Fig. 8: Single-neuron properties following inhibitory neuron ablation.

a, Baseline-normalized response reliability across trials of single neurons, averaged across all neurons per day (pink, inhibitory neuron ablation cohort; green, control cohort). Data were analyzed by one-way ANOVA (P = 0.027), two-sided t-test between baseline and postablation days with an FDR correction (top asterisks; *P < 0.05 and **P < 0.01) and permutation test for group comparisons (bottom asterisks; *P < 0.05). b, Baseline-normalized correlations averaged across nondiagonals in the similarity matrix constructed from trial-averaged population vectors. Data were analyzed by one-way ANOVA (P = 0.66), two-sided t-test with an FDR correction (P > 0.26) for all postablation days and permutation test (P > 0.21) for all postablation days. c, Fraction of sound-responsive neurons over days (one-way ANOVA, P = 0.90). d, Baseline-normalized best response amplitude of neurons responsive for each day. Data were analyzed by one-way ANOVA (P = 0.47), two-sided t-test (P > 0.13 for all postablation days) and permutation test (P > 0.13 for all postablation days). e, Change in normalized response amplitudes from baseline at the 15th largest stimulus index in the tuning curve (Extended Data Fig. 9d). Data were analyzed by two-sided t-test with an FDR correction (P = 0.34, 0.043, 0.17 and 0.34 for days 7, 9, 11 and 15; *P < 0.05) and permutation test (P > 0.085 for all postablation days). f, Change in fraction of neuron pairs with high signal correlation in the largest best response category (Extended Data Fig. 9f). Data were analyzed by two-sided t-test with an FDR correction (P = 0.079, 0.017, 0.10 and 0.15 for days 7, 9, 11 and 15; *P < 0.05) and permutation test (P > 0.17 for all postablation days). g, Response stability as the fraction of neurons categorized as sound responsive on two consecutive days. Data were analyzed by two-sided t-test: baseline (1 → 3, 3 → 5) versus postablation (5 → 7), P = 0.026; inhibitory neuron ablation versus control during 7 → 9 and 9 → 11, P = 0.048. *P < 0.05. h, Tuning shifts in the inhibitory ablation cohort relative to preablation population tuning. Left: same scheme as Fig. 5 but inhibitory neurons were ablated (black). Right: color map of baseline-subtracted fractions of spared neurons within each bin of signal correlations with day 5 high-category neurons (top: day 5 responsive spared neurons; bottom: spared neurons unresponsive on day 5 but responsive on other days). The bar plots next to the color map are baseline-averaged fractions of spared neurons in each split group (see Fig. 5). i, Baseline-subtracted change in fraction of spared neurons with high signal correlation (>0.6) toward high-category neurons in inhibitory neuron ablation and control cohorts. Top: day 5 responsive spared neurons. Bottom: spared neurons unresponsive on day 5 but responsive on other days. Data were analyzed by two-sided t-test: ablation versus control (top: P = 0.025, 6.1 × 10⁻³, 5.3 × 10⁻³, 2.8 × 10⁻³ and 0.018 for days 5, 7, 9, 11 and 15; bottom: P = 0.35, 0.028, 0.072 and 0.24 for days 7, 9, 11 and 15, respectively). Top asterisks, *P < 0.05. Bottom, data were analyzed by t-test between baseline and days 9–15 (bottom asterisk; *P < 0.05). Control cohort data were replotted for comparison; refer to Figs. 3a–f, 4g and 5h–i. Data are presented as mean ± s.e.m. across mice; inhibitory neuron ablation n = 6 and control n = 9, except for permutation tests. See Supplementary Table 1 for full statistics.