Dark exposure reduces high-frequency hearing loss in C57BL/6J mice

31 Plastic changes in the brain are primarily limited to early postnatal periods and have been mainly 32 studied within the same modality. Crossmodal plasticity refers to neural plasticity that allows 33 adaptation to the loss of a sensory modality. Sensory modality loss can occur during pathological 34 states of the peripheral sensory systems, such as loss of hair cells in the cochlea, resulting in 35 deafness, or damage of the eyes, resulting in blindness (1-4). Crossmodal plasticity is thought to 36 underlie the enhanced auditory abilities of the early- (1, 5-8) and late-blind (7, 9). Several circuit-37 level plasticity changes have been observed in adult mice with temporary visual deprivation (dark 38 exposure, DE). DE in adult mice has profound effects on the primary auditory cortex, such as 39 strengthening of thalamocortical synapses (10), refinement of excitatory and inhibitory 40 intracortical circuits (11-13), and selective reduction of thalamic -reticular nucleus -mediated 41 inhibition of the auditory thalamus (14). These circuit changes correlate with changes in the 42 sound-evoked responses, such as reduced thresholds, increased gain, increased frequency 43 selectivity (10), and decorrelation of spatiotemporal population responses (15), which together 44 should lead to increased coding fidelity. We thus investigated if DE leads to improved auditory 45 ability. 46 To understand the effects of DE on auditory processing, we placed 48 C57BL/6J (C57) 47 and 48 CBA/CaJ (CBA) mice in home cages fitted with an automated auditory behavior system 48 (“ToneBox”) that allowed continuous long-term observation (Fig. 1A) (16, 17). To test the effects 49 of DE, we take advantage of both the C57BL/6J mouse strain that develops progressive hearing 50 loss with age as well as CBA mice that retain normal hearing (18-24). Using the C57BL/6J strain 51 enables us to test the effects of DE on hearing frequency bands with normal (low and mid 52 frequencies) and reduced complement of hair cells (high-frequencies). 53 Once placed in the ToneBox, mice receive d a water reward for detecting tones from the 54 ToneBox speaker. After the tone was presented, mice had a reward window to lick the ToneBox 55 water spout. Licking was detected with a capacitive sensor. By placing multiple animals in the 56 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 4 cage, we avoided the effects of social isolation. Mice live d and performed the task in the cage 57 with no interaction with humans for the duration of the experiment, except for biweekly bedding 58 change. In the ToneBox, we continuously presented 88 different combinations of different sound 59 frequencies (11 total tones) and sound amplitudes (8 total amplitudes), which enabled us to 60 construct long-term “performance audiograms.” Figure 1B shows an example of a typical week -61 long timeline of hit activity showing, as expected, that mice were active in the dark cycle. 62 63

64 DE reduces the decline in performance in the high-frequency band of C57BL/6J mice. 65 Mice were 63 days old at the start of the experiment. We divided the mice of both 66 C57BL/6J and CBA lines into two groups respectively: Control (CT) and DE (8 ToneBoxes, 24 67 mice per group) (Fig. 1C). Mice were placed into the ToneBoxes for an initial 14-day habituation 68 and shaping phase to stabilize their performance (Fig. 1D). We divided the following 70-day-long 69 experiment timeline into five periods. Period I was 7 day-long while Periods II, III, IV, and V were 70 14 day -long. Periods I, III, and V were under a typical 12 -hour light/dark cycle for both 71 experimental groups. In periods II and IV, the DE group was hermetically sealed from the room 72 light, while the CT group remained in the normal Light/Dark cycle. Normal circadian rhythm was 73 present throughout the experiment, including DE periods. Example recordings of hit rates from 74 the complete 77 -day-long timeline are shown for all four groups: the C57BL/6J CT, CBA CT, 75 C57BL/6J DE, and CBA DE groups (Fig. S1). 76 In CT C57BL/6J ToneBoxes, we noticed that the hourly hit rates show a decrease in 77 performance for the highest frequency band (40kHz) with increasing age (Fig. 1E). We performed 78 a spot test on day 70 of the 40kHz band against performance from all other frequency bands and 79 the difference was statistically significant (t-test, p = 0.0273). This decreased performance is 80 consistent with the development of presbycusis in these mice due to peripheral hearing loss (18-81 20). Based on ABR measurements (18-20), the onset of hearing loss in C57BL/6J mice in the 82 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 5 high frequencies is approximately at the start of Period II (Postnatal day 84, P84), consistent with 83 the decreased