How the brain encodes sounds from the ear
- heloiseherve2
- 1 day ago
- 3 min read
The Central Auditory Circuit Plasticity team at the reConnect Institute at the Hearing Institute (Pasteur Institute/Inserm/CNRS) has revealed how the brain uses multiple neural coding strategies to preserve the reliability of sound information at every stage of the auditory pathway.
Decoding sound from neural activity in mice reveals a striking shift in how the brain encodes sound: from precise but redundant timing codes in early auditory areas to efficient, synergistic rate-based codes in the cortex. This transformation highlights a robust neural strategy for turning complex acoustic input into coherent perception across the auditory pathway.
In an article published in Advanced Science December 10th 2025, Boris Gourévitch, a researcher at the reConnect Institute and his team explain how the brain gradually recodes auditory information from the ear to the first auditory relay in the cortex.
A large-scale and real-time methodological approach
Sound information is converted by the sensory organ of the cochlea, in the inner ear, into an electrical message carried by neurons and therefore intelligible to the brain. “We wanted to understand how, despite environmental variability, since we never really hear the exact same sound twice, our perception remains so stable. This led us to study how neuronal coding at each relay between the ear and the brain could also remain stable in the face of sound perturbations”, explains Boris Gourévitch, lead author of the study. To investigate this, the team monitored the activity of auditory neurons in the brain of awake mice in real time, resulting in the recording and large-scale simulation of over 4,000 neurons across the auditory system. To analyze this large dataset, the team designed a new, simple and universal algorithm called Neural Coding Reliability (NCR), which decodes the auditory information contained in each neuron or set of neurons without resorting to machine learning.
Thanks to this pioneering algorithm, it is possible to reconstruct a sound by observing the responses of a set of neurons, and to analyze how this coding ability is refined throughout the different stages of the auditory system.
A neural language that adapts and becomes collaborative
The study revealed that in the auditory nerve, just behind the ear, neurons transmit information with great temporal precision, but they all respond in a similar way. This redundancy makes the coding robust to small sound variations.
As the signal travels through the brain to the inferior colliculus, the thalamus, and then the auditory cortex, the coding evolves into a strategy based on the total number of neuron impulses, also known as the discharge rate, and on the activation of neural circuits rather than individual neurons. This new neural coding is qualified as synergistic because it is based on the cooperation of multiple neurons to compensate for the lower precision of individual neuron. The team also showed that the absence of activity, or neural silence, carries significant information in these neural circuits, revealing a previously unknown dimension of auditory coding.
The team also found that the absence of activity, or neural silence, carries significant information in these neural circuits, revealing a previously unknown dimension of auditory coding.
Implications for hearing and sensory comprehension in general
These findings improve our understanding of sensory perception circuits at an individual level. Technologies for better diagnosis and treatment of hearing loss could eventually be inspired by this progressive neural recoding.
Beyond hearing, the strategy of studying neural responses to sound disturbances can be applied to the perception of complex sounds such as speech, and even to other sensory modalities, paving the way for new decoding algorithms applicable to EEG. The development of simple, high-performance algorithms that are less sensitive to noise and based only on neural activity could also inspire future hearing neuroprostheses and speech recognition technologies.
This work, providing multiple applications, was made possible thanks to the computational resources of the Pasteur cluster and the support of the French National Research Agency, the Fondation Pour l'Audition, and the Institut Pasteur.
Source
A. Buck, T. Dupont, R. Andrews Cavanagh, et al. “Neural Response Reliability as a Marker of the Transition of Neural Codes along Auditory Pathways” Adv. Sci. (December 10th 2025): e08777. https://doi.org/10.1002/advs.202508777
