Impaired prepulse inhibition in APP/PS1 mice is accompanied by substantial morphological changes in neurons of the central auditory system and hippocampus

Impaired prepulse inhibition in APP/PS1 mice is accompanied by substantial morphological changes in neurons of the central auditory system and hippocampus Jana Svobodov´ a Burianov´ a a,** , Daniela ˇ Cernotov´ a b , Tereza Klausov´ a b , Oliver Profant a , Jakub Fuksa a , Josef Syka a , Jan Svoboda b,* a Department of Auditory Neuroscience, Institute of Experimental Medicine of the Czech Academy of Sciences, Vídeˇ nsk´ a 1083, Praha 4 142 00, Czech Republic b Laboratory of Neurophysiology of Memory, Institute of Physiology of the Czech Academy of Sciences, Vídeˇ nsk´ a 1083, Praha 4 142 00, Czech Republic ARTICLE INFO Keywords: Alzheimer’s disease APP/PS1 Prepulse inhibition Morphometry ABSTRACT Auditory dysfunction is increasingly recognized as a non-cognitive feature of Alzheimer’s disease (AD). We examined auditory processing and neuronal morphology in APPswe/PSEN1dE9 (APP/PS1) transgenic mice, a model of AD. Ten-month-old male APP/PS1 and wild-type (WT) littermates were tested for auditory thresholds using auditory brainstem responses (ABR) and for sensorimotor gating using prepulse inhibition (PPI) of the acoustic startle reflex. ABR thresholds did not differ between groups across tested frequencies (2–16 kHz), indicating preserved peripheral hearing. In contrast, APP/PS1 mice showed significantly impaired PPI at 4, 12, and 20 kHz prepulses and displayed exaggerated startle responses to high-intensity stimuli. Exploratory and anxiety-like behavior, assessed in the open field and elevated plus maze, did not differ between groups. Morphological analysis of Golgi–Cox-stained neurons revealed widespread dendritic pathology in the inferior colliculus, medial geniculate body and auditory cortex, as well as in hippocampal CA1. Compared with WT, APP/PS1 neurons exhibited shorter dendrites, reduced branching, and lower spine density, accompanied by markedly decreased dendritic complexity in Sholl analyses. These findings demonstrate that sensorimotor gating deficits in APP/PS1 mice are accompanied by degeneration in central auditory and hippocampal circuits; however, a contribution of peripheral degeneration cannot be excluded. The data highlights the vulnerability of auditory midbrain and thalamic structures in this model of AD and suggests that dendritic alterations along the auditory pathway may contribute to central auditory dysfunction and serve as potential early biomarkers of disease. 1. Introduction Alzheimer’s disease (AD) is characterized not only by progressive cognitive decline but also by a range of non-cognitive symptoms, including auditory dysfunction. The APPswe/PSEN1dE9 (APP/PS1) mouse line is a widely used model that develops amyloid deposition at 4–6 months of age, accompanied by neuronal and dendritic alterations (Jankowsky et al., 2001) and spinal motor neurons (Garcia-Alloza et al., 2006). In these mice, reduced amplitudes of auditory cortex (AC) evoked responses have been reported, together with early accumulation of amyloid precursor protein (APP) in the AC as early as 2 months of age (Mei et al., 2021). However, reports on auditory thresholds in APP/PS1 mice are inconsistent: some studies describe progressive hearing loss (Liu et al., 2020), whereas others found no significant impairment (Na et al., 2023). Prepulse inhibition (PPI) of the acoustic startle response is a sensitive measure of auditory system integrity and sensorimotor coordination. The startle reaction is mediated by a neural circuit involving the ** Corresponding author at: Laboratory of Auditory Neuroscience, Institute of Experimental Medicine of the Czech Academy of Sciences, Vídeˇ nsk´ a 1083, Prague 142 00, Czech Republic. * Corresponding author at: Laboratory of Neurophysiology of Memory, Institute of Physiology of the Czech Academy of Sciences, Vídeˇ nsk´ a 1083, Prague 142 00, Czech Republic. E-mail addresses: [email protected] (J. Svobodov´ a Burianov´ a), [email protected] (J. Svoboda). Contents lists