performance we observed. In contrast, DE C57BL/6J ToneBoxes did not show 84 decreased performance at high frequencies and only showed a slight decline later in the 85 experimental timeline (Fig. 1F). This delay in decline will be described below in detail. At day 70 86 performance was similar between groups (t-test, p >> 0.05). Consistent with preserved peripheral 87 hearing in CBA mice, the 40kHz band did not show any decline with age, and groups performed 88 similarly at day 70 (t-test, p >> 0.05) (Fig. 1G and 1H). These observations suggest that the DE 89 periods in visually deprived animals reduced the decline of the processing of high -frequency 90 sounds in C57BL/6J mice. 91 92 DE increases tone detection performance in CBA and C57BL/6J mice. 93 To first identify how animal performance varied with frequency and amplitude and to 94 investigate whether DE affected the frequency and amplitude (FxA)-dependent hit rate, we 95 calculated the ratio of hit rates between the first vs. last period hit rates for each FxA bin. We first 96 plotted these heatmaps for C57BL/6J CT and DE groups (Fig. 2A). In the CT C57BL/6J group, 97 the hit rates decreased for the whole high -frequency band across sound amplitudes (Fig. 2A, 98 black arrows , 32 & 40kHz all SPL levels ). This decline in performance occurs in a similar 99 frequency band as presbycusis reported in C57BL/6J mice of this age (19). In contrast, such 100 decreased performance was not present in the C57BL/6J DE group. These results suggest that 101 DE attenuates the high-frequency-specific decline of detection performance in C57BL/6J mice. 102 We next investigated if there was an overall decreased performance in CT C57BL/6J mice 103 compared to DE mice. We calculated the total hit rates by merging all FxA bins in a given period, 104 which gave us a rough measure of change in overall tone detection performance, independent of 105 the tested stimulus (Fig. 2B, left). The changes in total hit rates in CT and DE C57BL/6J mice 106 were similar (CT: -0.53% ± 10.06% SEM, DE: + 2.91% ± 2.19% SEM, t-test, p = 0.7428). This 107 indicates that the overall performance of CT and DE mice was similar. Given that hit rates 108 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 6 decreased for high frequencies in CT mice, this suggests that CT mice have relatively more hits 109 at lower frequencies. In contrast, in C57BL/6J DE mice high-frequency performance is preserved. 110 Together, these results suggest CT C57BL/6J mice show decreased hit rates for high 111 frequencies and an increased hit rate at lower frequencies while in DE C57BL/6J mice high-112 frequency performance is preserved. 113 We next analyzed the rate -of-change audiograms for the CBA CT and DE groups (Fig. 114 2C). In contrast to C57BL/6J mice CT CBA mice do not show decreased performance at high 115 frequencies. However, after DE we observed a widespread increase in hit rates (Fig. 2C). In the 116 CBA DE group hit rates increased by +9.00% ± 4.60% SEM while in the CBA CT group hit rates 117 decreased by -4.15% ± 6.49% (Fig. 2D, left ; t-test, p = 0.1205). While this difference is not 118 significant when summed over the whole frequency spectrum, qualitative inspection suggests that 119 the frequency band-specific performance in low and mid frequencies is enhanced by DE. 120 To examine band-specific performance for both C57BL/6J and CBA groups in detail we 121 analyzed the relationship between sound amplitude and hit rates in each frequency band. All trials 122 for the C57BL/6J CT group were first binned based on the sound amplitude parameter, and the 123 global correlation coefficient was calculated to be 0.9679 ± 0.0058 SEM, while the remaining three 124 groups followed a similar pattern. We thus used linear regression analysis to test whether the 125 slope (gain, abbr. SL ) of this relationship and/or baseline performance (inter cept, abbr. I C) is 126 changed after DE for C57BL/6J mice (Fig. 2E) and CBA mice (Fig. 2F). We used linear regression 127 model fitting to estimate the parameters of the linear fit for each frequency band and each group. 128 These linear fit estimates, together with 95% confidence intervals (CI) are shown as lines in 129 respective colors (fit) and 95% CI as shaded areas for plots in both subpanels. To determine if 130 these model estimates differ, the difference between the two groups was also fitted and tested for 131 the significance of parameters using a Multiple comparison correction ( Bejnamini & Hochberg 132 false discovery rate procedure (25) ). In Fig. 2E and 2F, we plot results for 8kHz and 40kHz bands, 133 and the statistics for the remaining frequency bands are given in Table S1 . First, this analysis 134 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 7 confirms our visual observation of a major decline in the 40kHz band for the C57BL/6J group