Charly Lamothe
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Charly Lamothe has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Pascal Belin
Institut de Neurosciences de la Timone, Aix-Marseille-University, Marseille, Provence-Alpes-Côte d'Azur, UMR7289 CNRS, France
5 shared publications - Brice Bathellier
Centre National de la Recherche Scientifique
3 shared publications - Bruno L. Giordano
La Timone Neuroscience Institute UMR 7289, CNRS, Aix-Marseille University
3 shared publications - Etienne Thoret
Aix-Marseille University, CNRS, INT, UMR7289 1 , F-13005 Marseille
3 shared publications - Julien Sein
La Timone Neuroscience Institute UMR 7289, CNRS, Aix-Marseille University
3 shared publications - Régis Trapeau
Aix-Marseille Université, UMR 7289 CNRS, Institut de Neurosciences de la Timone, Marseille, Provence-Alpes-Côte d'Azur, 13005, France
3 shared publications
Latest Publications
- Open Access
A biomimetic deep encoder for auditory restoration by cortical stimulation
Cochlear and auditory brainstem implants are the primary treatments for profound deafness 1 . However, the inherently low resolution of electrical stimulation used by these devices limits information transfer to a small number of independent stimulation channels 2 and results in perceptual difficulties. An alternative approach is to target the auditory cortex, a brain region large enough to accommodate hundreds of stimulation channels and whose stimulation evokes sound percepts 3,4 . Although modern brain interfaces offer large channel counts and spatial resolution, developing encoding algorithms capable of accurately transmitting interpretable sound information remains challenging. Here we introduce Braincodec, a deep network that optimizes both sound information transmission and the alignment of stimulation patterns with the spatial, temporal, and high-level properties of the auditory cortex neural code. Braincodec operates online and requires only a few hundred channels to transmit sufficient information for the reconstruction of natural sounds with high fidelity, a performance inaccessible to cochlear implant encoders. Moreover, mice equipped with a Braincodec-driven auditory cortical prosthesis matched the accuracy of non-implanted mice in discriminating real sounds. Braincodec therefore enables the development of high-throughput cortical neuroprosthetics with the potential to overcome the limitations of current auditory implants and provides a generic framework applicable to other sensory modalities.
Open AccessA biomimetic deep encoder for auditory restoration by cortical stimulation
Cochlear and auditory brainstem implants are the primary treatments for profound deafness 1 . However, the inherently low resolution of electrical stimulation used by these devices limits information transfer to a small number of independent stimulation channels 2 and results in perceptual difficulties. An alternative approach is to target the auditory cortex, a brain region large enough to accommodate hundreds of stimulation channels and whose stimulation evokes sound percepts 3,4 . Although modern brain interfaces offer large channel counts and spatial resolution, developing encoding algorithms capable of accurately transmitting interpretable sound information remains challenging. Here we introduce Braincodec, a deep network that optimizes both sound information transmission and the alignment of stimulation patterns with the spatial, temporal, and high-level properties of the auditory cortex neural code. Braincodec operates online and requires only a few hundred channels to transmit sufficient information for the reconstruction of natural sounds with high fidelity, a performance inaccessible to cochlear implant encoders. Moreover, mice equipped with a Braincodec-driven auditory cortical prosthesis matched the accuracy of non-implanted mice in discriminating real sounds. Braincodec therefore enables the development of high-throughput cortical neuroprosthetics with the potential to overcome the limitations of current auditory implants and provides a generic framework applicable to other sensory modalities.
- 9 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Du signal vocal à la réponse neuronale : Modélisation computationnelle et perspective évolutive sur le traitement de la voix dans le cortex auditif
Les voix sont omniprésentes dans notre quotidien, essentielles pour communiquer et véhiculant des informations non verbales telles que l'identité, le genre ou l'émotion de nos congénères. De nombreuses espèces manifestent des aptitudes vocales sophistiquées. L'étude des mammifères, notamment des primates non humains, révèle des traits communs de traitement vocal. L'objectif de cette thèse est d'approfondir notre compréhension des mécanismes de traitement de la voix, en employant des méthodes computationnelles, y compris l'apprentissage profond, pour mieux comprendre l'encodage vocal et la communication vocale. Le premier chapitre offre une vue d’ensemble des primates, leur évolution et communication, et introduit les concepts clés et méthodologies pour les chapitres suivants. Le second chapitre présente une analyse comparative approfondie du cortex vocal chez les primates. Le troisième chapitre examine la corrélation entre l'activité cérébrale liée à l'identité vocale, mesurée à l'aide de techniques de neuroimagerie, et les représentations issues de l'apprentissage profond via l'encodage et le décodage. Enfin, le quatrième chapitre traite de l'utilisation de méthodes computationnelles pour étudier les vocalisations de primates non humains, offrant une ressource précieuse pour les futures études en neuroéthologie. Associant neuroimagerie, modélisation computationnelle et base de vocalisations, cette thèse enrichit notre compréhension du traitement vocal des primates, éclairant les origines du langage humain et offrant de nouvelles perspectives en neurosciences auditives.
