Internship offers

We have two fully funded M2 / PhD positions in computational neuroscience / NeuroAI at CERCO, Toulouse, France: 1) Modeling visual learning in bees with deep neural networks. We are looking for an M2 and/or PhD student. More details here: https://www.dropbox.com/scl/fi/5jnm96coewyiwv9szh6l9/Bees_M2-PhD_2026.pdf?rlkey=k5iymlh29mamfs92ep7uuczog&dl=0 2) From Synaptic Delays to Neural Sequences: Modeling Temporal Neural Coding with Spiking Neural Networks We are looking for an M2 student. More details here: https://www.dropbox.com/scl/fi/v9lvbs02uifz3feofscuz/CerCo_M2_Offer_2027.pdf?rlkey=r6e00qynzqn4wc7fq1gh0lnfu&dl=0

Supervisor

Timothée Masquelier

Host lab

Centre de recherche cerveau et cognition, UMR5549 CNRS - Univ. Toulouse

Duration

5

Posted

21 October 2026

Les littératures sur les stéréotypes et les biais sociaux ont démontré que les individus ne sont pas traités de la même façon dans un groupe ou même en société (Hilton 1996). Par exemple, des travaux en psychophysique computationnlle récents ont démontré que la forme du visage influence les premières impressions qu’on a de celui-ci: Des visages d’avatar sont jugés par des tiers personnes comme étant plus dominants si ceux-ci ont des traits plus masculins, ou ont des caractéristiques associées à la force physique (Oosterhof & Todorov, 2008). Le but de ce stage est d’investiguer à quel point ces caractéristiques faciales de dominance ou de masculinité influencent (ou biasent) les intereactions sociales humaines, afin de comprendre plus finement les mécanismes cognitifs qui soustendent les stéréotypes et les biais sociaux humains. Afin d’étudier cette question de façon rigoureuse expérimentalement, il est nécessaire de contrôler la forme du visage des participants pendant que ceux-ci interagissent librement. Pour ce faire, nous avons développé une plateforme de vidéoconférence, appelée DuckSoup, qui permet aux chercheur·euse·s d’utiliser des transformations algorithmiques (des filtres de réalité augmenté) de la voix et du visage en temps réel, afin de contrôler les signaux sociaux et les attributs physiques des participants (Figure 1). Cette plateforme ouvre la possibilité de comprendre comment certains signaux sociaux et caractéristiques physiques influencent causalement les interactions sociales humaines (Arias-Sarah 2024). Dans le cadre de ce stage, nous allons créer un algorithme de transformation de la morphologie du visage, capable de transformer la forme de vrais visages pour les rendre plus ou moins masculins, forts, mince, ou dominants. Nous tenterons ensuite de valider ces transformations dans des expériences de sciences cognitives, puis d’étudier à quel point ces caractéristiques physiques influencent et biaisent les jugements sociaux. Le but à moyen terme est d’utiliser ces transformations du visage de façon couverte et en temps-réel pendant des vrais interactions, afin de comprendre comment les caractéristiques du visage influencent le comportement social humain. Par exemple, il est question de comprendre si les personnes ont tendance à se comporter d’une façon spécifique avec des personnes qui ont des caractéristiques physiques données. Dans d’autres termes, on s’intéresse à répondre aux questions : Est-ce qu’une personne transformée pour avoir des caractéristiques plus masculines/dominantes dans son visage est favorisée par un groupe? Ou, comment se comportent les gens envers une personne une fois qu’on a augmenté ou dimunué la masculinité de son visage? La force de ce paradigme expériemental réside dans la possibilité de contrôler finement les caractéristiques physiques des participants, ce qui nous permet de tirer des conclusions causales sur les mécanismes cognitifs sous-jacents. Plus largement, ce stage s’encadre dans un projet qui s’intéresse à l’étude de l’impact psychologique des algorithmes de réalité augmenté sur le comportement humain, et s’encadre dans le projet PEPR COFFEE (Collaborative filters: Exploring the impact of augmented-reality filters on digital collaboration), coordonné par Pablo Arias-Sarah (LPNC). Celui-ci vise à explorer l’impact des filtres de réalité augmenté sur la psychologie humaine. Finalement, le sujet de ce stage étant très vaste, et touchant à plusieurs sujets, le travail spécifique à réaliser sera à définir avec le·a candidat·e·s retenu·e, en fonction de leur appétence scientifique. Ce travail pourra aussi être développé plus profondément lors d’une thèse de doctorat. Ce stage est en collaboration entre Pablo Arias-Sarah (LPNC, Grenoble) et Louise Goupil (LPNC, Grenoble) et a vocation à se dérouler au Laboratoire de Psychologie et neurocognition (LPNC) à Grenoble. Activités principales: - Recherche bibliographique sur la perception des visages et les premières impressions - Développement et calibration d’algorithmes de vision par ordinateur informés par les travaux en psychologie expériemental - Design d’expériences de sciences cognitives, si possible en interaction - Analyse de données Compétences pouvant être développées: - Développement python - Connaissances approfondies de la littérature sur les biais sociaux - Design expérimental de sciences cognitives - Analyse de données statistiques (en R et en python) - Connaissances de la littérature sur les interactions sociales Informations complémentaires: • Encadrants : Dr Pablo Arias-Sarah & Dr. Louise Goupil • Procédure de candidature : les candidatures (avec CV détaillé, lettre de motivation, et relevés de notes) doivent être transmises à Pablo Arias pablo.arias-sarah@univ-grenoble-alpes.fr et Louise Goupil louise.goupil@univ-grenoble-alpes.fr

Supervisor

Pablo Arias Sarah

Host lab

Laboratoire de Psychologie et NEurocognition

Duration

5 months

Posted

7 October 2026

Internship offer: Updating mental representations after occlusion: a common rate controller or separate controllers? Starting date: Early 2027 Duration: 5 to 6 months Supervisors: Robin Baurès, robin.baures@cnrs.fr and Jean-Baptiste Durand, jbdurand@cnrs.fr Location: CerCo, Toulouse Keywords: Extrapolation, occlusion, rate controller, time perception, spatial cognition, motor system, psychophysics. Scientific context In many everyday situations, we need to estimate the properties of an object after it has disappeared from view. When crossing a street, for instance, if an approaching car becomes hidden behind another vehicle, we must estimate how long it will take to reach us, and this estimate has to be continuously updated to remain accurate. This ability to dynamically update a mental representation is involved in many tasks: spatial, temporal and numerosity estimation, among others. How are these representations updated? Two hypotheses have been put forward. According to the common rate controller hypothesis, a single mechanism sets the rate at which all extrapolated properties are updated. At the brain level, motor structures have been proposed as the functional substrate of this controller. Conversely, the separate controllers hypothesis states that each property is updated by a distinct, dedicated process. Deciding between these two hypotheses would shed light on how the brain maintains dynamic representations in the absence of sensory input, and on the still-debated role of motor areas in perceptual and cognitive processes. Aim of the internship This internship is part of a project funded by the French National Research Agency (ANR). It aims to evaluate the common and separate controller hypotheses using behavioural extrapolation tasks targeting different properties (time, position, numerosity, etc.). Comparing performance across these tasks will make it possible to test whether they rely on a shared mechanism. Some tasks may also manipulate participants' motor engagement, in order to infer the role of motor areas in these processes. Concretely, the successful candidate will: - organise and run experimental sessions (participant recruitment, scheduling, data collection); - analyse the collected data; - take part in the life of the laboratory (seminars, team meetings). Depending on how the internship unfolds, the project may lead to a PhD. Required skills - Background in neuroscience, psychology or human movement science; - Interest in psychophysics and human experimental research; - Basic knowledge of statistics and programming (e.g. R) is a plus; - Rigour and autonomy in experimental work. Practical information The internship will take place at the Centre de Recherche Cerveau & Cognition (CerCo, CNRS UMR 5549) in Toulouse, France. The intern will receive the standard French internship allowance, approximately €700 per month for a full-time internship, revised each year on 1 January. Applications from non-French-speaking students are welcome. To apply or for further information, please contact Robin Baurès (robin.baures@cnrs.fr) and Jean-Baptiste Durand (jbdurand@cnrs.fr).