in 135 Fig. 2A where IC of CT and DE fits differed significantly ( F-test, p = 0.0027). Secondly, t he 136 frequency band-specific comparisons for CBA mice show differences of the intercept parameter 137 between the two groups for several low- and mid-frequency bands (4kHz: p = 0.0115; 6.3kHz: p 138 = 0.0014; 8kHz shown in Fig. 3B: p = 0.0079; 10kHz: p = 0.0005; 12.5kHz: p = 0.0107; 16kHz: p 139 = 0.0071; 25.0kHz: p = 0.0107; all other n.s. bands in Table S1). These rate -of-change 140 audiograms suggest that DE increased the total period hit rate across a broad frequency range in 141 CBA mice. 142 Together, these results suggest that DE increases performance on an auditory detection 143 task in both CBA and C57BL/6J mice. However, the details in which DE benefits audition seem 144 to differ between the two models: We hypothesize that in CBA mice with preserved hearing DE 145 facilitates the improvement of tone detections across frequency ranges, while in C57BL/6J DE 146 allows compensation to attenuate the age-dependent high-frequency hearing loss. 147 148 DE delays the effects of presbycusis in C57BL/6J mice by 12 days. 149 Our results suggest that C57BL/6J DE mice have higher behavioral performance at high 150 sound frequencies compared to CT at the end of our experimental time window. We next aimed 151 to identify the detailed behavioral time courses of these performance differences. We thus 152 analyzed the changes in the sound-frequency-dependent performance in the CT and DE groups 153 over time . For each ToneBox w e calculated the normalized hit rate (NHR) for each stimulus 154 (frequency & amplitude, F xA) condition in each hourly time bin for the entire 63 days of the 155 experiment. We normalized the stimulus-specific (FxA) hit rates to the total hit rate of a given 156 ToneBox for the same period across all conditions, meaning that the ToneBox with the normalized 157 hit rate of 1 for a given FxA band had the same hit rate as all the FxA bands combine d. This 158 normalization enabled us to investigate the distribution of hit preference and ability across the 159 sound spectrum. To minimize the effects of the 24-hour circadian rhythm, we averag ed these 160 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 8 normalized hit rates for both the CT and DE groups with a moving 168-hour window. NHR activity 161 for C57BL/6J mice is shown for 4, 8, and 40kHz bands (Fig. 3A) and all remaining frequency 162 bands (Fig. S2A). In C57BL/6J CT cages, a drop of the NHR for the 40kHz band is present starting 163 around Day 40. In the C57BL/6J DE group, a much weaker drop is present at much older ages 164 (Fig. 3A, right). To better define this difference in the onsets of performance decline for CT and 165 DE groups, we defined the onset of decline as the first day when mean performance drops two 166 standard deviations below baseline performance from period I. Respective days are labeled with 167 colored arrows on the x-axis. This difference turned out to be 12 days (Day 45 vs 57). For the 168 4khz and 8kHz bands, no decreases in NHR were observed. Thus, DE delays the development 169 of the behavioral effects of high -frequency hearing loss in C57BL/6J mice. Notably, significant 170 differences in the 8kHz band were observed during period one around days 14-20. This was likely 171 due to group fluctuations in the shaping phase and these differences ceased after day 20. Lastly, 172 there was an increased preference for 4 and 6.3kHz bands in the CT group (Fig. S2) that occurred 173 at the same time as the decline of the 40kHz band of this group. As discussed above, these 174 increases were likely the compensation for lost ability in the 40kHz band. 175 In contrast to C57BL/6J mice, CBA mice do not suffer from any systemic peripheral 176 hearing loss at this age. Consistent with this, our analysis shows that CBA CT and DE mice do 177 not show differences in NHR in the 4, 8, and 40kHz bands (Fig. 3B) and all other frequencies (Fig. 178 S2B). These results confirm that frequency preferences in CBA mice did not change with DE as 179 was the case in the C57BL/6J group. Thus, the increase in absolute hit rates we observe in CBA 180 mice (Fig. 2C, D) is relatively widespread across the frequency spectrum. 181 Together, these results support our hypothesis that DE selectively increases the relative 182 performance of C57BL/6J mice at high frequencies while providing a more general benefit across 183 a wide range of frequency spectrum in CBA mice. 184 185 The effect of DE on performance is present across sound amplitudes. 