Fig 1.
Large-scale recordings of neuronal activity in head-fixed mice.
A. Schematic of the experimental setup. B. Schematic of the ascending auditory pathway. Each area where sound responses have been obtained via electrophysiology or 2-photon imaging is colored, with indication of the recording method and the number of single fibers/neurons. C. Sample spectrograms from each of the 8 categories contained in the 307 sounds stimulation set. D-F. Methodologies of data collection for each area. D. i. Schematic of the targeting of the cochlear nucleus using Neuropixels 1.0 probes ii. assessed via post-mortem histology. iii. Waveforms, auto-correlograms, iv. and raster plots of responses to a 10 Hz AM of 3 example units. E. Same as D for inferior colliculus dataset collection. F. i. Schematic of the imaging of the auditory cortex ii. and example image of 4 planes recording. iii. Raw ΔF/F trace (black) and deconvolved trace (blue) in one example neuron from the imaging session in (i, ii). iv. Deconvolved response of 3 example neurons in response to a 10 Hz AM. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 2.
Single cell sound response samples across the auditory system.
Trial-averaged responses of example neurons from cochlear nucleus (CN, 5 neurons), inferior colliculus (IC, 5 neurons), and auditory cortex (AC, 7 neurons) to 12 sounds with spectral content at 12 kHz (2 pure tones, 2 ramps, 1 chord, 2 AMs, 2 chirps, 1 WN and 2 complex, represented with their spectrograms). Sound presentation periods are shaded in gray. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 3.
Decorrelation of sound representations across the auditory system identified with a generic population tuning measure.
A. Synthetic tuning curves representing the time-averaged firing rate responses of 16 modeled neurons to seven pure tones. B. Construction of the population similarity analysis matrix displaying the similarity of the population representations of the seven pure tones for the synthetic measures shown in A. C. For the same synthetic data, population tuning curves to 9 kHz (top) and for different frequency ratios (measured in octaves) as extracted from the population similarity matrix in B. D–F. Illustration of three advantages of population tuning measures against single cell measurements. D. While single cell measurements of tuning properties such as best frequency (BF) and frequency tuning band-width (BW) are imprecise and biased because of response variability, the effect of noise on population tuning measures can be corrected. E. While single cell tuning measures often depend on a response model (e.g., single frequency peak), population tuning provides a generic description of the similarity relationships between all sounds (e.g., all pure tone frequencies). F. Population tuning is a unified measure that generalizes across any sound feature or any sound. G. Matrices of spatial representation similarity for all regions. H. Average spatial representation similarity between all pairs of sounds computed from G for all regions. (Mean ± SEM: CN = 0.53 ± 1e − 3, IC = 0.41 ± 1e − 3, AC = 0.21 ± 1e − 3. Two-sample Wilcoxon sign-rank test for paired distribution between pairs of sounds across regions: CN against IC, p < 1e − 63, IC against AC, p < 1e − 63). I. Average population response of neurons to distinct categories of sounds, after standardization of each neuron by its maximum response and of the population by the mean response to complex sounds at 70 dB SPL. CN, cochlear nucleus, IC, inferior colliculus, AC, auditory cortex. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 4.
Identity and intensity tuning improves differentially across the auditory system for simple and complex sounds.
A. Spatial representation similarity matrices for pure tones at 50- and 70-dB SPL. B. Evolution of spatial representation similarity between pure tones dependent on their frequency difference. C. Spatial representation similarity matrices for complex sounds at 50- and 70-dB SPL. D. Spatial representation similarity between different complex sounds with different identity (i.e., dolphin vs. bird call) at the same average intensity. E. Spatial representation similarity between pure tones at the same frequency but at different intensities. F. Spatial representation similarity between identical complex sounds at different intensity. CN, cochlear nucleus, IC, inferior colliculus, AC, auditory cortex. Full statistics are provided in S1 Table. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 5.
Early emergence of amplitude modulation tuning.
A. Spatial representation similarity matrices for amplitude modulated (AM) sounds at 8 different carrier signals (pure tones and chords) and 6 modulation frequencies. B. Evolution of spatial similarity between AM sounds dependent on their modulation frequency difference. C. Similarity of the spatial representations of AM sounds and of the summed spatial representations of the pure tones corresponding to the carrier signal. D. Spatial representation similarity matrices for upward and downward linear intensity ramps (carrier signal = pure tone). E. Spatial representation similarity between upward and downward ramps at the same frequency. F. Spatial representation similarity between ramps and pure tones that have the same frequency and same start (left) or end (right) intensity as the ramp. CN, cochlear nucleus, IC, inferior colliculus, AC, auditory cortex. Full statistics are provided in S2 Table. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 6.
Decorrelation of broad-band and multi-frequency sounds.
A. Spatial representation similarity matrices of pure tones at 70 dB SPL and their combination into various chords, organized based on the frequency range covered (low, mid, high, or full range) and on whether they are harmonic (harmo). B. Spatial representation similarity between chords built from the same pool of pure tones (i.e., between all pairs of chords built from low, mid, high frequency pure tones but not between a low frequency chord and a mid frequency chord). C. Similarity between spatial representations of chords and the summed representations of the pure tones that compose them. D. Spatial representation similarity matrices of pure tones, narrow- and broad-band noises. E. Mean spatial representation similarity between broadband noises dependent on their difference in bandwidth across recorded areas. F. Schematic of the reconstruction of a 4.8-28 kHz broadband noise using spectrograms of sounds used G. Schematic of the reconstruction of a 12 + 25 kHz narrowband noise using spectrograms of sounds used. H. Similarity of spatial representations between broadband (left) or narrowband (right) noises and the summed response of pure tones included in their frequency range. CN, cochlear nucleus, IC, inferior colliculus, AC, auditory cortex. Full statistics are provided in S3 Table. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 7.
Cortical decorrelation of sound direction.
A. Spatial representation similarity matrices of frequency modulated chirps at 70 dB SPL and pure tones at 70 dB SPL. B. Similarity of spatial representations between time-symmetric up- and down-frequency sweeps. C. Similarity of spatial representations of sweeps and of the summed spatial representations of pure tones traversed by the sweep. D. Spatial representation similarity matrices of forward and backward complex sounds at 70 dB SPL. E. Spatial representation similarity between forward and backward renditions of the same complex sound. CN, cochlear nucleus, IC, inferior colliculus, AC, auditory cortex. Full statistics are provided in S4 Table. The source data have been uploaded to https://doi.org/10.5281/zenodo.14421103.
Fig 8.
Evolution of sound feature tuning along the auditory system.
Sketch representing population tuning strength across 4 different stages of the auditory (1 − similarity) for different acoustic features from pure frequency tones to amplitude (AM) or frequency (FM) modulations to complex sounds. The double arrow indicates a possible link between tuning to AM and to noises. id. = identity; direct. = direction; freq. = frequency.