available at ScienceDirect Hearing Research journal homepage: www.elsevier.com/locate/heares https://doi.org/10.1016/j.heares.2025.109484 Received 2 September 2025; Received in revised form 19 November 2025; Accepted 19 November 2025 Hearing Research 469 (2026) 109484 Available online 20 November 2025 0378-5955/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). auditory nerve, cochlear root neurons, the caudal pontine reticular formation, and spinal motor neurons, and represents a motor response of the body to a sudden, intense acoustic stimulus. This startle reaction can be modulated by sensory stimulation that precedes the startle-eliciting stimulus. When the preceding (prepulse) stimulus is auditory, the inferior colliculus also participates in the modulation of the response (for review, see Lauer et al., 2017; Gom´ ez-Nieto et al., 2020). A reduction in PPI reflects impaired sensorimotor gating, which normally suppresses excessive behavioral responses to salient stimuli. The hippocampus and entorhinal cortex are critically involved in regulating PPI, as well. Wang et al. (2012) reported reduced PPI in 7and 22-month-old APP/PS1 mice compared with age-matched controls, while startle responses to pulse-alone stimuli were comparable between groups. In contrast, Ewers et al. (2006) found reduced PPI only in APP/PS1 mice with a high amyloid burden in the cortex. These discrepancies motivated us to investigate whether PPI of the startle response is consistently altered in APP/PS1 mice compared with controls. Morphological analyses of APP/PS1 mice have demonstrated the presence of senile plaques and cerebral amyloid angiopathy (Garcia-Alloza et al., 2006). Subsequent studies described swollen dystrophic neurites, loss of dendritic spines, and reduced dendritic area in these animals (Moolman et al., 2004). Three-dimensional analyses of dentate gyrus spines further revealed a marked decrease in the frequency of large spines, even in plaque-free regions (Knafo et al., 2009). Despite these observations, no systematic study has examined dendritic morphology in APP/PS1 mice, particularly within central auditory structures. Given the functional changes reported in the auditory system of this model, we hypothesized that structural alterations may also be present in the inferior colliculus (IC), medial geniculate body (MGB), and auditory cortex (AC), as well as in the hippocampus. Consistent with this hypothesis, we found pronounced reductions in dendritic arborization across these regions, in parallel with impaired sensorimotor integration reflected in reduced PPI. In addition, exploratory and anxiety-like behaviors were assessed in the open field and elevated plus maze as complementary non-auditory tasks. 2. Methods 2.1. Animals APPswe/PSEN1dE9 (APP/PS1) transgenic mice were maintained as hemizygotes and crossed with C57BL/6 J ×C3H wild-type mice to generate experimental animals. APP/PS1 offspring carried one copy each of the human APP swe and PSEN1 dE9 transgenes, whereas wild-type littermates lacked both transgenes and served as controls. Genotyping was performed to identify carriers of both mutant alleles (APP swe and PSEN1 dE9 ), referred to as APP/PS1 mice, and non-carriers, referred to as wild-type (WT) littermates. Only male mice aged 10 months were used in this study. For behavioral testing, 10 WT and 9 APP/PS1 mice were included; one APP/PS1 animal was excluded from elevated plus maze analysis due to technical issues. For auditory brainstem response (ABR) recordings, 9 WT and 9 APP/PS1 mice were tested, as one WT animal was not evaluated. For morphological analyses, brains from 10 WT and 9 APP/PS1 mice were processed, except for the inferior colliculus, where usable data were obtained from 9 WT and 9 APP/PS1 animals due to tissue damage in one WT brain. Animals were housed in groups of 2–5 in an accredited facility at the Institute of Physiology of the Czech Academy of Sciences, under controlled temperature and humidity, with a 12/ 12 h light/dark cycle (lights on at 6:00 AM). All experiments were approved by the local Animal Care Committee (Project