computational-auditory-periphery-models

# Computational Auditory Periphery Models: the Return of the Rodent ## Overview Cross-species computational models of the auditory periphery bridge the gap between non-invasive human diagnostics and experimental evidence from animal studies. This work adapts a 1-D nonlinear cochlear transmission-line (TL) model originally designed for the human auditory periphery to mouse and gerbil, enabling a single computational framework for cross-species research on sensorineural hearing loss (SNHL). ## Key Contributions 1. **Species-Specific Parameterization**: Adapted anatomical and physiological parameters — including basilar membrane (BM) length and width, stapes area, middle-ear transfer functions, and frequency range — to match each species' auditory periphery and hearing range. 2. **Cross-Species Validation**: Validated against experimental BM velocity level-growth characteristics, auditory-nerve (AN) tuning curves, and DPOAEs (distortion product otoacoustic emissions). 3. **Cochlear Synaptopathy Simulation**: Reproduced observed differences in recorded auditory brainstem responses (ABR) and envelope following responses (EFR) from mice and gerbils with SNHL. 4. **Model Limitations Identified**: OHC individualization based on DPOAEs failed to faithfully reproduce individual measurements, though intergroup differences in OHC damage were captured. ## Core Methodology ### Transmission-Line (TL) Model The TL model is based on a time-domain implementation of Zweig's description of cochlear admittance, grounded in local scaling symmetry. Key equations: - **Series impedance**: Z_{s_n}(s) = ω_n M_{s0} s - **Shunt admittance (BM admittance)**: Y_{p_n}(s) = 1/Z_{p_n}(s) = s[ω_n M_{p0}(s² + δn + 1 + ρ_n e^{-μ_n s})]⁻¹ where `s = iω/ω_c` is normalized by CF, `n` is the cochlear section number (1–1000), and δ, ρ, μ specify damping, stiffness, and delay. ### Cochlear Nonlinearity Zweig's TL linear equations are extended to a nonlinear version by dynamically shifting the double-pole α* in the s-plane. This shift broadens filters in response to increasing stimulus intensity while maintaining BM velocity zero-crossings. Key parameters per CF: 1. **Active pole α*_A**: Sharpest filters at low stimulation 2. **Passive pole α*_P**: Broader filters at high stimulation or post-mortem 3. **Compression slope C**: 0.31 dB/dB for human model 4. **Compression threshold**: Level at which BM velocity begins compressive growth ### Species Translation Species-specific adjustments include: | Parameter | Human | Gerbil | Mouse | |-----------|-------|--------|-------| | BM Length | ~35 mm | ~11 mm | ~7 mm | | BM Width | Variable | Narrower | Narrowest | | Frequency Range | 20–20,000 Hz | 0.1–50 kHz | 1–80 kHz | | Stapes Area | ~3.2 mm² | ~0.8 mm² | ~0.5 mm² | | AN Fiber Count | ~30,000 | ~24,000 | ~12,000 | ### Auditory Nerve and Brainstem Modeling AN synapse model computes vesicle release and firing probability → single-fiber firing rate. ANFs divided into three subtypes by spontaneous rate (LSR=1, MSR=10, HSR=68.5 spikes/s). Responses summed across CFs and passed to ABR generators (cochlear nucleus, inferior colliculus) to model EFRs. ### Hearing Loss Simulation - **OHC Loss**: Reduce mechanical cochlear gain in TL model - **Cochlear Synaptopathy**: Remove subtypes of ANFs ## Key Results 1. Simulated AN outputs reasonably matched empirical AN thresholds and frequency selectivity 2. Discrepancy larger for cochlear sections near base or apex 3. Cochlear synaptopathy simulations reproduced species-specific ABR/EFR differences 4. OHC individualization via DPOAEs limited in reproducing individual measurements but effective for intergroup differences ## When to Use Use this skill when: - Building or adapting computational models of the auditory periphery - Conducting cross-species hearing research (human ↔ rodent translation) - Simulating sensorineural hearing loss or cochlear synaptopathy - Validating auditory models against physiological data (BM velocity, AN tuning, DPOAEs, ABR) ## Related Skills - [[computational-neuroscience-in-llm-era]] - General computational neuroscience - [[multi-scale-info-geometry-neural]] - Information geometry approaches ## References - Thienpont, M. et al. (2026). Computational Auditory Periphery Models: the Return of the Rodent. arXiv:2605.19070v2 [q-bio.NC] - Zweig, G. (1991). Finding the impedance of the organ of Corti. JASA. - Verhulst, S. et al. (2012, 2018). Nonlinear time-domain cochlear model. Hearing Research. - Shera, C.A. (2001). Frequency glides in click-evoked otoacoustic emissions. JASA. - Altoè, A. et al. (2014, 2018). Cochlear transmission-line models. JASA.