Supervisor

Robin Baurès

Host lab

CerCo

Duration

5-6 mois

Posted

28 September 2026

Scientific context The neural bases of familiar face recognition have been investigated for more than two decades, yet the literature remains strikingly inconsistent, whichever technique is considered. In fMRI, several studies report stronger responses to familiar than to unfamiliar faces along the ventral visual pathway (e.g., Visconti di Oleggio Castello et al., 2017), whereas others find no such difference in face-selective regions (e.g., Davies-Thompson et al., 2009), and reviews of this literature converge on the same conclusion: the effect of familiarity on ventral occipito-temporal responses is far from settled (Natu & O'Toole, 2011). In EEG, the debate takes a similar turn around the amplitude of the N170, the earliest face-selective component. Long considered insensitive to familiarity, and therefore taken as the marker of a purely perceptual stage preceding access to identity representations (e.g., Eimer, 2000), it has since been reported to be modulated by long-term familiarity, even if these effects remain small and variable across studies (Caharel & Rossion, 2021). This uncertainty matters well beyond the laboratory. Clarifying how familiarity shapes face-selective responses is a prerequisite for understanding disorders of face recognition such as prosopagnosia, and more broadly for building a mechanistic account of how we identify the people around us. Part of the problem lies with the methods themselves. Conventional neuroimaging designs contrasting blocks or events of familiar and unfamiliar faces suffer from limited sensitivity, from ambiguity as to which comparison isolates the process of interest, and from the difficulty of separating identity recognition from the semantic and affective knowledge attached to famous faces. To address these limitations, we are developing at CerCo, in collaboration with Bruno Rossion (CRAN, CNRS & Université de Lorraine), an innovative stimulation approach known as frequency tagging (Liu-Shuang et al., 2014; Norcia et al., 2015). First developed in electroencephalography, it has more recently been adapted to functional imaging with considerable success (Gao et al., 2018; Laurent et al., 2023; Zheng et al., 2026). In this framework, a rapid periodic stream of face images is presented while a specific property of interest is manipulated at a lower, precisely defined frequency. Any brain response occurring exactly at that frequency provides an objective, high signal-to-noise index of the process under study, without requiring an explicit task from the participant. Our team has already demonstrated the value of this approach for identifying the neural bases of individual face recognition, of familiar face processing using pictures of celebrities, and of the processing of dynamic facial properties such as head orientation. Our current work extends it to faces that participants learn in the laboratory, which offers a decisive advantage: familiarity is built under controlled conditions, so that pre-existing semantic and emotional associations no longer confound the comparison. Several experiments are currently being conducted in parallel, in fMRI and in EEG, so that the same question can be addressed with complementary spatial and temporal resolutions. Beyond familiarity itself, the approach opens the way to finer investigations of social cognition, including how group membership or affective states modulate the neural discrimination of faces. Aim of the internship This internship is part of an ongoing research programme on the neural mechanisms of face familiarity. The stimulation protocols and the familiarisation procedures will already be in place when the internship starts, so that the successful candidate can focus directly on data collection, data analysis and scientific writing. The exact balance between these activities will depend on the state of the different projects at that time, on the experiments that are then a priority for the team, and on the interests of the candidate. Concretely, the successful candidate will: • Organise and run experimental sessions, from participant recruitment and scheduling to data acquisition, in fMRI and/or in EEG; • Analyse the collected data using in-house documented pipelines that have been designed to make students autonomous in the processing of frequency-tagging data; • Contribute to the writing of a scientific article with a view to publication, to an extent that will depend on the results obtained and on the progress of the work. The candidate will be integrated into an active research team and will gain hands-on experience with human neuroimaging, from data acquisition to analysis and scientific communication. Application Applications should include a CV, a short cover letter describing your research interests, and the contact details of one referee. They should be sent to: alan.guyomarch@ube.fr and benoit.cottereau@cnrs.fr.

Supervisor

Benoit Cottereau

Host lab

Centre de Recherche Cerveau & Cognition

Duration

6 months

Posted

25 September 2026

Background Assistants based on large language models (LLMs) offer new ways to support justice professionals in tasks such as finding information, summarizing documents, and analyzing case files. In this context, a system’s technical performance alone does not determine its usefulness. A central question is how users assess the assistant’s reliability and adapt their behavior accordingly. Too little trust may lead users to underuse a reliable tool. Conversely, excessive trust may lead them to accept incorrect answers without checking them. The goal is therefore less to maximize trust than to foster appropriate trust: use that reflects the system’s actual capabilities and limitations. This issue is particularly important in the justice system, where some errors can have significant consequences and where traceability, source verification, and continued human oversight are essential. The internship is part of a collaboration between the École normale supérieure and the French Ministry of Justice on reliability, uncertainty, and human–AI interaction. Aim of the internship The main aim is to explore methods for measuring and characterizing users’ trust in an AI assistant, with particular attention to professional uses in the justice system. The project will distinguish between several concepts often grouped under the term “trust”: - the trust users report; - the trust revealed by their behavior; - the confidence expressed by the system; - the objective reliability of the answer produced; - users’ ability to adjust their trust when the system is correct or makes a mistake. The central research question is the following: How can we determine whether a user’s trust in an AI assistant is appropriate to its actual reliability? Areas of work The internship may be organized around three complementary areas. 1. Review of research on human–AI trust. The intern will conduct a focused literature review on topics including: - trust in AI and actual reliance on system recommendations; - trust calibration, or the match between users’ trust and the system’s actual reliability; - automation bias and overreliance; - uncertainty management, the system’s ability to abstain, and human verification; - explainability and the effects of explanations on trust; - the effects of sources, citations, and justifications on perceived reliability. A particular aim will be to identify the experimental methods and measures used to study these phenomena. 2. Identifying indicators of trust Drawing on the literature, the intern will propose a set of indicators applicable to an AI assistant. These may combine subjective and behavioral measures, such as: - reported confidence in an answer; - the decision to accept, correct, or reject a system suggestion; - how often a user checks the system’s answers; - time spent checking; - whether the user consults the sources provided; - ability to identify incorrect answers; - changes in a decision after consulting the assistant; - the gap between user trust and the system’s actual reliability. Particular attention may be given to trust calibration: the extent to which trust in the system corresponds to its actual performance. 3. Designing an experimental protocol The intern will propose an initial protocol for studying these phenomena through tasks representative of how an AI assistant might be used. For example, an experiment could present participants with several answers generated by an assistant. Participants would evaluate each answer, decide whether to use or verify it, and report their confidence. The protocol should allow comparisons between experimental conditions - for example, an answer presented alone or accompanied by sources, an uncertainty indicator, or an explanation or justification. The aim is not necessarily to conduct a large-scale user study during the internship, but to design a sufficiently robust protocol to support future research. If time permits, the intern may also identify directions for the future design of the Ministry’s AI assistants: - which reliability indicators should be shown to users? - how should the system’s uncertainty be represented? - when should users be explicitly encouraged to verify an answer? - how can explanations be designed without creating unwarranted trust? - which user behaviors should be monitored to detect possible overreliance? Supervision and research environment The internship will take place within a collaboration between the École normale supérieure and the French Ministry of Justice. For this reason, a good command of French is required. The intern will work at the intersection of artificial intelligence, human–AI interaction, evaluation of generative systems, and cognitive science. The work may contribute to a broader scientific discussion of the question: “How can we trust AI in the justice system?”

Supervisor

Valerian Chambon

Host lab

Institut Jean Nicod

Duration

4-5months (Feb-June/July)

Posted

23 September 2026

The Scientific Question: The research investigates how the human brain's Language Network encodes natural language, aiming to understand how core regions of this network represent abstract, language-independent meaning and structure. It also explores neural coding across different resolutions, ranging from single-cell activity up to large neuronal populations using intracranial recordings. The Methods: The project utilizes intracranial data collected from electrodes implanted directly in the brains of human patients. This includes multi-resolution data ranging from single-neuron spiking activity and broadband-gamma activity (BGA) from microwires to broader signals from macro contacts across various brain regions (such as the left fusiform gyrus and superior temporal gyrus). Responsibilities: The student will design, implement, test, and deploy advanced machine learning models (using Python and deep learning frameworks like PyTorch) to decode brain signals into language and linguistic features. They will also analyze and process large datasets on High-Performance Clusters (HPC) with GPU optimization to train and fine-tune their models.

Supervisor

Yair Lakretz

Host lab

LSCP, DEC, ENS-Paris

Duration

6 months

Posted

18 September 2026

Monitoring and control processes allow us to adjust our behavior and to learn from it. Think of adjusting your finger pressure on a guitar string the moment a note starts to sound off, or catching yourself mid-melody and fixing the next note — or, in speech, adjusting your tongue position when a sound starts to slip, or catching yourself mid-sentence and rephrasing. Over time, these same signals also drive learning — the gradual improvement behind becoming a fluent guitarist or speaker. This project asks whether this loop — monitoring, control and learning — operates similarly across actions differing in hierarchical level (motor to conceptual) and domain (linguistic and non-linguistic), or whether separate, specialized systems are involved instead. Carried out within this broader project, the M2 internship draws on two complementary paradigms — perturbing established actions (inducing errors or increasing demands) and learning new ones — to design and validate a task for later use in fMRI. The student will work closely with Elin Runnqvist, PhD student Jakob Scherm Eikner, and the wider team. We are looking for someone collaborative and highly motivated for research. A fully funded PhD position starts September 2027; the internship does not guarantee it, but is good preparation for applying.