No. 51–2022-P) and conducted in accordance with the Animal Protection Code of the Czech Republic and EU Directive 2010/63/EU. 2.2. The open field and elevated plus maze apparatuses Exploratory and anxiety-like behaviors were assessed in the open field (OF) and elevated plus maze (EPM) tests. The OF apparatus was a square wooden chamber (45 ×45 ×30 cm) with white walls and floor. Mice were placed in the center and allowed to explore freely for 5 min. The EPM consisted of four white plexiglass arms (each 30 cm long, 5 cm wide) elevated 70 cm above the floor. Two opposing arms were open, while the other two were enclosed by 16 cm high opaque walls. Each mouse was placed in the central platform and allowed to explore for 5 min. Illumination was set to 10 lx. Behavior was recorded at 30 FPS using an overhead video camera (Logitech). 2.2.1. Data analysis In the OF test, the total distance traveled and the time spent in the central zone (defined as 70 % of the arena area) were quantified. In the EPM, total distance traveled and time spent in the open arms, including the central platform (5 ×5 cm), were measured. Behavioral sessions were analyzed using EthoVision XT 17.5 (Noldus Information Technology, Wageningen, The Netherlands). One APP/PS1 mouse was excluded from EPM analysis due to recording failure. 2.3. Prepulse inhibition test Prepulse inhibition (PPI) was assessed using a startle reflex system (Med Associates Inc.). The standard short protocol (6 min) provided by the manufacturer was employed, consisting of acoustic startle stimuli presented with or without prepulses. The startle stimulus was a 10 ms white noise burst at 100 dB. Prepulses were 10 ms pure tones (4, 12, or 20 kHz) at 70 dB, presented 100 ms before the startle stimulus. Each session began with a 60 s habituation period without stimulation, followed by 55 trials separated by randomized inter-trial intervals (3–8 s). In block 1, startle-alone trials (16) and startle +prepulse trials (8 per frequency) were presented in random order. In block 2, startle stimuli were delivered at varying intensities (70, 80, and 90 dB; 5 trials each) to assess startle amplitude as a function of intensity. PPI was expressed as the percentage reduction of the startle response in prepulse trials relative to startle-alone trials using the formula: PPI ( %) =[1 – (prepulse trial / startle-alone trial)] ×100. 2.4. ABR auditory thresholds Auditory thresholds were determined using auditory brainstem responses (ABRs) in mice anesthetized with isoflurane. ABRs were recorded in a sound-attenuated room, with body temperature maintained at 37 ◦C using an electrically controlled heating pad. Stainless steel needle electrodes were positioned subcutaneously: one over the mastoid ipsilateral to the stimulated ear, one at the vertex (Cz, positive), and a ground electrode in the neck. Recordings were obtained with an eABR system (BioMed Jena GmbH) and Sennheiser Momentum M2 headphones calibrated to IEC 318–4 standards. Tone pips (2, 4, 8, and 16 kHz; 5–10 dB steps) were presented. Pure tone bursts were used for acoustic stimulation, defined by a 2–3–2 cycle envelope (two cycles rise, three cycles at full amplitude, and two cycles decay); therefore, the stimulus rise time varied with stimulation frequency. Stimuli were delivered via earphones coupled to an ear-canal tube. The acoustic system was calibrated using a Brüel & Kjær 4134 microphone. The repetition rate was 21.3 Hz, with a 10 ms data acquisition window. Both ears were stimulated separately. Signals were band-pass filtered (180 Hz–1.8 kHz), averaged 200–400 times, and analyzed with eAudio software. 2.5. Histology Immediately after ABR recordings, mice were overdosed with isoflurane and transcardially perfused with heparinized saline (5 ml/L). Brains were removed and processed using the FD Rapid GolgiStain™ kit (FD NeuroTechnologies, USA) following the manufacturer’s instructions, as described previously (Svobodov´ a Burianov´ a and Syka, J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 2 2020). Coronal sections (180 µm) were cut on a freezing microtome (Microm HM400R, Germany), mounted on gelatin-coated slides, air-dried, developed, dehydrated, cleared in xylene substitute, and coverslipped with Permount (Fisher Scientific). 