186 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 9 So far, we have lumped the performance at all sound amplitudes. We next evaluated if the 187 increased performance of C57BL/6J DE mice was present at all sound amplitudes. We thus 188 computed performance audiograms. We averaged cage NHRs for each period and plotted 189 smoothed heatmaps for Period I and V of the C57BL/6J CT and DE groups (Fig. 4A; see Fig. S3 190 for data from all periods). The interquartile range (IQR) contours for high-frequency bands differ 191 between C57BL/6J CT and C57BL/6J DE, and the CT group shows a shift of the IQR contours 192 toward lower frequencies. To evaluate this further, we plotted the ratios of C57BL/6J DE vs. CT 193 NHR (Fig. 4A bottom). The results from the C57BL/6J group indicate that DE effects are present 194 across high-frequency bands (32 and 40kHz). A similar analysis from the CBA group shows very 195 stable audiograms where the contours of the audiograms of the NHRs are nearly identical for 196 Periods I and V (Fig 4B; see Fig. S4 for data from all periods). This is also visible in the ratios of 197 CBA DE vs. CBA CT NHR (Fig. 4B bottom) where only minor differences are observed. We will 198 next quantify these differences in detail. 199 DE enhancement across performance levels suggested th at DE had effects across tone 200 amplitudes. We thus next investigated the DE enhancement effect in the amplitude domain by 201 plotting the NHRs of the FxA bands along the amplitude dimension for two frequency bands of 202 the C57BL/6J group from the period I and V: 8 and 40kHz, comparing the relationship between 203 sound amplitude and NHR with our linear regression model as previously done on raw hit rate 204 rations in Fig. 2E and F (Fig. 5A; see Fig. S5A for data from all periods and frequencies). 205 As can be seen from Fig. S5A and Table S1, the significance of the difference between 206 CT and DE groups for C57BL/6J mice is restricted to only 4 cases, 8kHz band in Period I that 207 likely originated in group-wide fluctuations of performance in late days of the training (F-test, 208 p=0.0465), 40kHz band for Period IV and V that signifies the rescuing effects of DE on 40kHz 209 band (F-test, P.IV: p = 0.0003; P.V: p < 0.0001, and, lastly, 4kHz band in Period V which is 210 compensatory effect of lost performance in the 40kHz band during the same period. For the last 211 case, the slope parameter was also significant, meaning that CT mice increased their 4kHz band 212 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 10 performance with increased gain compared to the DE group (F-test, IC: p = 0.0082; SL: p = 213 0.0117). In contrast, the CBA group revealed no significant differences between the two models 214 (Fig. 5B and Fig. S5) for any frequency band or period (See Table S1). 215 Together, these analyses suggest that DE -induced auditory behavioral enhancement 216 seen in the C57BL/6J group leads to decreased thresholds across the high-frequency spectrum 217 after the onset of hearing loss. 218 219 The effect of DE emerges gradually. 220 We next investigated in detail the time course of the effects of DE. We thus generated NHR across 221 the experimental periods for all stimulus combinations for both the C57BL/6J CT, C57BL/6J DE, 222 CBA CT, and CBA DE groups and consequently plotted the ratio s of CT vs. DE for both the 223 C57BL/6J group (Fig. 6A) and CBA group (Fig. 6B). This analysis shows that the onset of DE 224 enhancement for the C57BL/6J group emerged around day 40 for the softest high -frequency 225 sounds and that improvements in the 32kHz bin emerged at around day 60. Given that day 40 226 was between our first and second DE periods, these data suggest that the first DE period could 227 already have a long-lasting effect on tone detection performance. 228 To test if a single period of DE could have a preventative effect, we trained an additional 229 cohort of C57BL/6J animals starting at P84 up to P140 (4 cages). Then, we performed two weeks 230 of DE timed to match Period IV in a postnatal reference. A single period of DE also resulted in 231 preventing the decreased performance in the 40kHz band (Fig. S6). Thus, a single period of DE 232 was able to reduce the effect of presbycusis. 233 234

235 Our results show that temporary visual deprivation via DE in adults enhances the behavioral 236 performance of C57BL/6J mice in tone detection tasks in high-frequency bands where the effects 237 of presbycusis are usually evident. Additionally, we observed broad increases in the performance 238 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 11 of low and mid frequencies in CBA mice that do not suffer from any systemic hearing loss at this 239 age. 240 Our automated design allowed us to gather hundreds of thousands of trials per cage . 241 Because of the minimalistic impact of the experiment design on mice’s daily routine, we eliminated 242 several confounds commonly appearing in rodent behavior studies, such as repeated handling of 243 animals (26). The hearing of both C57BL/6J and CBA mouse lines w as previously studied 244 extensively by several methods, most notably auditory brainstem response (ABRs) (19, 20, 22). 