Supervisor

Elin Runnqvist

Host lab

Laboratoire Parole et Langage, Aix-en-Provence (LPL, UMR 7309; Aix-Marseille Université)

Duration

4 mois

Posted

17 September 2026

Context: Understanding brain function requires bridging multiple analysis scales, from synaptic dynamics to large-scale neuronal population activity. Neuronal networks generate large-scale activity patterns that support cognitive function. However, understanding how the anatomical and physiological properties of the network shape network activity and function is still a major challenge. Working memory is a cognitive process well suited to study this link as it requires transient maintenance of information over short delays (seconds), bridging fast sensory processing (milliseconds) and long-term memory retention (days). It is defined as the ability to temporarily store and manipulate behaviourally relevant information. The objective of our project is to elucidate how synaptic properties give rise to population activity patterns in working memory, which could reveal fundamental principles linking synaptic physiology and network computation. The project relies on a hybrid approach combining neuronal recordings and theoretical modelling. M2 project: The intern will work on the construction of predictive neuronal network models constrained by available experimental evidence both from the existing scientific literature and from the dataset collected by the project experimental partner. He/she will work exclusively at the interface between models and large-scale data analysis, without participating in animal handling. Depending on the student’s interests, the project can emphasize modelling or data analysis. Key tasks may include: o Development of models: Implementation and analysis of attractor models (rate-based and spiking neurons networks) incorporating realistic physiological parameters (short-term plasticity, excitation/inhibition balance) to test hypotheses regarding memory state stability. o Electrophysiological data analysis: Processing and spike-sorting of Neuropixels 2.0 recording data acquired in the mediodorsal prefrontal cortex and the posterior parietal cortex. o Application of decoding methods (Machine Learning, SVM, Principal Component Analysis) to identify neuronal activity patterns associated with information maintenance.

Supervisor

Arthur Leblois

Host lab

Institut des maladies neurodégénératives (IMN), Neurocampus, Bordeaux

Duration

6 months (Jan-June)

Posted

16 September 2026

Background: The human brain can process single words during different tasks and via different modalities. We can hear the word ‘apple’ and repeat it out loud or write it down (dictation), or we can see the word written on a page and read it aloud or copy it. These four tasks (word repetition, reading, dictation, and copying) require different computations at early stages, but share others at higher, amodal, levels. Decades of research in neuropsychology and cognitive sciences have produced a detailed description of the various information-processing stages during each of these tasks, based on examination of patients with brain damage, often highly localized. Each of these processing steps is considered a separate module, which can be selectively impaired following a focal brain lesion, which would lead to specific error patterns in one or more of these tasks (Friedmann & Coltheart, 2016). Neuroimaging studies have isolated cortical regions putatively associated with several of these processing stages; however, they often target a specific component of the model in a specific modality (Kearns et al., 2019), and thus could not provide a comprehensive view on single-word processing in the human brain. A similar limitation persists in computational work on single-word processing. Most work addressed a single task, such as reading (Norris, 2013) or word repetition (Sajid et al., 2022), often using symbolic, discrete models, without exploiting the advances made by modern deep neural models. In this project, we will develop a new experimental design to study the neural basis of these four tasks jointly, using a combination of non-invasive fMRI and invasive intracranial recordings. This multimodal approach allows us to map large-scale networks with fMRI while leveraging intracranial data to capture high-speed neural dynamics and fine-grained behavioral outputs, such as writing kinematics via a digital tablet. By adapting these paradigms for clinical settings (e.g., awake brain surgery), we can map causal circuits with high precision. Throughout the project, you will work closely with students who work on the development of deep neural models that simulate the four tasks jointly. Predictions from the models about neural mechanisms will be tested against the fMRI and intracranial data. Main Goal: Create a global map of the processing pathways in the brain for single-word processing and identify its amodal parts. The ultimate goal of the project is to provide a global view of single-word processing in the brain, delineating the various processing stages and, in particular, disentangling modality-specific and amodal processing of words. We will test the hypothesis that the unified single-word processing model (Dotan et al., 2015) can be mapped onto the human language network, such that each processing component (boxes and arrows in the model) can be mapped onto dedicated brain regions. Responsibilities: Data Collection: Conduct pilot and collect fMRI and intracranial data from participants. Analyze the neural data. Requirements: Strong programming skills in Python. Familiarity with data structures and statistical analyses. Familiarity with fMRI recordings (advantage).

Supervisor

Yair Lakretz

Host lab

LSCP, DEC, ENS-Paris

Duration

6 months full time

Posted

15 September 2026

Our laboratory regularly offers M2 internship positions in the broad field of learning and decision-making. We are interested in a range of related topics: human reinforcement learning, that is, how the history of past successes and failures shapes our decisions; behavioural economics, or how we make decisions based on described information about risk, delay and loss; and the question of how decision-making can be improved, through applied research that seeks to shape, nudge and boost better decisions in real life. We also study human–machine interaction, examining how we take decisions jointly with artificial systems and how we process decisions taken by such systems; machine psychology, which uses cognitive science to understand how large language models (LLMs) work; and the epistemological foundations of cognitive science and AI cognition. To investigate these questions, we combine several methods: computational cognitive modelling and simulation, including agent-based modelling; experimental investigation of behaviour, through both online and laboratory experiments as well as eye tracking; and the use of LLMs, both as a subject of study and as a tool. We welcome applications from students with a background in cognitive science, psychology, economics, computer science, or a related field. A good working familiarity with programming (for example Python, R or MATLAB) and with scientific writing is important, as both are central to the day-to-day work of the internship. During the internship, the M2 student will generally be paired with a PhD student or postdoctoral researcher, who will provide close guidance alongside the overall supervision.

Supervisor

Stefano Palminteri

Host lab

Laboratoire de neurosciences cognitives et computationnelles

Duration

6 months

Posted

1 September 2026

Les abeilles sont confrontées à une pression croissante liée aux facteurs de stress d'origine anthropique. Parmi ceux-ci, les pesticides figurent au rang des menaces les plus répandues pour la santé des pollinisateurs. Le fipronil, un insecticide de la famille des phénylpyrazoles ciblant le système GABAergique, suscite une inquiétude particulière. Même à des concentrations sublétales représentatives des conditions de terrain, le fipronil altère les capacités cognitives des abeilles mellifères adultes (El Hassani et al., 2005) et réduit la densité des complexes synaptiques au niveau de leurs corps pédonculés (Privitt et al., 2023). L'exposition larvaire via du nectar et du pollen contaminés constitue une voie d'intoxication encore mal comprise, mais susceptible d'entraîner de graves conséquences neurodéveloppementales. Il a été démontré que l'exposition du couvain aux néonicotinoïdes — qui ciblent le système cholinergique — réduit le volume cérébral et altère les performances cognitives à l'âge adulte chez Bombus terrestris (Smith et al., 2020). On ignore encore si une perturbation du système GABAergique durant le développement larvaire entraîne des altérations neuroanatomiques et cognitives équivalentes ou distinctes chez Bombus terrestris. Combler ces lacunes constitue non seulement un enjeu majeur en écotoxicologie, mais pourrait également révéler un mécanisme inédit par lequel la contamination larvaire, induite par une classe distincte d'insecticides neurotoxiques, fragilise insidieusement les colonies et contribue à leur effondrement. L'hypothèse centrale est que l'exposition larvaire au fipronil laisse une empreinte neuroanatomique mesurable chez les adultes, se traduisant par des déficits quantifiables au niveau de l'apprentissage et de la mémoire olfactifs. Cette hypothèse soulève trois questions de recherche spécifiques : - L'exposition larvaire au fipronil modifie-t-elle le volume de structures cérébrales clés, telles que les lobes antennaires et les corps pédonculés, chez les adultes fraîchement émergés ? - Les individus exposés présentent-ils une altération de l'apprentissage associatif olfactif et de la rétention mnésique par rapport aux témoins ? - Les différences inter-individuelles de volume des structures cérébrales permettent- elles de prédire la vitesse d'acquisition de l'apprentissage ? Pour répondre à ces questions, le projet associe l'expérience de l'équipe PTSN (IBMM) en neuro-toxicologie du développement et en imagerie cérébrale par micro-tomographie (micro-CT) chez les insectes sociaux, à celle de l'équipe IBT (CEFE) en écologie comportementale et conditionnement olfactif de Bombus terrestris. Les larves seront exposées par injection intra-alvéolaire directe. Les adultes issus de ces larves seront évalués à deux niveaux complémentaires. Sur le plan neuroanatomique, une imagerie cérébrale par micro-CT à rayons X sera réalisée à l'IBMM (Smith et al., 2020), où la segmentation des cerveaux d'abeilles sera facilitée par des algorithmes de deep learning. Sur le plan fonctionnel, l'apprentissage associatif et la mémoire seront évalués au moyen du conditionnement par réponse d'extension du proboscis (PER) au CEFE (Gómez-Moracho et al., 2022). Ces deux types de mesures seront effectués sur les mêmes individus. Références: El Hassani, A.K., Dacher, M., Gauthier, M., Armengaud, C. (2005). Effects of sublethal doses of fipronil on the behavior of the honeybee (Apis mellifera). Pharmacology, Biochemistry and Behavior, 82(1), 30–39. doi:10.1016/j.pbb.2005.07.008 Privitt, J.J., Van Nest, B.N., Fahrbach, S.E. (2023). Altered synaptic organization in the mushroom bodies of honey bees exposed as foragers to the pesticide fipronil. Frontiers in Bee Science, 1, 1219991. doi:10.3389/frbee.2023.1219991 Smith, D.B., Arce, A.N., Ramos Rodrigues, A., Bischoff, P.H., Burris, D., Ahmed, F., Gill, R.J. (2020). Insecticide exposure during brood or early-adult development reduces brain growth and impairs adult learning in bumblebees. Proceedings of the Royal Society B, 287(1922), 20192442. doi:10.1098/rspb.2019.2442 Gómez-Moracho, T., Durand, T., Lihoreau, M. (2022). The gut parasite Nosema ceranae impairs olfactory learning in bumblebees. Journal of Experimental Biology, 225(13), jeb244340. doi:10.1242/jeb.244340