- 9 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Du signal vocal à la réponse neuronale : Modélisation computationnelle et perspective évolutive sur le traitement de la voix dans le cortex auditif
Les voix sont omniprésentes dans notre quotidien, essentielles pour communiquer et véhiculant des informations non verbales telles que l'identité, le genre ou l'émotion de nos congénères. De nombreuses espèces manifestent des aptitudes vocales sophistiquées. L'étude des mammifères, notamment des primates non humains, révèle des traits communs de traitement vocal. L'objectif de cette thèse est d'approfondir notre compréhension des mécanismes de traitement de la voix, en employant des méthodes computationnelles, y compris l'apprentissage profond, pour mieux comprendre l'encodage vocal et la communication vocale. Le premier chapitre offre une vue d’ensemble des primates, leur évolution et communication, et introduit les concepts clés et méthodologies pour les chapitres suivants. Le second chapitre présente une analyse comparative approfondie du cortex vocal chez les primates. Le troisième chapitre examine la corrélation entre l'activité cérébrale liée à l'identité vocale, mesurée à l'aide de techniques de neuroimagerie, et les représentations issues de l'apprentissage profond via l'encodage et le décodage. Enfin, le quatrième chapitre traite de l'utilisation de méthodes computationnelles pour étudier les vocalisations de primates non humains, offrant une ressource précieuse pour les futures études en neuroéthologie. Associant neuroimagerie, modélisation computationnelle et base de vocalisations, cette thèse enrichit notre compréhension du traitement vocal des primates, éclairant les origines du langage humain et offrant de nouvelles perspectives en neurosciences auditives.
- Open Access
- 28 March 2026
- Published by Elsevier in Trends in Cognitive Sciences
Voice information processing by the primate brain
- 28 March 2026
- Published by Elsevier in Trends in Cognitive Sciences
Open AccessVoice information processing by the primate brain
- Open Access
- 8 January 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 13
Reconstructing voice identity from noninvasive auditory cortex recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, ‘temporal voice areas’ (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.
- 8 January 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 13
Open AccessReconstructing voice identity from noninvasive auditory cortex recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, ‘temporal voice areas’ (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.
- Open Access
- 20 November 2025
- Published by eLife Sciences Publications, Ltd in eLife preprints
Reconstructing Voice Identity from Noninvasive Auditory Cortex Recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, “temporal voice areas” (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.
- 20 November 2025
- Published by eLife Sciences Publications, Ltd in eLife preprints
Open AccessReconstructing Voice Identity from Noninvasive Auditory Cortex Recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, “temporal voice areas” (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.
- Open Access
- 24 September 2025
- Published by Wiley in The Journal of Physiology
- Vol. 604 (21) , 8702-8720
Sound offset responses become highly informative in the auditory cortex
The entire auditory system downstream of the cochlea features pronounced offset responses that follow the termination of sounds. Because of their ubiquity, it remains an unsolved question whether offset responses are generated early in the auditory system and subsequently propagated downstream or recomputed at each processing stage. Here, we analysed large‐scale sound response datasets acquired in the cochlear nucleus, inferior colliculus, medial geniculate nucleus and auditory cortex of awake mice. All brain regions showed a significant proportion of offset responses, often combined with onset and sustained responses in the same neuron. However, using population activity decoders, we observed that neural representations after the sound offset show a three‐fold increase in sound encoding accuracy in the cortex relative to subcortical areas. This result indicates that cortical offsets encode a more precise short‐term memory of the elapsed sound than subcortical offsets and that they probably result from specific computational steps. image Key points: Offset responses are found throughout the central auditory system. Offset responses are often combined with sustained and onset responses at all central auditory system stages. Offset responses provide richer information about elapsed sounds in the auditory cortex.
- 24 September 2025
- Published by Wiley in The Journal of Physiology
- Vol. 604 (21) , 8702-8720
Open AccessSound offset responses become highly informative in the auditory cortex
The entire auditory system downstream of the cochlea features pronounced offset responses that follow the termination of sounds. Because of their ubiquity, it remains an unsolved question whether offset responses are generated early in the auditory system and subsequently propagated downstream or recomputed at each processing stage. Here, we analysed large‐scale sound response datasets acquired in the cochlear nucleus, inferior colliculus, medial geniculate nucleus and auditory cortex of awake mice. All brain regions showed a significant proportion of offset responses, often combined with onset and sustained responses in the same neuron. However, using population activity decoders, we observed that neural representations after the sound offset show a three‐fold increase in sound encoding accuracy in the cortex relative to subcortical areas. This result indicates that cortical offsets encode a more precise short‐term memory of the elapsed sound than subcortical offsets and that they probably result from specific computational steps. image Key points: Offset responses are found throughout the central auditory system. Offset responses are often combined with sustained and onset responses at all central auditory system stages. Offset responses provide richer information about elapsed sounds in the auditory cortex.