245 Threshold intensity shifts caused by presbycusis in C57BL/6J mice were observed as early as 246 P30 (27). While we did not measure the ABRs of individual mice, our experimental design 247 minimized variability. First, we used large cohorts of animals enabled by our automatic system. 248 Second, all animals were subject to the same developmental conditions until they were distributed 249 to two experimental groups at the same age (P63) before the start of the experiment. Third, 250 animals were group-housed in the ToneBox, thus each ToneBox recording represents a 251 composite of the performances of the three individual mice within a given cage. This within-cage 252 averaging further reduced the effect of the population variability on the hearing capabilities of 253 individual animals. 254 While we here use a tone -detection task, the circuit and functional changes of DE are 255 widespread and include the sharpening of tuning curves (10). We predict that DE affects a variety 256 of auditory tasks. Indeed, training on auditory temporal discrimination tasks can also improve 257 spectral tuning (28), suggesting that mechanisms engaged by training are affecting general sound 258 processing. Given that DE has an effect on thalamic (14), thalamocortical (10), and intracortical 259 (11-13) auditory circuits we expect that performance in a variety of auditory task s is improved. 260 Our data show an enhancement of the performance at high frequencies in C57BL/6J mice. This 261 enhancement is consistent with the increased number of neurons responsive to high frequencies 262 after DE in C57BL/6J mice (15). 263 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 12 C57BL/6J mice have early onset of presbycusis – gradual age -related hearing loss, 264 evident in ABRs, otoacoustic emissions, and startle behavior for the high-frequency range starting 265 around 10 weeks of age (18-20). When mice reached an age corresponding to when high -266 frequency ABR hearing threshold shifts were evident in this strain (19), we observed a decline in 267 performance in our operand conditioning task in the high-frequency band. We noted that Control 268 mice show an increase in the relative amount of hits to low frequencies, indicating that they shift 269 their behavior to relatively “easier” stimuli to keep their water consumption constant. In contrast, 270 DE mice showed better performance at high frequencies. What mechanism could underlie this 271 improved performance? Hearing loss in C57BL/6J is caused by degeneration of the cochlea and 272 this degeneration starts at high frequencies (22-24). Such degeneration results in reduced 273 ascending sound-evoked activity and subsequently reduced activation of the auditory cortex for 274 high-frequency stimuli (29). Mechanistically, our behavioral data could be explained by the circuit 275 level plasticity we reported in previous studies. DE induces an increase in the strength of auditory 276 thalamocortical synapses (10), which can counteract the reduced afferent drive which leads to 277 enhanced sensitivity and increased responsiveness to sound stimuli in the thalamocortical 278 recipient layers of the auditory cortex (10). Consistent with the increase in thalamocortical 279 synaptic gain, we observed that DE leads to a steep increase in firing rates with changes in sound 280 amplitudes (10). Intracortical circuits and thalamic circuits can alter gain and adult DE has been 281 shown to affect both. After DE, ascending and recurrent intracortical circuits change synaptic 282 strength (11-13), and refine their connections, leading to a more efficient information transmission 283 (12). In addition to changes on the cortical and thalamocortical level, DE reduces inhibition from 284 the thalamic reticular nucleus to the auditory thalamus, enhancing the ascending transmission of 285 sound information through the thalamus (14). DE could also have effects on spectral contrast 286 tuning sensitivity (30). As previously shown, DE induces decorrelation of the sound -evoked 287 population activity in A1 which can lead to increased encoding fidelity of represented stimuli in 288 the cortex (15, 31). Together, all these circuit changes, both at the thalamic and cortical levels, 289 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 13 enhance the transmission of the weakened high-frequency ascending signals to the auditory 290 cortex and lead to an increased representation of high-frequency tones in the auditory cortex after 291 DE (15). This potentiation of the feedforward circuit and refinement of the intracortical circuit could 292 allow better detection and sharper tuning needed to allow the processing of reduced auditory 293 signals arising from age -related peripheral hearing loss . We reason that these