2.6. Neuronal morphometry Sections containing the IC, MGB, primary AC, and ventral hippocampal CA1 were identified according to Franklin and Paxinos (2007). IC sections were sampled from bregma −5.0 to −5.3 mm, MGB from −3.1 to −3.3 mm, AC from −2.8 to −3.1 mm, and CA1 from −3.4 to −3.6 mm. The IC was divided into the dorsal cortex (DIC), external cortex (EIC), and central nucleus (CIC), and the MGB into dorsal (MGB-D) and ventral (MGB-V) subdivisions. Tracing was performed under bright-field microscopy (Leica DMRXA) using a video camera coupled to a high-resolution monitor (Wacom, 3840 ×2160 px) and Neurolucida software (version 2021.1.3, MicroBrightField, USA). For each animal, 15 IC neurons (5 per subdivision), 10 MGB neurons (5 per subdivision), and 10 pyramidal neurons each from AC and CA1 were reconstructed in 3D Pyramidal cells were identified by the presence of a basal dendritic tree and a distinct apical dendrite. To minimize truncation artifacts, only neurons with somata located in the middle third of the section were analyzed. Inclusion criteria required well-impregnated neurons with at least tertiary branches, minimal overlap with neighboring stained cells, and absence of major staining artifacts. Tracing was performed manually by one experimenter blind to group identity. From 3D reconstructions of Golgi-stained neurons, the following morphometric parameters were quantified: soma area, dendritic volume, dendritic length (apical and basal), number of primary dendrites, branch points (nodes), dendritic terminals (ends), and spine counts and density. Dendritic complexity was further evaluated using Sholl analysis. All procedures followed our previously published protocol (Svobodov´ a Burianov´ a and Syka, 2020). 2.7. Statistical analysis All statistical analyses were performed in GraphPad Prism 10 (GraphPad Software, San Diego, CA). For neuronal morphometry, five neurons were randomly sampled from each subdivision of the IC and MGB, and ten pyramidal neurons from the AC and CA1. Neuronal parameters were averaged per animal, and these animal means served as the unit of analysis. Cell body area, number of varicosities, and dendritic parameters (total length, volume, dendrite number, branch points/nodes, and terminal ends) were analyzed with two-way mixed ANOVAs with Group (WT vs. APP/PS1; between-subjects) and Subdivision (IC: CIC, DIC, EIC; MGB: dorsal, ventral; within-subjects). Spine density was analyzed in the MGB, AC, and CA1 using the same statistical design. For AC and CA1, the same parameters were analyzed with Group (between-subjects) and Dendrite type (apical vs. basal; within-subjects). Significant effects or interactions were followed by Sidak post hoc tests (WT vs. APP/PS1 within each subdivision/type). The number of basal primary dendrites, which could not be included in the two-way design, was compared between groups using Welch’s unequal-variance t-test. Sholl analyses of dendritic intersections, dendritic length, and spine counts as a function of radius were evaluated using two-way mixed ANOVAs with Group (between-subjects) and Radius (within-subjects), followed by Sidak post hoc tests at each radius; significant ranges are reported. ABR thresholds were analyzed using two-way mixed ANOVAs with Group (WT vs. APP/PS1; between-subjects) and Frequency (2, 4, 8, 16 kHz; within-subjects). ABR wave latencies (I–V) and interwave intervals (I–III, III–V, I–V) were analyzed using two-way mixed ANOVAs with Group (between-subjects) and Wave/Interval (within-subjects) separately at 8 and 16 kHz and at 70 dB SPL and 10 dB above individual threshold. Sidak’s multiple comparisons test was used for post hoc contrasts; significance was set at p <0.05. For PPI, open field, and elevated plus maze data were analyzed using unpaired t-tests or two-way ANOVAs as appropriate. Normality of residuals was verified with Shapiro–Wilk tests, and homogeneity of variance with Levene’s test. Statistical significance was set at p <0.05. 