- Open Access
Sound offset responses become highly informative in the auditory cortex
The entire auditory system downstream of the cochlea features pronounced offset responses, which follow the termination of sounds. Because of their ubiquity, it is still an unsolved question whether offset responses are generated early in the auditory system and then propagated or recomputed at each processing stage. Here, we analysed large-scale sound responses datasets acquired in the cochlear nucleus, inferior colliculus, medial geniculate nucleus and auditory cortex of awake mice. All brain regions showed a significant proportion of offset responses often combined with onset and sustained responses in the same neuron. However, using population activity decoders, we observed that neural representations after the sound offset show a three-fold increase in sound encoding accuracy in the cortex relative to subcortical areas. This result indicates that cortical offsets encode a more precise short-term memory of the elapsed sound than subcortical offsets and that they likely result from specific computational steps. Key points summary: Offset responses are found throughout the central auditory system Offset responses are often combined to sustained and onset responses at all central auditory system stages Offset responses provide richer information about elapsed sounds in the auditory cortex
Open AccessSound offset responses become highly informative in the auditory cortex
The entire auditory system downstream of the cochlea features pronounced offset responses, which follow the termination of sounds. Because of their ubiquity, it is still an unsolved question whether offset responses are generated early in the auditory system and then propagated or recomputed at each processing stage. Here, we analysed large-scale sound responses datasets acquired in the cochlear nucleus, inferior colliculus, medial geniculate nucleus and auditory cortex of awake mice. All brain regions showed a significant proportion of offset responses often combined with onset and sustained responses in the same neuron. However, using population activity decoders, we observed that neural representations after the sound offset show a three-fold increase in sound encoding accuracy in the cortex relative to subcortical areas. This result indicates that cortical offsets encode a more precise short-term memory of the elapsed sound than subcortical offsets and that they likely result from specific computational steps. Key points summary: Offset responses are found throughout the central auditory system Offset responses are often combined to sustained and onset responses at all central auditory system stages Offset responses provide richer information about elapsed sounds in the auditory cortex
- Open Access
- 13 May 2025
- Published by Springer Nature in Scientific Data
- Vol. 12 (1) , 1-9
A large annotated dataset of vocalizations by common marmosets
Non-human primates, our closest relatives, use a wide range of complex vocal signals for communication within their species. Previous research on marmoset (Callithrix jacchus) vocalizations has been limited by sampling rates not covering the whole hearing range and insufficient labeling for advanced analyses using Deep Neural Networks (DNNs). Here, we provide a database of common marmoset vocalizations, which were continuously recorded with a sampling rate of 96 kHz from an animal holding facility housing simultaneously ~20 marmosets in three cages. The dataset comprises more than 800,000 files, amounting to 253 hours of data collected over 40 months. Each recording lasts a few seconds and captures the marmosets’ social vocalizations, encompassing their entire known vocal repertoire during the experimental period. Around 215,000 calls are annotated with the vocalization type. We offer a trained classifier to assist future investigations. Finally, we validated our dataset by sampling 700 representative recordings and cross-examining them with four experts.
- 13 May 2025
- Published by Springer Nature in Scientific Data
- Vol. 12 (1) , 1-9
Open AccessA large annotated dataset of vocalizations by common marmosets
Non-human primates, our closest relatives, use a wide range of complex vocal signals for communication within their species. Previous research on marmoset (Callithrix jacchus) vocalizations has been limited by sampling rates not covering the whole hearing range and insufficient labeling for advanced analyses using Deep Neural Networks (DNNs). Here, we provide a database of common marmoset vocalizations, which were continuously recorded with a sampling rate of 96 kHz from an animal holding facility housing simultaneously ~20 marmosets in three cages. The dataset comprises more than 800,000 files, amounting to 253 hours of data collected over 40 months. Each recording lasts a few seconds and captures the marmosets’ social vocalizations, encompassing their entire known vocal repertoire during the experimental period. Around 215,000 calls are annotated with the vocalization type. We offer a trained classifier to assist future investigations. Finally, we validated our dataset by sampling 700 representative recordings and cross-examining them with four experts.
- Open Access
Reconstructing Voice Identity from Noninvasive Auditory Cortex Recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, “temporal voice areas” (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.
Open AccessReconstructing Voice Identity from Noninvasive Auditory Cortex Recordings
The cerebral processing of voice information is known to engage, in human as well as non-human primates, “temporal voice areas” (TVAs) that respond preferentially to conspecific vocalizations. However, how voice information is represented by neuronal populations in these areas, particularly speaker identity information, remains poorly understood. Here, we used a deep neural network (DNN) to generate a high-level, small-dimension representational space for voice identity—the ‘voice latent space’ (VLS)—and examined its linear relation with cerebral activity via encoding, representational similarity, and decoding analyses. We find that the VLS maps onto fMRI measures of cerebral activity in response to tens of thousands of voice stimuli from hundreds of different speaker identities and better accounts for the representational geometry for speaker identity in the TVAs than in A1. Moreover, the VLS allowed TVA-based reconstructions of voice stimuli that preserved essential aspects of speaker identity as assessed by both machine classifiers and human listeners. These results indicate that the DNN-derived VLS provides high-level representations of voice identity information in the TVAs.