extensive circuit 294 changes compensate for decreased ascending drive and lead to the observed reduced behavioral 295 performance declines after DE. 296 Studies investigating early and life-long visual deprivations have shown various functional 297 and circuit changes that can give rise to improved auditory performance (32-34). We find that DE 298 can also induce such changes in adult animals. The changes we here see with DE improve 299 auditory behavior in a model of presbycusis is consistent with the idea that the behavioral deficits 300 in presbycusis are not solely due to loss of inner hair cells in the cochlea but also due to changes 301 in the brain. Indeed, the aging auditory cortex in CBA mice that do not suffer from peripheral 302 hearing loss also shows altered sound -evoked activity, such as increased correlations and 303 reduced ability to control activity correlations (35, 36). Our observation of increased hit rates in 304 DE CBA mice across low- and mid-frequencies is consistent with these findings. In conclusion, 305 this study suggests that the changes in central auditory processing lead to the increased ability 306 of animals to perform auditory tasks after DE. Furthermore, our data collectively suggest that DE 307 could be a simple method to reduce some of the effects of central aging and enhance the efficacy 308 of auditory performance with cochlear implants. 309 310 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 14 311 312 Figure 1: DE reduces loss of detection performance for high-frequency tones. 313 (A) Automated home-cage training system with ToneBox. Tones are randomly presented from 4-314 40kHz and 30-65dB amplitude. Three animals were placed in each training cage, either C57BL/6J 315 or CBA mice. These groups are labeled with a crossed-out ear pictogram for the C57BL/6J group 316 and a non-crossed-out ear pictogram to label the CBA group. This notation is used for all figures. 317 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 15 (B) Example performance during one week. Gray bars indicate the hit rate within each hourly time 318 bin. (C) 8 ToneBoxes under control 12h/12h light/dark conditions or during DE. (D) Experimental 319 timeline. DE cohort receives two 2-week DE periods (II & IV). Animals begin the Habituation and 320 Shaping phase at postnatal day 63 (P63). (E, F) Moving average hit rates for 40kHz or all other 321 tones for C57BL/6J CT and DE groups (N=8 for both). The black line in (E) on day 70 shows the 322 difference between the two observed means and this difference is statistically significant (t-test, 323 p < 0.05). Data averaged over all sound amplitude levels. (G, H) Same as in (E, F) but for CBA 324 CT and DE groups (N=8 for both). 325 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 16 326 Figure 2: DE causes broad increases in performance in CBA and C57BL/6J mice. 327 (A) Rate-of-change audiograms for the CT (Left) and DE (Right) C57BL/6J groups. Black arrows 328 indicate the area of high-frequency band performance which significantly deteriorated throughout 329 the experiment. (B) Bar plot showing the average change of total period hit rates between periods 330 I and V. Red C57BL/6J CT, Blue DE C57BL/6J group. Black lines indicate SEM. ‘n.s’, ‘*’, indicate 331 statistical significance ( t-test, non-significant, p < 0.05 respectively). (C) Rate-of-change 332 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 17 audiograms for the CT (Left) and DE (Right) CBA groups. (D) Same as in B, but for CBA CT and 333 CBA DE groups. (E) Scatter points show hit rate ratios for 8kHz and 40kHz as a function of 334 amplitude in the C57BL/6J CT (red) and DE (blue) groups between Periods I and V. Vertical lines 335 show SEM. Scatter points are overlapped with linear regression model fit in matching colors. 336 Shaded areas are 95% confidence intervals of the fit. Inter cept (I C) or slope (SL) Difference 337 appears if the p-value of the F-test showed significance for either parameter of the group 338 difference linear fit model (F-test, Bejnamini & Hochberg false discovery rate procedure applied. 339 Does not appear for non -significant, * for p < 0.05, & ** for p < 0.01). Dashed black horizontal 340 lines outline an NHR level of 1. Dashed black horizontal lines outline hit rate ratio 1 (Where mean 341 performance in both periods was the same) (F) Same as in (E) but for CBA CT and DE groups. 342 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 18 343 Figure 3: Rescuing effects of DE emerge in Period 3 of the C57BL/6J group. 