Supervisor

Maxime Linard

Host lab

Institut des Biomolécules Max Mousseron (Montpellier)

Duration

5 months (Jannuary to June)

Posted

28 August 2026

This internship is part of a research project on the mathematical modeling of adaptive decision-making, being developed within Etienne Koechlin’s team at the Laboratory of Cognitive and Computational Neuroscience (Inserm U960, ENS-PSL). The proposed theoretical framework formalizes reward maximization as an optimization problem on a statistical variety, drawing on tools from information geometry and optimal transport. In particular, this model allows us to predict how an agent allocates its cognitive resources between exploration and exploitation depending on the informational structure of its environment. The goal of the internship is to experimentally test the predictions of this theory through the design and implementation of behavioral experiments, both online and in the laboratory. The intern will participate in the entire process: designing experimental paradigms, collecting data from participants, and comparing the results with the model’s predictions. Required Skills: Basic knowledge of Python Basic knowledge of cognitive science Ability to understand complex mathematical formalisms (no mathematical work will be required, but an understanding of the variables being tested is necessary) Desirable Skills (not required): Experience with Pavlovia Knowledge of JavaScript

Supervisor

Constantin Vaillant Tenzer

Host lab

LNC2 (ENS - PSL, Inserm U960)

Duration

Flexible (from October 2026). At least 1 day/week. Open to gap years.

Posted

19 July 2026

Many important questions in education require analysing large and complex datasets including measures of educational achievement, cognitive skills, environmental factors, and sometimes even genetic factors. We have access to several such datasets: The Eden cohort (N=400 to 1400), the Elfe cohort (N=9000), the DEPP 2007 panel (N=35000), the DEPP 2011 panel, the DEPP maternelle panel (N=35000), as well as more recent data from the national evaluations in many grades (N=800 000 per grade). Other international datasets may also be available (e.g., PISA, PIRLS, TIMMS, TEDS, BiB, Millenium cohort…). Projects may bear on the following questions, among others: - What are the early predictors of reading and math ability? to what extent are they mediated by early cognitive skills? to what extent do they differ between boys and girls? - What are the characteristics of schools, teachers, pedagogical practices and manuals that produce superior achievement? - Bring your own question! Student profile: data science, statistics, econometrics, psychology, cognitive science. Any other profile, but with strong analytical skills. Also available for M1.

Supervisor

Franck Ramus

Host lab

Laboratoire de sciences cognitives et psycholinguistique

Duration

6 months

Posted

2 July 2026

Current interventions for reading disability (dyslexia) show variable effects: they are often effective for some but not all of the children. It is difficult to predict which children will respond well to an intervention and why. It therefore remains difficult to optimise referral to different types of intervention. This project takes genetic factors into account to enrich our understanding of individual differences in reading trajectories and responses to intervention. Questions addressed will be: To what extent is the home literacy environment confounded with genetic factors? To what extent do they interact? What are the genetic, environmental and cognitive predictors of reading progress, in normal-reading and in dyslexic children? What are the genetic, environmental and cognitive predictors of dyslexic children’s response to educational or speech and language therapy? Other projects regarding gene-environment interplay can be considered.

Supervisor

Franck Ramus

Host lab

Laboratoire de sciences cognitives et psycholinguistique

Duration

6 months

Posted

2 July 2026

Five-month internship for a Master’s 2 student 2026-2027 2nd semester Callyope’s mission: monitoring brain disorders through speech At Callyope, we believe that patient-centered technology can improve the quality of life of millions of people by making brain health monitoring more objective and responsive. Callyope develops remote patient-monitoring solutions to help healthcare professionals detect relapses and identify optimal treatment strategies. To achieve this, we analyze hundreds of clinically validated speech biomarkers in order to provide insights into psychiatric, cognitive, and motor impairments associated with brain disorders. Our initial focus is psychiatry, starting with schizophrenia in the context of the research protocol for which the student will be recruited. This internship is part of a clinical research protocol in adult psychiatry. Patient recruitment will take place over the five-month duration of the Master’s 2 internship at the Centre Hospitalier de Versailles, which is partnering with Callyope for this research protocol. ------------------------ The Adult Psychiatry University Department at the Versailles Hospital Our university department includes several units. Patients living with schizophrenia will be recruited from the Chaslin unit, an open inpatient unit. Recruitment will concern patients whose hospital discharge has been scheduled and whose disorder is therefore clinically stable. Recruitment will also take place at the outpatient psychiatric consultation center, located on Rue Berthier in Versailles. This university department regularly hosts medical students and residents and therefore has extensive experience in supervision and teaching. The Centre Hospitalier de Versailles also has a dedicated research office, which will act as the sponsor of the study and includes statistical support. --------------------------- The MOODs research team, DisAp axis The Master’s 2 student will join the DisAp axis of the MOODs research team, affiliated with the CESP. The DisAp axis is geographically located within the Centre Hospitalier de Versailles. Four offices are dedicated to research activities and include workstations with computers. Two PhD students currently work there, as well as one third-year undergraduate student who will spend two months in the team. Two senior researchers / principal investigators also work there part-time. The MOODs team organizes four lab meetings per year. It is a research team of approximately one hundred members, including other Master’s 2 students, with whom the selected student will be able to interact. In addition, the CESP — Centre for Research in Epidemiology and Population Health — is a dynamic research structure with a doctoral students’ association that organizes various social events. ---------------------------- Internship description The Master’s 2 student will work closely with the team of investigators and will take an active role in the research protocol. In collaboration with the clinical research staff involved at the study site, the student will recruit patients and coordinate the pathway of participants included in the research protocol. He/she will be responsible for administering questionnaires and automated tests. Subsequently, the student will clean the database and perform statistical analyses of the psychometric data, while data related to the vocal samples will be processed by Callyope. Finally, the student will begin drafting scientific articles in English, which will be integrated into his/her final internship thesis. -------------------------------------------- Activities * Organizing patient recruitment * Administering the cognitive tasks to be completed by patients * Data processing * Data interpretation ---------------------------- Required skills * Strong organizational skills * Good interpersonal skills, particularly in contact with patients * Knowledge of ethics and professional conduct in clinical research * Knowledge of clinical research methods --------------------------- Preferred skills * Good command of French * Knowledge of a statistical software package, ideally R * Scientific English for article writing ------------------------ Benefits Paid internship: full-time internship, 7 hours per day, 22 days per month, with the legal monthly stipend for Master’s 2 internships. --- Workplace Centre Hospitalier de Versailles 177 rue de Versailles 78150 Le Chesnay, France Accessible by bus from the train stations of Vaucresson, Versailles Chantiers, and Versailles Rive Droite. Carpooling may also be possible. --- Contact To apply, please send a CV and cover letter, or contact us for further information: Dr Frileux Solène sfrileux@ght78sud.fr

Supervisor

Solène Frileux

Host lab

CESP (Centre d'Epidémiologie et de Santé des Populations), MOODs team

Duration

5 months (2nd semester, 2026-2027 year)

Posted

30 June 2026

Many tools that we use to think and communicate — letters, numbers, calendars, artistic forms, and other symbols — have a cultural history. They evolved by being transmitted from one mind to another, then another, and this process was shaped by cognitive constraints, some symbols being easier to learn than others. Cognitive science and cultural evolution make predictions about the structure of cultural symbols, which can be tested in two main ways. First, cultural evolution can be modeled and these models can be tested against historical or anthropological data. Second, we can design experiments to reproduce, in the lab, mechanisms that could have played a role in the history of symbols.