3. Results 3.1. Open field and elevated plus maze Behavioral analysis using unpaired t-tests revealed no significant differences between APP/PS1 and WT groups in all measured parameters (Fig. 1). In the open field test, the total distance traveled was comparable between APP/PS1 (1595 ±679.9 cm) and WT (1563 ± 332.9 cm) mice, with no significant difference observed (t₁₇ =0.14, p = 0.894). The time spent in the center zone also showed no significant difference between APP/PS1 (139.1 ±67.68 s) and WT (150.5 ±45.5 s) mice (t₁₇ =0.44, p =0.667). Similarly, assessment in the elevated plusmaze showed comparable performance between groups, with APP/PS1 (1186 ±432.3 cm) and WT (1129 ±442.2 cm) mice showing no significant differences in total distance traveled (t₁₆ =0.28, p =0.786) or time spent in the open arms and center (APP/PS1: 73.84 ±42.72 s vs. WT: 65.98 ±47.56 s; t₁₆ =0.36, p =0.721). 3.2. Prepulse inhibition of the acoustic startle reflex APP/PS1 mice showed a pronounced disruption of sensorimotor gating (Fig. 2). PPI was significantly reduced at all prepulse frequencies tested. For example, with a 4 kHz prepulse, APP/PS1 mice exhibited markedly lower PPI than WT controls (t(12.8)=4.33, p =0.0008). Similar reductions were found at 12 kHz (t(14.9)=4.38, p =0.005) and 20 kHz (t(12.6)=3.63, p =0.003). APP/PS1 mice also displayed exaggerated startle responses to pulsealone stimuli. Startle amplitude increased more steeply with stimulus intensity than in WT mice, as confirmed by a significant group ×intensity interaction (F(3,36)=4.97, p =0.005). 3.3. Auditory brainstem response thresholds Auditory thresholds, assessed by ABR recordings (Fig. 3), did not differ between APP/PS1 and WT mice at any tested frequency (2–16 kHz). Two-way mixed ANOVA revealed no significant main effect of Group or Group ×Frequency interaction (all p >0.3), confirming preserved peripheral auditory sensitivity in both genotypes. Analysis of ABR wave latencies and interwave intervals (I–III, III–V, I–V) at 8 and 16 kHz (70 dB SPL and 10 dB above individual threshold) showed no significant main effects or interactions (all p >0.05; Fig. 4; Supplementary Table S5). APP/PS1 mice displayed small, nonsignificant prolongations of late components (waves IV–V) and longer III–V and I–V intervals at 16 kHz/70 dB. 3.4. Neuronal morphology Golgi–Cox staining and reconstruction revealed consistent morphological alterations across all analyzed brain structures, including the inferior colliculus, medial geniculate body, auditory cortex, and hippocampal CA1. APP/PS1 mice uniformly exhibited reduced dendritic complexity compared with WT, with differences detected in virtually all parameters assessed. Representative examples of reconstructed neurons are shown in Figs. 5–8. For clarity, we refer to primary dendrites, branch points, and terminals as dendrite number, dendritic nodes, and dendritic ends in the Results and figure legends. Inferior colliculus (IC). APP/PS1 neurons exhibited simplified dendritic arbors across IC subdivisions (Fig. 9). Total dendritic length was J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 3 markedly shorter (F(1,16)=40.47, p <0.0001), and dendrite number was also reduced (F(1,16)=10.89, p =0.0045). Number of nodes and terminal ends showed Group ×Subdivision interactions (nodes: F (2,32)=4.98, p =0.013; ends: F(2,32)=3.94, p =0.030), with post-hoc tests confirming significant WT>APP/PS1 differences in multiple subdivisions. Sholl analyses revealed extensive reductions in dendritic complexity: WT>APP/PS1 at 20–130 µm in CIC, 30–120 µm in DIC, and 20–150 µm in EIC (intersections and length, all adj. p <0.05). Medial geniculate body (MGB). Dendritic degeneration was evident across both dorsal (MGB-D) and ventral (MGB-V) subdivisions (Fig. 10). Total dendritic length was reduced by ~68 % in APP/PS1 mice (WT 1197 µm vs. APP/PS1 385 µm; F(1,17)=146.3, p <0.0001), and spine density was halved (WT 0.331 vs. APP/PS1 0.158 µm⁻¹; F(1,17)=48.1, p <0.0001). Node numbers and terminal ends were also strongly reduced (nodes: F(1,17)=100.3, p <0.0001; ends: F(1,17)=157.3, p <0.0001). Dendrite number was significantly lower in both subdivisions (MGB-D p <0.0001; MGB-V p <0.0001, Sidak post-hocs). Sholl analyses confirmed uniform deficits: WT>APP/PS1 across 20–140 µm in MGB-D and 20–150 µm in MGB-V (intersections and length, all adj. p <0.05). Auditory cortex (AC). APP/PS1 pyramidal neurons displayed shortened dendritic trees in both apical and basal compartments (Fig. 11). Total dendritic length was significantly reduced (F(1,17)=94.4, p < 0.0001), with a Group ×Type interaction (F(1,17)=7.95, p =0.012) reflecting greater loss in basal trees (WT basal 767 µm vs. APP/PS1 basal 332 µm). Dendrite nodes were fewer in APP/PS1 (F(1,17)=57.3, p < 0.0001), while terminal ends were strongly reduced (F(1,17)=59.8, p < 0.0001), with both apical and basal showing significant differences (Sidak p <0.0001 each). Spine density was lower in APP/PS1 (F(1,17)= 22.8, p =0.0002), with an apical>basal effect (F(1,17)=17.0, p = 0.0007). The number of basal primary dendrites was also reduced (Welch’s t(9.7)=5.90, p =0.0002; WT 5.0 vs APP/PS1 3.3). Sholl analyses indicated reduced intersections at 30 and 70–150 µm (apical) and 20–130 µm plus 150 µm (basal), and shortened per-radius length at 30–50 and 70–140 µm (apical) and 20–150 µm (basal). Hippocampal CA1. APP/PS1 pyramidal neurons showed profound dendritic degeneration in both apical and basal trees (Fig. 12). Total dendritic length was reduced by more than half (apical: WT 1086 vs. APP/PS1 466 µm, p <0.0001; basal: WT 1141 vs. APP/PS1 340 µm, p < 0.0001). Dendrite nodes were fewer (apical: WT 9.47 vs. APP/PS1 3.53, p <0.0001; basal: WT 7.90 vs. APP/PS1 2.77, p <0.0001), and terminal ends were also reduced (apical: WT 10.69 vs. APP/PS1 4.57, p <0.0001; basal: WT 12.96 vs. APP/PS1 5.96, p <0.0001). Spine density was halved in both compartments (apical: WT 0.464 vs. APP/PS1 0.205 µm⁻¹, p <0.0001; basal: WT 0.449 vs. APP/PS1 0.203 µm⁻¹, p <0.0001). The number of basal primary dendrites was also significantly reduced (t (13.35)=2.85, p =0.0135). Sholl analyses demonstrated broad deficits: WT>APP/PS1 at 20–150 µm for basal intersections, 30–150 µm for basal length, 40–150 µm for apical intersections, and 50–150 µm for apical length. Analyses of cell body area, varicosities counts, dendritic parameters (including dendritic volume), and Sholl analysis of spine counts across all structures are presented in the Supplementary File. Fig. 1. Open field and elevated plus maze performance. (A, B) In the OF test, WT and APP/PS1 mice did not differ in total distance traveled (A) or time spent in the center zone (B). (C, D) In the EPM test, no significant differences were found between genotypes in the total distance (C) or time spent (D) in the open arms +center. Data are mean ±SEM; unpaired t-tests. J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 4 4. Discussion Auditory function and related neural morphology were systematically assessed in 10-month-old APP/PS1 mice. ABR thresholds across Fig. 2. Prepulse inhibition (PPI) of the acoustic startle reflex. (A–C) PPI at prepulse frequencies of 4 kHz (A), 12 kHz (B), and 20 kHz (C). Startle pulse intensity was 100 dB, and prepulse intensity 70 dB. WT mice showed robust inhibition of the startle response, whereas APP/PS1 mice exhibited markedly reduced PPI at all frequencies. (D) Startle amplitude as a function of stimulus intensity (70–100 dB). APP/PS1 mice displayed significantly larger startle responses compared with WT. Data are mean ±SEM. **p <0.01, ***p <0.001 (unpaired t-test for PPI; two-way ANOVA with Sidak post hoc test for startle response). Fig. 3. Auditory brainstem response (ABR) thresholds. Hearing thresholds were determined at 2, 4, 8, and 16 kHz in WT and APP/PS1 mice. No significant differences were observed between groups across the tested frequencies. Data are mean ±SEM; two-way ANOVA revealed no significant main effect of genotype or genotype ×frequency interaction. Fig. 4. Latencies of ABR waves I–V and interwave intervals in WT and APP/PS1 mice at 16 kHz stimulation (70 dB). Data are mean ±SD. Group differences