344 (A) (Left axis) Normalized hit rates for 4kHz, 8kHz, and 40kHz tones for CT C57BL/6J (red) and 345 DE C57BL/6J (blue) ToneBoxes. Data averaged over all SPL levels. The shaded error bar 346 represents the standard error of the mean. Red and blue arrowheads in the right panel for 40kHz 347 mark the point in the timeline (days 45 and 57) where mean NHR deviated from the baseline by 348 two standard deviations for CT and DE groups respectively. (Right axis) The absolute value of t-349 statistic from a two-sample t-test, if the t-test was performed at a given point in time. Green bars 350 show data points of statistical significance (t-test, p < 0.05). (B) Same as in (A), but for the CBA 351 CT and DE groups. 352 353 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 19 354 Figure 4: Frequency-amplitude-dependent performance is not altered by DE in CBA mice. 355 (A) Normalized hit rate audiograms for all stimulus conditions for periods I (left) and V (right) of 356 the C57BL/6J CT group (upper row) and C57BL/6J DE group (middle row). Dashed lines indicate 357 quartiles of smoothed NHR. Red IQR lines for the DE group are plotted as CT reference. (lower 358 row) The ratio of normalized hit rates between CT and DE C57BL/6J. (B) Same as in (A) but for 359 the CBA CT and DE groups. 360 361 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 20 362 Figure 5: DE performance increases are multiplicative across levels. 363 (A) Scatter points show normalized hit rates for 8kHz and 40kHz as a function of amplitude in the 364 C57BL/6J CT (red) and DE (blue) group during Period I (left column) and Period V (right column). 365 Vertical lines show SEM. Scatter points are overlapped with linear regression model fit in 366 matching colors. Shaded areas are 95% confidence intervals of the fit. Intercept (IC) or slope (SL) 367 Difference appears if the p-value of the F-test showed significance for either parameter of the 368 group difference linear fit model ( F-test, Bejnamini & Hochberg false discovery rate procedure 369 applied. Does not appear for non -significant, * for p < 0.05, & ** for p < 0.01). Dashed black 370 horizontal lines outline an NHR level of 1. (B) Same as in (A) but for CBA CT and DE groups. 371 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 21 372 Figure 6: DE performance for quietest high-frequency tones increases after the first DE 373 period. 374 (A) The ratio of hit rates between CT and DE for all stimulus combinations of the C57BL/6J group. 375 Arrows: Differences for 40kHz emerge after day 40, while differences for 32kHz emerge around 376 day 65. Dashed vertical lines mark the individual periods. 377 (B) Same as in (A) but for the CBA group. 378 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 22

379 1. D. H. Ashmead et al., Spatial hearing in children with visual disabilities. Perception 27, 380 105-122 (1998). 381 2. C. J. Sabourin, Y. Merrikhi, S. G. Lomber, Do blind people hear better? Trends Cogn Sci 382 26, 999-1012 (2022). 383 3. A. K. Singh, F. Phillips, L. B. Merabet, P. Sinha, Why Does the Cortex Reorganize after 384 Sensory Loss? Trends Cogn Sci 22, 569-582 (2018). 385 4. M. Teichert, J. Bolz, How Senses Work Together: Cross-Modal Interactions between 386 Primary Sensory Cortices. Neural Plast 2018, 5380921 (2018). 387 5. A. Dufour, Y. Gerard, Improved auditory spatial sensitivity in near-sighted subjects. Brain 388 Res Cogn Brain Res 10, 159-165 (2000). 389 6. N. Lessard, M. Pare, F. Lepore, M. Lassonde, Early-blind human subjects localize sound 390 sources better than sighted subjects. Nature 395, 278-280 (1998). 391 7. P. Voss et al., Early- and late-onset blind individuals show supra-normal auditory abilities 392 in far-space. Curr Biol 14, 1734-1738 (2004). 393 8. F. Gougoux et al., Neuropsychology: pitch discrimination in the early blind. Nature 430, 394 309 (2004). 395 9. C. Holig, J. Focker, A. Best, B. Roder, C. Buchel, Crossmodal plasticity in the fusiform 396 gyrus of late blind individuals during voice recognition. Neuroimage 103, 374-382 (2014). 397 10. E. Petrus et al., Crossmodal induction of thalamocortical potentiation leads to enhanced 398 information processing in the auditory cortex. Neuron 81, 664-673 (2014). 399 11. E. Petrus et al., Vision loss shifts the balance of feedforward and intracortical circuits in 400 opposite directions in mouse primary auditory and visual cortices. J Neurosci 35, 8790-401 8801 (2015). 402 12. X. Meng, J. P. Kao, H. K. Lee, P. O. Kanold, Visual Deprivation Causes Refinement of 403 Intracortical Circuits in the Auditory Cortex. Cell Rep 12, 955-964 (2015). 404 13. X. Meng, J. P. Kao, H. K. Lee, P. O. Kanold, Intracortical Circuits in Thalamorecipient 405 Layers of Auditory Cortex Refine after Visual Deprivation. eNeuro 4 (2017). 406 14. J. L. Whitt et al., Visual Deprivation Selectively Reduces Thalamic Reticular Nucleus-407 Mediated Inhibition of the Auditory Thalamus in Adults. J Neurosci 42, 7921-7930 408 (2022). 409 15. K. Solarana, J. Liu, Z. Bowen, H. K. Lee, P. O. Kanold, Temporary Visual Deprivation 410 Causes Decorrelation of Spatiotemporal Population Responses in Adult Mouse Auditory 411 Cortex. eNeuro 6 (2019). 