Supervisor

Olivier Morin

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

2 to 5 months

Location

ENS, DEC, 29 rue d'Ulm, 75005 Paris, France

In military environments, where high-stress situations are common, maintaining or enhancing cognitive performance is critical for ensuring both the safety and success of combat operations. Similar to physical training, emerging techniques like neurofeedback offer the potential to optimize cognitive abilities in soldiers. Neurofeedback relies on a brain-computer interface that allows individuals to consciously modulate specific brain activities, training their brains to function more efficiently. By promoting brain plasticity, neurofeedback training can lead to lasting functional changes in the brain, ultimately enhancing cognitive performance. However, despite its promise, neurofeedback has shown variable efficacy: studies report that between 30% and 50% of participants do not experience cognitive improvement after neurofeedback training. This variability among individuals poses a significant challenge to developing effective cognitive optimization strategies through neurofeedback. This internship project will test a novel hypothesis: that the quality of interoception—the processes by which the nervous system processes bodily signals—may be a key factor in determining the effectiveness of neurofeedback. Specifically, individuals with better interoceptive abilities may achieve greater control over brain activity during training, leading to improved learning outcomes. In contrast, those with lower interoceptive quality may struggle to benefit from neurofeedback, limiting its efficacy for them. Objective: The primary objective of the internship project is to determine whether the quality of interoception can serve as a predictor of neurofeedback efficacy. Method: The comparison between neurofeedback responders and non-responders will draw on an existing dataset of 106 participants, with multilevel measures of cardiac interoception (concurrent EEG and ECG enabling analysis of the heartbeat-evoked potential, heartbeat counting task performance, and self-reported interoception via the MAIA-2 questionnaire). Statistical models (cluster-based permutation, frequentist and Bayesian) and machine-learning techniques will be applied. The internship will be jointly supervised by Michael Quiquempoix, PhD, and Charles Verdonk, MD, PhD. Profile: A motivated student in neuroscience or cognitive science with programming skills (MATLAB, R and/or Python) and a solid understanding of standard statistical methods. Must be enrolled at Master 2 level and able to pass the primary security clearance required by the French Ministry of Army. Application deadline: October 1, 2026.

Supervisor

Michael Quiquempoix, PhD; Charles Verdonk, MD, PhD

Host lab

French Armed Forces Biomedical Research Institute (IRBA), Brétigny-sur-Orge (91)

Duration

6 months

Location

French Armed Forces Biomedical Research Institute, Brétigny-sur-Orge (91), France

When we listen to speech, the brain constantly anticipates what comes next. Recent work suggests that beta-band oscillations (~13 to 30 Hz) carry these predictions, but their effect on behaviour may stay hidden when listening is easy. The question this project tackles is what beta activity actually looks like in the brain during clear, natural speech, and what mechanisms organise it. The project builds on two recent papers from our group: a theoretical framework that classifies language-brain modelling approaches into four families (Bouton et al., 2026, Imaging Neuroscience), and a computational model that predicts specific neural signatures of beta-mediated prediction (Platonova, Dogonasheva, Giraud, Bouton, 2026). Both call for a direct empirical test on naturalistic neural data, which is what this internship will do. The student will use the open Podcast ECoG dataset (Zada et al., 2025) to ask: what mechanisms organise beta activity in the cortical language network during natural narrative listening? Each of the four model families of the taxonomy will be applied to the same recordings (temporal response functions, banded ridge regression with language-model embeddings and RSA, spectral Granger and phase-amplitude coupling, and HMM-MAR on multichannel beta). Profile. M2 in cognitive neuroscience, computational neuroscience, or a related quantitative discipline. Strong interest in language and oscillatory dynamics. Required: Python (MNE-Python an asset), comfort with signal processing and statistical modelling. No prior ECoG experience required. To apply. Send a CV and a short motivation paragraph to sophie.bouton@pasteur.fr.

Supervisor

Sophie Bouton

Host lab

Département d'Études Cognitives (DEC), ENS; affiliated with Institut Pasteur

Duration

6 months

Location

ENS, DEC, 29 rue d'Ulm, 75005 Paris, France

How the brain consolidates auditory memories during sleep remains poorly understood. A key open question is whether the hippocampus contributes to this process via sharp-wave ripple (SWR) events, the fast oscillatory sequences that drive memory replay in spatial tasks, or whether auditory cortex operates through distinct, cortex-intrinsic mechanisms. Addressing this question requires combining large-scale recordings across the auditory system and hippocampus with reliable vigilance-state monitoring. The intern will contribute to high-density electrophysiological recordings in freely-moving ferrets instrumented with Neuropixels 2.0 probes spanning primary auditory cortex (A1), non-primary auditory areas, and hippocampus, alongside olfactory bulb (OB) local field potential electrodes used for vigilance-state classification. The work will involve participating in recording sessions, preprocessing and spike-sorting electrophysiological data (Kilosort pipeline), and contributing to the analysis of sound-evoked and spontaneous population activity across wake and sleep states. A specific focus will be placed on identifying candidate SWR events in ferret hippocampus and assessing their temporal relationship to cortical reactivations. Skills gained: chronic in vivo electrophysiology, high-density silicon probe recordings, spike sorting, LFP analysis, sleep-state classification, supervised animal handling. Candidate profile: M2 student in neuroscience or biology, comfortable with quantitative data analysis (Python or Matlab); prior exposure to electrophysiology or signal processing is a plus.

Supervisor

Yves Boubenec

Host lab

Laboratoire des Systèmes Perceptifs (LSP), Audition team, École Normale Supérieure

Location

29 rue d'Ulm, 75005 Paris, France

Sleep exhibits structured organization into cycles of substages (wakefulness, NREM, and REM sleep) shaped by internal brain processes and external experience. Social context's influence on these transitions remains largely unexplored, particularly in carnivores whose sleep patterns resemble humans more closely than rodents. Ferrets—highly social with well-characterized auditory cortex and extended postnatal development—provide an ideal model for these investigations. The intern will conduct and analyze extended electrophysiological recordings (>24 hours) in freely moving ferrets equipped with olfactory bulb electrodes, using the lab's OB gamma oscillation-based vigilance classification method. Work will characterize sleep architecture's fine structure at naturalistic timescales, including substage distribution and sequencing, and quantify how proximity to and interactions with conspecifics modulate wake-to-sleep transitions and inter-substage dynamics, combining electrophysiological data with behavioral tracking. Skills Gained: Chronic in vivo electrophysiology in freely moving animals, LFP signal processing, sleep-state classification, behavioral tracking and quantification, longitudinal experimental design. Candidate Profile: M1/M2 student in neuroscience or biology with systems neuroscience and animal behavior interest; Python or Matlab familiarity expected.

Supervisor

Yves Boubenec

Host lab

Laboratoire des Systèmes Perceptifs (LSP)

Location

29 rue d'Ulm, 75005 Paris, France

The ability to perceive and categorize speech sounds develops early in life, well before language acquisition. The neural mechanisms underlying this early plasticity, and in particular the role of auditory experience in shaping the maturation of auditory cortical areas, remain poorly understood. Our group addresses these questions using the ferret as an animal model, whose protracted postnatal brain development makes it especially well suited for experimental investigation. The intern will participate in functional ultrasound imaging (fUSI) experiments conducted on young ferrets exposed to speech sounds across varying durations and exposure protocols. The work will include conducting fUSI acquisition sessions, processing and analyzing hemodynamic data, and contributing to the interpretation of exposure-dependent effects in relation to the cortical localization of responses across primary and non-primary auditory fields. Skills gained: functional brain imaging in vivo, neuroimaging signal processing, auditory functional neuroanatomy, supervised animal handling. Candidate profile: M1/2 student in neuroscience or biology, with an interest in systems neuroscience and in vivo experimental approaches.

