Fig. 4: Major changes in single-neuron response properties are primarily mediated by neurons gaining responsiveness after microablation.

a, Response stability quantified as the fraction of neurons classified as sound responsive on two consecutive imaging days. Data were analyzed by one-way ANOVA across days (right asterisk; *P < 0.01) and two-sided t-test between baseline change and change from days 5 to 7 (bottom asterisk; *P < 0.01). b, Top: schematic illustrating the categorization of neurons based on responsiveness on day 5 before microablation. One group (left) included neurons responsive on day 5, tracking their responses on all other days regardless of continued responsiveness. The other group (right) included neurons unresponsive on day 5 but classified responsive on other days. Bottom: normalized best response amplitude across cohorts. Left: data were analyzed by t-test (day 3 versus day 7, sound-responsive cohort: P = 4.54 × 10⁻⁴; ***P < 0.001; group comparison on day 7, one-sided t-test, FDR correction, *P < 0.05). Right: data were analyzed by one-way ANOVA across days (right asterisk; *P < 0.01). c, Change in tuning width for neurons responsive on day 5 (left) and for neurons unresponsive on day 5 but responsive on another day (right), measured as normalized response amplitudes of the 15th stimulus in the sorted tuning curve (same as Fig. 3f). Data were analyzed by two-sided t-test (sound-responsive cohort: P = 0.011, 0.25, 0.011 and 1.0 × 10⁻³ for days 7, 9, 11 and 15; *P < 0.05 and ***P < 0.001) and permutation test across groups (bottom asterisks; *P < 0.05). d, Change in the fraction of neuron pairs with high signal correlation, both responsive on day 5. Color maps show the difference in fraction normalized to baseline for four categories of neuron pairs sorted by their mean best response amplitude (left: sound responsive; middle: non-sound responsive; right: control). For the sound-responsive cohort, data were analyzed by two-way ANOVA (P = 9.5 × 10⁻⁵ across days, P = 0.0010 across amplitude bins) and t-test (days 9–15, P = 8.7 × 10⁻⁶). For the non-sound-responsive and control cohorts, P > 0.37, except the control cohort (P = 0.039 across days). e, Change in fraction of neuron pairs with high signal correlation in the largest best response category (d). Data were analyzed by permutation test (sound-responsive cohort: P = 0.010, 0.035, 0.13 and 0.21 for days 7, 9, 11 and 15, respectively); *P < 0.05. f,g, Same as d and e but for pairs of neurons categorized as unresponsive on day 5 but responsive on the other day (sound-responsive cohort: two-way ANOVA (f; left: P = 0.012 across days, P = 0.028 across amplitude bins) and t-test (days 9–15, P = 0.034); non-sound-responsive and control cohorts: (f; middle and right) P > 0.22, except non-sound-responsive (P = 0.044 across amplitude bins); permutation test (g; sound-responsive cohort: P = 0.91, 0.21, 0.040 and 0.010 for days 7, 9, 11 and 15, respectively); *P < 0.05). h,i, Same as d and e but for mixed pairs (one responsive on day 5 and one categorized as unresponsive on day 5 but responsive on the other day). Note the strong increase in pairs with high signal correlation that were categorized as unresponsive on day 5 (sound-responsive cohort: two-way ANOVA (h; left: P = 0.89 across days, P = 0.038 across amplitude bins); non-sound-responsive and control cohorts: (h; middle and right) P > 0.059; permutation test (i; all postablation days: P > 0.11 for all cohorts)). Data are presented as mean ± s.e.m. across mice (except for d, f and h); n = 10 (sound-responsive ablation), n = 10 (non-sound-responsive ablation) and n = 9 (control). Yellow vertical lines indicate the time point of ablation procedure. See Supplementary Table 1 for detailed statistics.