412 16. N. A. Francis, K. Bohlke, P. O. Kanold, Automated Behavioral Experiments in Mice 413 Reveal Periodic Cycles of Task Engagement within Circadian Rhythms. eNeuro 6 414 (2019). 415 17. N. A. Francis, P. O. Kanold, Automated Operant Conditioning in the Mouse Home Cage. 416 Frontiers in neural circuits 11, 10 (2017). 417 18. K. Parham, Distortion product otoacoustic emissions in the C57BL/6J mouse model of 418 age-related hearing loss. Hear Res 112, 216-234 (1997). 419 19. J. R. Ison, P. D. Allen, W. E. O'Neill, Age-related hearing loss in C57BL/6J mice has 420 both frequency-specific and non-frequency-specific components that produce a 421 hyperacusis-like exaggeration of the acoustic startle reflex. J Assoc Res Otolaryngol 8, 422 539-550 (2007). 423 20. Q. Zhang et al., Identifying microRNAs involved in degeneration of the organ of corti 424 during age-related hearing loss. PLoS One 8, e62786 (2013). 425 21. J. F. Willott, S. Carlson, H. Chen, Prepulse inhibition of the startle response in mice: 426 relationship to hearing loss and auditory system plasticity. Behav Neurosci 108, 703-713 427 (1994). 428 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 23 22. K. R. Johnson et al., Effects of Cdh23 single nucleotide substitutions on age-related 429 hearing loss in C57BL/6 and 129S1/Sv mice and comparisons with congenic strains. Sci 430 Rep 7, 44450 (2017). 431 23. S. N. Park et al., Comparison of cochlear morphology and apoptosis in mouse models of 432 presbycusis. Clin Exp Otorhinolaryngol 3, 126-135 (2010). 433 24. V. P. Spongr, D. G. Flood, R. D. Frisina, R. J. Salvi, Quantitative measures of hair cell 434 loss in CBA and C57BL/6 mice throughout their life spans. J Acoust Soc Am 101, 3546-435 3553 (1997). 436 25. Y. Benjamini, Y. Hochberg, Controlling the False Discovery Rate: A Practical and 437 Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B 438 (Methodological) 57, 289-300 (1995). 439 26. J. P. Balcombe, N. D. Barnard, C. Sandusky, Laboratory routines cause animal stress. 440 Contemp Top Lab Anim Sci 43, 42-51 (2004). 441 27. S. Allan, P. Rémy, Age-related changes in the C57BL/6J mouse cochlea. I. Physiological 442 findings. Developmental Brain Research 2, 65-75 (1981). 443 28. Y. Cheng et al., Positive impacts of early auditory training on cortical processing at an 444 older age. Proc Natl Acad Sci U S A 114, 6364-6369 (2017). 445 29. J. F. Willott, Effects of aging, hearing loss, and anatomical location on thresholds of 446 inferior colliculus neurons in C57BL/6 and CBA mice. J Neurophysiol 56, 391-408 447 (1986). 448 30. D. L. Barbour, X. Wang, Contrast tuning in auditory cortex. Science 299, 1073-1075 449 (2003). 450 31. B. B. Averbeck, P. E. Latham, A. Pouget, Neural correlations, population coding and 451 computation. Nat Rev Neurosci 7, 358-366 (2006). 452 32. J. P. Rauschecker, M. Korte, Auditory compensation for early blindness in cat cerebral 453 cortex. J Neurosci 13, 4538-4548 (1993). 454 33. J. U. Henschke et al., Early sensory experience influences the development of 455 multisensory thalamocortical and intracortical connections of primary sensory cortices. 456 Brain Struct Funct 223, 1165-1190 (2018). 457 34. J. P. Rauschecker, U. Kniepert, Auditory localization behaviour in visually deprived cats. 458 Eur J Neurosci 6, 149-160 (1994). 459 35. K. Shilling-Scrivo, J. Mittelstadt, P. O. Kanold, Altered Response Dynamics and 460 Increased Population Correlation to Tonal Stimuli Embedded in Noise in Aging Auditory 461 Cortex. J Neurosci 41, 9650-9668 (2021). 462 36. K. Shilling-Scrivo, J. Mittelstadt, P. O. Kanold, Decreased Modulation of Population 463 Correlations in Auditory Cortex Is Associated with Decreased Auditory Detection 464 Performance in Old Mice. J Neurosci 42, 9278-9292 (2022). 465 466 467 Acknowledgments 468 We thank members of the Kanold lab for their comments on the manuscript. 469 We thank Dr. Behtash Babadi for his advice regarding interpreting the results of the manuscript. 470 Conflict of interest: The authors declare no competing financial interests. 471 Funding: Supported by NIH R01DC018790 (POK, HKL) 472 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint 24 Author contributions: 473 Conceptualization: POK, HKL, PJ 474 Methodology: PJ, POK 475 Investigation: PJ 476 Visualization: PJ, POK 477 Funding acquisition: POK, HKL 478 Project administration: PJ, POK 479 Supervision: POK 480 Writing – original draft: PJ, POK 481 Writing – review & editing: PJ, HKL, POK 482 483 Data and materials availability: All data supporting the findings from this study will be available 484 upon publication at the Johns Hopkins Data Archive (https://archive.data.jhu.edu). 485 Supplementary Materials 486

487 Supplementary Figures S1-S6 488 Supplementary Table S1 489 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.02.592252doi: bioRxiv preprint