Supervisor

Yves Boubenec

Host lab

Laboratoire des Systèmes Perceptifs (LSP)

Location

29 rue d'Ulm, 75005 Paris, France

At 6 months of age, infants prefer speech over non-speech sounds (e.g., Vouloumanos & Werker, 2004) and recognize that speech can communicate information without accessing its lexical content (Vouloumanos, Martin & Onishi, 2014). It has been argued that attentional biases for speech together with an abstract understanding of its communicative function predict language development. An open question is whether these early biases are specific to speech or whether they extend to other communicative signals. In a recent study, work from our lab (Gupta et al., submitted) shows that 18-month-olds preferentially attend to novel gestures than motor actions, much like the preference for speech over non-speech, suggesting that infants are tuned to detect communicative signals at large. But do they understand the value of communication in these signals or is it simply due to low-level attentional biases that make these signals special? Research environment: This project is part of the ERC ThoughtOrigins led by the PI. The candidate will join a dynamic team of PhDs and post-docs all involved in projects aiming at uncovering the format and structure of thought in the absence of language. The babylab is located within the lab but there is a possibility of staying in Paris and collaborating with a Paris-based baby lab for this project. Expected profile: Eligible applicants are currently enrolled in a master's degree or equivalent in a relevant field such as cognitive science, psychology, etc., with excellent academic results and some experience with programming languages (Python, MATLAB, R). The candidate should be fluent in French to communicate with participating families. They will be involved in all aspects of the project: experimental design, stimuli creation, scripting, data collection, data analysis, data presentation and writing.

Supervisor

Isabelle Dautriche, Shreejata Gupta

Host lab

CRPN (Centre de Recherche en Psychologie Cognitive et Neurosciences), Aix-Marseille University

Duration

4-6 months

Location

Aix-Marseille (CRPN babylab); possibility of staying in Paris with a Paris-based baby lab

Previous studies have shown that infants understand their first words as early as six months of age (Bergelson & Swingley, 2012). The question that then arises is: from what point can babies combine these different words to understand sentences? The current state of the literature shows that empirical data on infants' understanding of linguistic compositionality are limited and often difficult to interpret. While several studies indicate that children understand multi-word expressions from their second year, this ability can often be attributed to non-compositional strategies such as linear heuristics, statistical associations, or one-to-one mappings between forms and situations. This project therefore aims to explore when children transition from a primarily lexical understanding to a more compositional understanding of utterances, particularly in adjective–noun combinations.

Supervisor

Isabelle Dautriche, Eulalie Pequay

Host lab

ERC ThoughtOrigins; CRPN (Centre de Recherche en Psychologie et Neurosciences), Aix-Marseille University; babylab

Duration

4-6 months

Location

Aix-Marseille University (with possibility of Paris-based collaboration)

Two opposing perspectives on emotion's role in decision-making are examined: the Cartesian view emphasizing rational decision-making independent of emotion versus the Darwinian view highlighting emotions' value in option evaluation. The laboratory investigates mood fluctuations within these frameworks. Mood is defined as unidimensional, varying from negative to positive states. The MAGNETO computational model represents mood as the average of all possible action values, which are learned from their outcomes. This mood influences foraging decisions, with positive mood increasing likelihood of costly actions. Model simulations demonstrate adaptive advantages in correlated reward/cost environments. The M2 internship involves developing a behavioral task with three actions featuring oscillating contingencies—actions A and B correlated, action C anti-correlated. The project includes task development, testing healthy volunteers, and updating the computational model accordingly. Potential PhD continuation would explore neuroimaging mechanisms, psychiatric distortions, and sophisticated inference extensions.

Supervisor

Mathias Pessiglione

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

Long

Location

ENS, DEC, 29 rue d'Ulm, 75005 Paris

In noisy environments, our ability to spatially separate the voice we are trying to understand from competing sound sources can significantly improve intelligibility. This spatial unmasking is based on binaural listening, meaning we use both ears rather than the monaural signal in only one ear. Numerous models exist for predicting speech intelligibility. These models are crucial tools for evaluating and developing hearing aids and cochlear implants that restore some audibility to deaf people. Monaural intelligibility models cannot predict spatial unmasking. To address this issue, we recently proposed a preprocessor that processes binaural signals into enhanced monaural signals (in that the advantages of binaural listening are then taken into account) which can then be used by these models. The aim of this internship is to evaluate the performance of this preprocessor by designing a listening test to compare speech intelligibility under standard binaural conditions with that obtained for monaural sounds processed by the preprocessor. It will also involve varying the signal characteristics that influence intelligibility: the spatial configuration of the sources, the room reverberation level, and the type of competing sources. There are several possibilities for extending this topic in a PhD: improving the preprocessor and testing it on new data; comparing it to other intelligibility models; and predicting intelligibility for people wearing hearing aids and/or cochlear implants. Type of work: experimentation with audio signal analysis and processing (MatLab), implementation of controlled listening tests, statistical analysis of the data produced (R), publication. Pursuing this research into a PhD is possible.

Supervisor

Mathieu Lavandier

Host lab

Université de Lyon, ENTPE, LTDS; UMR CNRS 8248 - Équipe Audition (D. Pressnitzer) du DEC

Duration

5-6 months

Location

Vaulx-en-Velin (ENTPE), affiliation at ENS, DEC, 29 rue d'Ulm, 75005 Paris

Humans do not perceive space uniformly: the brain encodes space near the body differently from distant space. The area immediately surrounding the body is called peripersonal space (PPS). This space is particularly important at the behavioral level because it is through this space that we interact with the outside world. At least two functions are attributed to PPS: protecting the body and planning goal-directed actions. Studies have shown that the size of PPS is flexible and can be influenced by emotional, social, and motor factors. In particular, emotional states related to the perception of danger, such as anxiety, may modulate how this protective space is represented around the body. In this context, the aim of this internship will be to examine the influence of anxiety on the representation of PPS in healthy subjects. To do this, we will use the Threat-of-Scream paradigm, which induces anxiety through the presentation of unpredictable aversive stimuli (Beaurenaut et al., 2020), and a multisensory (audio-tactile) interaction paradigm to assess the size of PPS (Canzoneri et al., 2012; Hobeika et al., 2020). These effects will be compared between an anxiety-inducing condition and a neutral condition. This internship will take place between the Department of Neuroscience and Cognitive Science (IRBA, Brétigny-sur-Orge) and the Cognition, Action, Emotion Interaction Laboratory (Université Paris Nanterre). It will be supervised by Marine Taffou (IRBA) and Morgan Beaurenaut (Université Paris Nanterre). Responsibilities: setting up/designing the experiment; programming the paradigm; data collection and analysis; participation in cross-cutting laboratory tasks. Interested candidates should send a CV and cover letter to Marine Taffou (marine.taffou@def.gouv.fr) and Morgan Beaurenaut (morgan.b@parisnanterre.fr).

Supervisor

Marine Taffou, Morgan Beaurenaut

Host lab

Department of Neuroscience and Cognitive Science (IRBA, Brétigny-sur-Orge) and Cognition, Action, Emotion Interaction Laboratory (Université Paris Nanterre)

Duration

6 months

Location

Brétigny-sur-Orge and Nanterre, France

Le stage s'inscrit dans le cadre d'un projet de recherche interdisciplinaire porté par le consortium Fans4all et visant à améliorer la sécurité et l'accessibilité en aviation, notamment pour des pilotes dits HSI (pilotes privés, aviation légère, avec des difficultés d'audition et d'élocution). Afin de permettre à ces pilotes de voler de façon autonome en espace aérien contrôlé, le consortium FANS4all développe actuellement un équipement de cockpit inclusif et multisensoriel (vue haute, toucher, reconnaissance des gestes, communication adaptée). Le projet de stage associe recherche fondamentale (modélisation et analyse de la conscience situationnelle en vol) et recherche appliquée (développement et intégration de briques technologiques avec feedback utilisateur). L'objectif est de concevoir, développer et tester un prototype fonctionnel intégrant : la transcription vocale automatisée des communications de contrôle aérien ; l'analyse de l'information pour cartographier dynamiquement la situation aérienne ; le filtrage contextuel et personnalisé de l'information pertinente pour les pilotes sourds ; l'affichage et l'interaction utilisateur adaptées (UX/UI). Profil recherché : Master 2 Informatique, Interaction Homme-Machine, Sciences Cognitives, Ingénierie ou équivalent. Bonne maîtrise des outils de développement web/plurimédia (JavaScript/HTML/CSS, frameworks UX/UI, prototypage rapide), compétences de base en data science et IA (reconnaissance vocale, analyse sémantique, traitement du signal), intérêt pour l'accessibilité et l'expérience utilisateur. Modalités : Durée 6 mois (dates à adapter), lieu Uttop Tarbes, stage en présentiel, gratification 4.5€/h soit environ 630€/mois. Candidatures à adresser à : philippe.marrast@uttop.fr

Supervisor

Philippe Marrast, Etienne Ceretto, Cédrick Béler, Isabelle Bazet, Marie Bénéjean, Florian Hémont

Host lab

CERTOP, LGP/ICE (Uttop, Tarbes), Fans4All, PREFICS (Rennes)

Duration

6 mois

Location

Uttop, Tarbes (présentiel)

The medial olivocochlear (MOC) system represents a crucial neural feedback mechanism enabling dynamic brain control over auditory periphery sensitivity, particularly affecting outer hair cells. This system proves essential for hearing in noisy settings, though individual effectiveness varies significantly and declines with age. Precise MOC assessment is critical for understanding hidden hearing loss and establishing future clinical audiology diagnostic protocols. Internship Objectives: This position aims to develop and compare two non-invasive approaches measuring MOC system effects in humans. Empirical measurements will evaluate the MOC reflex across participant cohorts using objective methods based on transient evoked otoacoustic emissions and subjective psychoacoustic detection thresholds in forward-masking paradigms. Computational peripheral auditory system models will optimize experimental protocols and strengthen links between physiological and perceptual measurements. Candidates gain experience in human psychophysical and physiological auditory testing, experiment programming (MATLAB/Python), computational auditory modeling, and data analysis and statistical methods.

