1.96*SEM) of two cells obtained from tone evoked responses. B, C, Cumulative distribution functions of response amplitudes (B) and bandwidth. B, Response amplitude measured by peak ΔF/F increased in L4 and L2/3 after DE [mean ± SEM, L4 NR = 65.9 ± 0.9%, DE = 83.0 ± 1.4%; Kolmogorov–Smirnov (KS) test, p < 10−16; L2/3 NR = 61.3 ± 1.0%, DE = 67.35 ± 1.6%; KS test, p = 0.027]. C, Bandwidth decreased in L4 and L2/3 after DE (mean ± SEM, L4 NR = 1.17 ± 1.14, DE = 0.68 ± 0.79 octaves; KS test, p < 10−5; L2/3: NR = 0.98 ± 0.04, DE = 0.86 ± 0.06, KS test; p = 0.01). D, E, Response amplitudes (D) and bandwidth (E) in octave frequency bins. E, Response amplitudes in L2/3 were increased for cells with BFs of 8–16 kHz (4–8 kHz p = 0.066; 8–16 kHz p = 0.004; 16–32 kHz p = 0.07; 32–64 kHz p = 0.32). Response amplitudes in L4 were increased for cells with BFs of 4–8, 8–16, and 32–64 kHz (4–8 kHz p = 5.2 × 10−8; 8–16 kHz p = 2.4 × 10−6; 16–32 kHz p = 0.7; 32–64 kHz p = 1.4 × 10−5). D, Bandwidth in L2/3 was similar in each bin (4–8 kHz p = 0.073; 8–16 kHz p = 0.45; 16–32 kHz p = 0.075; 32–64 kHz p = 0.089). Bandwidth in L4 was decreased for cells with BFs of 8–16 and 16–32 kHz (4–8 kHz p = 0.039; 8–16 kHz p = 0.029; 16–32 kHz p = 0.29; 32–64 kHz p = 0.86).
- Figure 3.
DE alters the representation of sound frequencies in A1. Distribution of BFs in NR and DE in all imaging fields from L4 (A) and L2/3 (B) and across mice (Table 1). A, top panel, Cumulative distributions showing the spread of BFs in NR (red) and DE (black) in imaging fields from L4. The BF distribution of cells differs between DE and NR [Kolmogorov–Smirnov (KS) test; L4 p < 10−23]. Lower panels, Same data as in top panel by animals (nine mice NR; six mice DE) and binned into octaves. The mean differences for the comparisons are shown by Cumming estimation plot. The raw data are plotted on the upper axes; summary measurements (mean ± SD) are shown as lines. Mean differences for each frequency bin are plotted on the lower plot as a bootstrap sampling distribution (DABEST). Mean differences are depicted as horizontal lines; 95% confidence intervals are indicated by the ends of the vertical error bars (4–8 kHz: 18.2% [95.0%CI, 4.34, 30.7], p = 0.0432 Mann–Whitney; 8–16 kHz: –24.9% [95.0%CI –50.2, –3.86], p = 0.0518; 16–32 kHz: –10.2% [95.0%CI, –25.5, 7.41], p = 0.377; 32–64 kHz: 17% [95.0%CI, –4.21, 39.8], p = 0.0872). Effect size [CI width, lower bound, upper bound]. B, top panel, Cumulative distributions showing the spread of BFs in NR (red) and DE (black) in imaging fields from L2/3. The BF distribution of cells differs between DE and NR (KS test; L2/3 p < 10−40). Lower panels, Same data as in top panel by animals (eight mice NR; six mice DE) and binned into octaves. The mean differences for the comparisons are shown by Cumming estimation plot. Mean differences are depicted as in A (4–8 kHz: 1.8% [95.0%CI, –24.3, 14.4], p = 0.651 Mann–Whitney; 8–16 kHz: –17.1% [95.0%CI, –26.9, 2.41], p = 0.175; 16–32 kHz: –14.3% [95.0%CI, –33.3, 4.33], p = 0.22; 32–64 kHz: 32.28% [95.0%CI, 21.68, 44.99], p = 0.0024).
- Figure 4.
DE decreases pairwise activity correlations in L4. A, B, CDFs of pairwise noise (NC) and signal correlations (SC) in L4 from NR and DE animals [Kolmogorov–Smirnov (KS) test; L4 NC p = 3.7 × 10−4; L4 SC p = 6.5 × 10−28]. Lower panels show magnified view of center of distributions. C, D, Differences between the CDFs show a broad decrease in SC in L4.
- Figure 5.
DE decreases pairwise activity correlations in L2/3. A, B, CDFs of pairwise noise (NC) and signal correlations (SC) in L2/3 from NR and DE [Kolmogorov–Smirnov (KS) test; L2/3 NC p = 0.37; L2/3 SC p = 7.7 × 10−9]. Lower panels show magnified view of the center of distributions. C, D, Differences between the CDFs show a broad decrease in SCs in L2/3.
- Figure 6.
DE decreases pairwise SCs for both co-tuned and non-co-tuned neurons. A, L4 NCs in NR and DE for cells with similar and different BF (p = 0.95; p = 0.97). B, L4 SCs in NR and DE for cells with similar and different BF (p < 0.0036; p = 5.4 × 10−35). C, L2/3 NCs in NR and DE for cells with similar and different BF (p = 0.37; p = 0.96). D, L2/3 SCs in NR and DE for cells with similar and different BF (p < 1.2 × 10−5; p < 0.0094).
- Figure 7.
DE causes frequency specific effects on SCs and NCs. A, Boxplots showing NCs for L4 cell pairs with BFs of 4–8, 8–16, 16–32, and 32–64 kHz from NR and DE (p = 7.3 × 10−18, p = 1.2 × 10−10, p = 8.3 × 10−5, p = 1.9 × 10−22). B, Boxplots showing SCs for L4 cell pairs with BFs of 4–8, 8–16, 16–32, and 32–64 kHz from NR and DE (p = 7.5 × 10−17, p = 4.4 × 10−12, p = 5.4 × 10−20, p = 8.38 × 10−10). C, Boxplots showing NCs for L2/3 cell pairs with BFs of 4–8, 8–16, 16–32, and 32–64 kHz from NR and DE (p = 0.006, p = 0.48, p = 0.0004, p = 0.73). D, Boxplots showing SCs for L2/3 cell pairs with BFs of 4–8, 8–16, 16–32, and 32–64 kHz from NR and DE (p = 0.002, p = 0.0001, p = 0.22, p = 0.16).
- Figure 8.
DE-induced changes in pairwise correlations can depend on bandwidth sum. A, NCs for L4 cell pairs as a function of summed bandwidth from NR and DE. B, SCs for L4 cell pairs as a function of summed bandwidth from NR and DE from NR and DE (** indicates significant difference at p < 0.01). C, NCs for L2/3 cell pairs as a function of summed bandwidth from NR and DE (* indicates significant difference at p < 0.05). D, SCs for L2/3 cell pairs as a function of summed bandwidth from NR and DE (* indicates significant difference at p < 0.05).
Tables
Group Layer Animals Fields Mean depth Total cells Responding cells Control L2/3 8 15 189 ± 4μm 1573 989 L4 9 14 370 ± 11 μm 1202 846 DE L2/3 6 19 191 ± 4 μm 1919 682 L4 6 14 354 ± 41 μm 1099 710 Responding cells denote neurons from the total cell population that showed a significant tone-evoked response to at least one frequency (ANOVA across 10 repetitions, p < 0.001).
In this issue