were analyzed using two-way mixed ANOVAs followed by Sidak’s multiple comparisons test (p <0.05). No statistically significant differences were detected between genotypes. J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 5 2–16 kHz were comparable to wild-type controls, and wave latencies and interpeak intervals showed only minor, nonsignificant prolongations of late components at 16 kHz. In contrast, prepulse inhibition of the acoustic startle reflex was profoundly reduced, and morphometric analyses revealed extensive dendritic degeneration within the inferior colliculus, medial geniculate body, auditory cortex, and ventral hippocampal CA1. Our data showed only minimal differences in ABR wave latencies and interpeak intervals, with a mild trend toward prolonged wave V latency and extended I–V and III–V intervals at 16 kHz at the suprathreshold level; however, these changes were not statistically significant. Previous reports have yielded variable results. Liu et al. (2020) found prolonged wave II latency and extended I–II and I–V intervals in 3-month-old APP/PS1 mice, whereas Na et al. (2023) observed delayed wave IV latency and an extended I–IV interval in 13-month-old animals—both consistent with our trend toward delayed central Fig. 5. Inferior colliculus (IC) neurons (Golgi–Cox staining and 3D tracings). (A) Atlas view of the inferior colliculus (IC) with its dorsal (DIC), central (CIC), and external (EIC) subdivisions, from which neurons were sampled. (B) Low-magnification image of Golgi–Cox-stained IC neurons. (C–E) Higher magnification images of representative neurons from WT mice in DIC (C), CIC (D), and EIC (E). (F–H) Corresponding images from APP/PS1 mice in DIC (F), CIC (G), and EIC (H). (I–K) Neurolucida tracings of IC neurons from WT mice in DIC (I), CIC (J), and EIC (K). (L–N) Tracings from APP/PS1 mice in DIC (L), CIC (M), and EIC (N). Stained and traced neurons are independent examples. Scale bars: 200 µm (B), 50 µm (C–N). J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 6 components, suggesting reduced neural synchrony and possible demyelination within the central auditory pathway. Similar prolongations were reported by Burke et al. (2024) in the PS19/P301S model. In contrast, Na et al. (2023) described shorter wave I latencies and reduced I–IV intervals in 5xFAD mice, indicating enhanced central gain, while O’Leary et al. (2017) noted minimal latency changes in the same strain. Increased central gain (hyperexcitability) in 5xFAD mice is further supported by decreased wave I and increased wave IV amplitudes (Na et al., 2023). In the APP/PS1 model, Na et al. (2023) reported reduced wave I amplitude without change in wave IV, whereas Liu et al. (2020) observed decreases in waves I, III, and V. Collectively, ABR findings point to peripheral hypoactivity in some AD mouse models and heterogeneous central responses—from slowed conduction to increased central gain—underscoring modeland age-dependent differences in auditory processing; in our APP/PS1 cohort, such central timing effects were subtle and not statistically significant.. Fig. 6. Medial geniculate body (MGB) neurons (Golgi–Cox staining and 3D tracings). (A) Atlas depiction of the medial geniculate body (MGB), showing the dorsal (MGB-D) and ventral (MGB-V) subdivisions targeted for analysis. (B) Low-magnification image of Golgi–Cox-stained neurons in the MGB. (C, E) Higher magnification of Golgi–Cox-stained neurons in MGB-V from a WT mouse (C) and an APP/PS1 mouse (E). (D, F) Higher magnification of neurons in MGB-D from a WT mouse (D) and an APP/PS1 mouse (F). (G–J) Neurolucida tracings of representative neurons: WT (G, H) and APP/PS1 (I, J), corresponding to MGB-V and MGB-D, respectively. Stained and traced neurons are independent examples. Scale bars: 200 µm (B), 50 µm (C–J). J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 7 In terms of ABR-measured hearing thresholds, these appeared to remain intact, contrasting with Liu et al. (2020), who reported elevated thresholds in APP/PS1 mice. By contrast, Na et al. (2023) found no significant ABR threshold