Supervisor

Emmanuel Ponsot

Host lab

Perception and Sound Design Team, STMS lab (UMR 9912), IRCAM, Paris

Duration

5-6 months

Location

IRCAM, Paris; ENS, DEC, 29 rue d'Ulm, 75005 Paris

Our research group is seeking a motivated Master's student to join a project investigating how early adverse experiences influence the development of mental and behavioral disorders in nonhuman primates. By integrating animal welfare and evolutionary psychiatry, the study seeks to uncover the evolutionary, ecological, and developmental foundations of psychopathology. Our main hypothesis is that human dimensional models of adversity—deprivation, threat, and unpredictability—can predict the emergence of psychopathologies in primates. To test these hypotheses, the study will classify early life experiences in primates and analyze their impact on socio-emotional skills and personality traits, gathered through behavioral observations and questionnaires. One key outcome will be the creation of an innovative adversity measurement scale. The selected student will work within a multidisciplinary team composed of experts in primatology, developmental psychology, and psychometrics. We particularly welcome students with backgrounds or strong interests in psychology, cognitive science, psychopathology or primatology. The internship can be arranged from fully on-site to fully remote. Funding opportunities may also be available through the ERASMUS+ program. To apply, contact comparativeminds@udg.edu with your CV and academic transcripts.

Supervisor

Thomas Castelain

Host lab

Comparative Minds, Department of Psychology, University of Girona (Girona, Spain)

Duration

4.5 months

Location

University of Girona, Spain (flexible: on-site to fully remote)

Our research group is seeking a highly motivated Master's student to join a project on multimodal communication in early childhood, with a specific focus on the role of gestures in peer interactions. The thesis will investigate how children between 3 and 6 years old use gestures when communicating with peers, and how their gestural repertoires (frequency, types, and integration with speech) evolve during this critical stage of social and linguistic development. While much existing research has focused on gesture–speech combinations in child–caregiver interactions or controlled experimental settings, far less is known about how gestures function in peer-to-peer contexts. By adopting a cross-sectional approach, the student will explore how gestures support complex peer interactions—such as argumentation and discourse structuring—during early socialization. The student will be trained in advanced annotation systems (e.g., M3D labeling system) and quantitative analysis techniques. We particularly welcome students with backgrounds or strong interests in linguistics, education, developmental psychology, primatology, or cognitive science. Prior experience in qualitative (ELAN software) or quantitative research methods will be considered an asset. The internship can be arranged from fully on-site to fully remote. Funding opportunities may also be available through the ERASMUS+ program.

Supervisor

Thomas Castelain

Host lab

Comparative Minds research group, Department of Psychology, University of Girona

Duration

4.5 months

Location

University of Girona (flexible: fully on-site to fully remote)

Our research group is seeking a motivated Master's student to join a project investigating gestural communication and lateralization in non-human primates. The study will examine lateralization in communicative gestures across several primate species housed in wildlife centers. Specifically, it aims to quantify lateralization at individual and population levels, assess the influence of rearing conditions on motor asymmetry, and explore potential evolutionary patterns of gestural lateralization across diverse taxa. The field of gestural communication is taxonomically biased: most research has focused on chimpanzees and other great apes, with fewer studies on African and Asian monkeys and only scarce investigations on neotropical monkeys. Broadening the taxonomic scope is essential to identify shared evolutionary traits and species-specific adaptations. The selected student will work within a multidisciplinary team and be trained in advanced observational techniques, behavioral coding, and the analysis of gestural repertoires and lateralization patterns. We particularly welcome students with backgrounds or strong interests in psychology, cognitive science, primatology or language sciences. The internship can be arranged from fully on-site to fully remote. To apply, contact comparativeminds@udg.edu with your CV and academic transcripts.

Supervisor

Thomas Castelain

Host lab

Comparative Minds research group, Department of Psychology, University of Girona

Duration

4.5 months

Location

Girona, Spain (flexible: on-site to fully remote)

Our research group is seeking a motivated Master's student to join our ongoing projects at the intersection of primatology, psychology, and artificial intelligence. Building on our recent work (Llorente, Guitton & Castelain, 2025), this project aims to investigate how concepts from primatology can be applied to the study and design of social robots. The selected student will develop an original thesis project focusing specifically on how large language models (LLMs) and generative AI systems perceive and represent the association between primatology and social robotics. The study will employ a combination of prompt engineering and content analysis, providing a unique methodological perspective on the current strengths and limitations of AI in bridging these domains. The student will work within a multidisciplinary team composed of experts in primatology, developmental psychology, and cyberpsychology. We particularly welcome students with backgrounds or strong interests in psychology, primatology, cognitive science, human–robot interaction, or artificial intelligence. The internship can be arranged from fully on-site to fully remote. To apply, contact comparativeminds@udg.edu with your CV and academic transcripts.

Supervisor

Thomas Castelain

Host lab

Comparative Minds research group, Department of Psychology, University of Girona

Duration

4.5 months

Location

Girona, Spain (flexible: on-site to fully remote)

Sleep is an essential biological process, critically involved in learning, especially in infancy. Throughout development, brain activity during sleep undergoes important changes before reaching its mature organization. In particular, it is initially characterized by two different stages: quiet and active sleep. While we know that sensory processing is disrupted during sleep, we still lack a quantitative description of how it reduces sensory responses in these two different sleep stages. The goal of this project is to investigate the impact of different sleep stages on infant auditory processing using EEG. The internship will consist in: literature review on the early development of sleep; analyses of an EEG dataset recorded from infants presented with speech sounds during their sleep; automated identification and characterization of sleep spindles; report writing. For any further question or discussion, please reach out (claire.kabdebon@cnrs.fr).

Supervisor

Claire Kabdebon

Host lab

DEC (Département d'Études Cognitives), ENS

Duration

6 months

Location

Paris, France

What are the psychological factors that drive charitable donations? Is there a psychological gap between rich and poorer donors in preferences for charitable interventions? These are the questions we sought to answer in a vignette study conducted on more than 700 U.S. participants. Consistent with predictions from the psychology of poverty literature, we found that socio-economic status drives preferences for and donations to unconventional, long-oriented charities with distant beneficiaries, and that this association is partly explained by different psychological characteristics (openness to experience, time discounting, trust) in donors. In the coming year, we plan to replicate this across more than a dozen countries, aiming at showing that there is a systematic psychological gap in charitable preferences according to socio-economic status cross-culturally. To achieve this, we will need to extend our experimental design to country-specific needs. We are also eager to understand whether a similar gap can be observed between donors' preferences and beneficiaries' needs. You can join this project now, and continue during the second semester, with the possibility of becoming an author of the final article. By joining our team, you will also be invited to the weekly ESC lab meetings (with pizza), and become part of lab life. Please contact us if you have any questions or would like to chat!

Supervisor

Coralie Chevallier

Host lab

Département d'Études Cognitives (DEC), ENS

Duration

6 months

Location

Paris, France

L'humour est un type de jeu présent dans toutes les cultures humaines (Apte, 1985 ; Lefcourt, 2001). Il peut être défini comme le fait de ressentir ou d'exprimer quelque chose de surprenant et de drôle, accompagné d'une réponse émotionnelle positive telle que la joie. Nos récents travaux ont montré que l'humour peut avoir une influence positive sur l'apprentissage chez les bébés (e.g. Esseily et al, 2015 ; di Stasi et al, 2025). Dans ce projet, nous mesurerons à l'aide d'une technologie avancée de suivi oculaire (eyetracking) la distribution de l'attention lors d'une démonstration d'une action humoristique ou neutre, et la mettrons en lien avec l'apprentissage par imitation chez des bébés. Nature de la mission : Recrutement actif et passations d'expériences en babylab avec des bébés âgés de 18 mois. Recueil et analyses de données comportementales et de suivi de regard (The Observer XT, Tobii Pro Lab, BabyFace Reader ; analyses sous R Studio ou Python). Le stage implique de travailler certains samedis pour les passations. Les passations auront lieu en majorité au BabyLab de l'Université Paris Nanterre, mais pourront également avoir lieu dans d'autres structures (PMI, Cité des bébés, crèche ou domicile des participants). Profils recherchés : étudiant.e en M2 ou césure extrêmement motivé.e, avec statut étudiant, en neurosciences, sciences cognitives, psychologie, éthologie, biologie du comportement. Une expérience préalable avec l'eye-tracker est un plus. Encadrement : Dr Lauriane Rat-Fischer et Prof. Rana Esseily. Durée : 5 à 6 mois à temps complet, entre janvier et décembre 2026. Rémunération : gratification de stage selon la législation en vigueur. Contact : Lauriane Rat-Fischer (l.ratfischer@parisnanterre.fr) et Rana Esseily (resseily@parisnanterre.fr).