differences, consistent with our results, and also reported unchanged DPOAEs—findings that seem more plausible, given the recorded ABR waveforms, than those of Liu et al. (2020). Among the factors that may underlie discrepancies between studies, mouse age appears critical: 3 months in Liu et al. (2020), 10 months in our study, and 13 months in Na et al. (2023). Sex may also contribute, as only males were used here and by Na et al. (2023), whereas Liu et al. (2020) did not report animal sex. Differences in anesthesia represent another variable—our recordings were obtained under isoflurane, Liu et al. (2020) used ketamine–xylazine, and Na et al. (2023) used ketamine–acepromazine and likewise observed no threshold differences. Isoflurane has been reported to elevate auditory thresholds compared with ketamine–xylazine in mice (Verdoodt et al., 2021) and rats (Ruebhausen et al., 2012), although other studies found no effect in either species (Kim et al., 2012; Bielefeld et al., 2014). Because both APP/PS1 and WT mice in our study were tested under isoflurane, any anesthesia-related threshold shift should have affected both groups similarly. This interpretation aligns with Na et al. (2023), who also reported no genotype differences under ketamine–acepromazine. Fig. 7. Auditory cortex (AC) pyramidal neurons (Golgi–Cox staining and 3D tracings). (A) Atlas representation of the auditory cortex (AC), highlighting the primary auditory area where pyramidal neurons were examined. (B) Low-magnification image of Golgi–Cox-stained AC neurons. (C, D) Higher magnification images of individual stained neurons from a WT mouse (C) and an APP/PS1 mouse (D). (E, F) Neurolucida tracings of representative AC pyramidal neurons from a WT mouse (E) and an APP/PS1 mouse (F). The stained and traced neurons are independent examples. Scale bars: 200 µm (B), 50 µm (C, D), 100 µm (E, F). J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 8 Additional methodological differences include stimulus delivery, with earphones used here and by Na et al. (2023) but free-field loudspeakers by Liu et al. (2020). Collectively, these factors—particularly age, which strongly influences auditory function—likely account for the contrasting results, a conclusion further supported by Na et al. (2023). While the primary auditory function of sound detection remained intact, as demonstrated by normal ABRs, a more complex function—- prepulse inhibition (PPI) of the acoustic startle response—was severely impaired. We cannot exclude the possibility that pathology of the peripheral auditory system contributed to the changes observed in APP/ PS1 mice, since our study did not focus on the basic circuitry of the startle reflex and PPI. However, PPI is a complex process involving multiple brain regions including primary auditory cortex, ventral hippocampus, and basolateral amygdala (Rohleder et al., 2016). In our WT mice, PPI reduced the startle amplitude by approximately 50 %, consistent with previous reports (Wang et al., 2012; Cheng et al., 2014; O’Leary et al., 2018). In contrast, APP/PS1 mice showed almost no inhibition, independent of prepulse frequency. Previous studies reported smaller PPI in APP/PS1 compared with controls, but the difference was less pronounced than in our data, likely due to methodological differences. For example, those studies used a 120 dB startle pulse, whereas we used 100 dB, which may have increased the sensitivity of our Fig. 8. Hippocampal CA1 neurons (Golgi–Cox staining and 3D tracings). (A) Atlas schematic of the ventral hippocampal CA1 region; “Py” marks the pyramidal layer, the sampling site of pyramidal neurons. (B) Low-magnification image of Golgi–Cox-stained CA1 pyramidal neurons. (C, D) Higher magnification images of individual stained neurons from a WT mouse (C) and an APP/PS1 mouse (D). (E, F) Neurolucida tracings of representative CA1 pyramidal neurons from a WT mouse (E) and an APP/PS1 mouse (F). The stained and traced neurons are independent examples. Scale bars: 200 µm (B), 50 µm (C, D), 100 µm (E, F). J. Svobodov´ a Burianov´ a et al. Hearing Research 469 (2026) 109484 9