Supervisor

Lauriane Rat-Fischer, Rana Esseily

Host lab

Laboratoire Éthologie Cognition Développement, Université Paris Nanterre

Duration

5 à 6 mois

Location

Nanterre, France

As human beings, we spend a large part of our mental life imagining possibilities—situations that are not real. We often consider alternative possible worlds or reflect on how different versions of the past might have influenced the present. Reasoning about possibilities is essential for human survival, as it allows us to predict the future, plan our actions, and make better decisions. Despite its importance, the origins of reasoning about possibilities have been debated for centuries. Currently, a major disagreement remains: is this kind of reasoning available to non-human animals and prelinguistic infants in the absence of language, or does it depend heavily on language? The latter hypothesis is supported by recent experimental findings: when an object is dropped into an inverted Y-shaped tube and participants must catch the ball, animals and children under the age of 4 consistently fail to use both hands to cover the two exits (Leahy et al., 2022; Redshaw & Suddendorf, 2016). However, the Y-tube task is challenging because it requires producing an optimal action regardless of which outcome occurs. We hypothesize that even younger infants and some animals might succeed if they were not required to produce such an action — a hypothesis supported by recent data on chimpanzees (Engelmann et al., 2021). As part of this project, we will develop a task that implicitly measures whether children and infants—before the development of a fully-fledged linguistic system—are capable of representing multiple possibilities when no action is required from them.

Supervisor

Isabelle Dautriche

Host lab

DEC (Département d'Études Cognitives), ENS

Duration

6 months

Location

Paris, France

As developmental science becomes more inclusive and ecologically valid, remote and online testing of infants is gaining momentum. Studying infants in their home environments holds promise: it reduces the burden on families, increases accessibility to underrepresented populations, and may yield more naturalistic behavior compared to lab settings. However, testing preverbal infants remotely comes with significant methodological challenges. Unlike older children or adults, infants cannot press buttons or follow verbal instructions, so we must rely on indirect measures such as gaze patterns. While online platforms are improving in their capacity to record and analyze infant gaze (e.g., via webcam), capturing sensitive indicators of language development such as speech sound discrimination remains difficult. This internship aims to pilot and evaluate new methods for conducting remote studies on early speech perception. Students will: review the existing literature on infant gaze-based paradigms; identify promising designs for remote implementation; select and adapt speech stimuli for online presentation; program and pilot an online experiment (e.g., using Lookit); conduct exploratory data collection and analysis to assess feasibility and sensitivity. This project sits at the intersection of developmental psychology, experimental design, and digital methods. Interns will gain hands-on experience in experimental programming, psycholinguistic theory, and webcam-based data collection.

Supervisor

Sho Tsuji

Host lab

Département d'Études Cognitives (DEC), ENS

Location

Paris, France

Young children are active communicators well before they become proficient speakers. Even in preverbal stages, they engage in complex turn-taking behaviors, and their development is shaped by the responsiveness of their social environment. Yet not all responses are equal: caregivers often switch between affiliative responses (e.g., emotional mirroring, encouragement) and informative responses (e.g., labeling, correction). The relative impact of these communicative styles on early learning and engagement remains surprisingly underexplored. This internship will investigate how different styles of social feedback from a virtual agent influence young children's communicative behaviors. Students will contribute to the development of an interactive agent capable of producing either affiliative or informative responses, potentially using generative AI or scripted interactions. We will then experimentally evaluate how these differing interaction styles shape children's vocalizations, gaze behavior, and/or learning outcomes. The project offers hands-on experience at the intersection of developmental psychology, human-computer interaction, and AI. Depending on interest and background, students may contribute to designing and prototyping the virtual agent's behavior, experiment design and study with child participants, annotating and analyzing recordings of child-agent interaction, and theoretical integration with frameworks of language acquisition, social learning, and joint attention.

Supervisor

Sho Tsuji

Host lab

Département d'Études Cognitives (DEC), ENS, PSL

Location

Paris, France

We are offering a paid six-month research internship during the academic year 2025-2026. The project focuses on Visual Perspective Taking (VPT), investigated through the N170 ERP component measured via EEG, with concurrent fNIRS recordings. The goal is to determine whether humans perceptually simulate what others see, using an innovative dual-participant paradigm featuring four conditions (active vs. passive observer and eyes open vs. closed), and test the hypothesis of an intermediate N170 amplitude when simulating another's viewpoint. This work builds on prior evidence that the N170 is sensitive to face inversion. Intern Missions: Set up and synchronize dual EEG + fNIRS recordings; program paradigm in MATLAB or Python; collect and analyze data (N170 amplitude/latency and fNIRS targeting FFA, OFA, STS regions); engage in shared lab duties across both laboratories. Required Skills: M1/M2 enrollment or equivalent in cognitive neuroscience; EEG/ERP and/or fNIRS experience; Python and/or MATLAB proficiency; strong motivation for social perception research; autonomy and scientific rigor.

Supervisor

Morgan Beaurenaut

Host lab

LICAÉ and Dysco labs, Paris Nanterre University

Duration

6 months

Location

Paris/Nanterre, France

When pairs of participants work together, an exchange of information is possible through different modalities (visual, auditory...). These sensory channels support communication, making it more efficient. Although the influence of the visual and auditory modalities has been widely described, the role of kinesthetic cues in the interaction between two participants remains little studied. Our current study aims to fill this gap and proposes to inspect the influence of the physical link, which translates into an exchange of haptic information, in the effectiveness of dyadic collaboration. To this end, pairs of individuals will be subjected to collaborative movement tasks by means of a robotic system allowing an exchange of haptic information only. We rely on measures of brain synchronization between two individuals (hyperscanning), an emerging metric established to be related to the quality of collaboration, and we assess the sense of joint agency. Expected role of the intern: support a PhD student in the collection, analysis, and interpretation of EEG hyperscanning data, structured around: a literature review on inter-brain synchrony and sensorimotor communication; participation in EEG hyperscanning experiments for data acquisition; analyzing EEG data using Python. Prior experience in EEG, robotics, or programming is appreciated but not a prerequisite. Interested candidates should email guillemet@isir.upmc.fr.

Host lab

ISIR (Institut des Systèmes Intelligents et de Robotique), Sorbonne University

Duration

3-6 months

Location

Paris, France

This internship examines how children and adults resolve linguistic ambiguity during sentence comprehension. The research focuses on understanding cognitive mechanisms underlying processing of ambiguous words—investigating whether individuals access multiple interpretations simultaneously or selectively. The project comprises two main research lines. First, it explores the timeline of mental access to possible word meanings, determining whether children have parallel access to all meanings for ambiguous words or selective access depending on syntactic structure. Second, it investigates how children identify relevant linguistic cues for resolving syntactic ambiguities and whether they can flexibly adjust importance attributed to different cues based on context. The overall goal is understanding how humans develop ambiguity resolution abilities and constrain interpretations using stored linguistic knowledge and real-time expectations, thereby advancing knowledge of language acquisition mechanisms.

Supervisor

Alex de Carvalho

Host lab

Département d'études cognitives (DEC), ENS

Location

Paris, France

Reading research establishes that 4-to-5-year-olds' speech perception abilities (i.e., phonemic awareness) predict reading readiness and the ability to read individual words. Moreover, teaching methods that promote phonemic awareness (e.g., Phonics) are effective at improving word reading. Notably however, fluid understanding of connected text often emerges later, around 7-9 years of age, the same age at which children become able to parse and interpret sentence ambiguities in spoken language in an adult-like manner. In this project we will ask: how well does a child's ability to understand and analyze spoken sentences relate to their ability to read sentences and other forms of connected text? No existing research examines children's real-time ability to deal with ambiguity in both listening and reading within the same children, while also assessing cognitive control skills. In this project, we aim to relate children's processing of spoken sentences to their ability to read connected text (by recording eye movements) and their ability to control their thoughts in non-linguistic tasks. Our goal will be to identify commonalities and differences in processing abilities in speech and text, within the same child, and identify cognitive control abilities that support speech and text comprehension.

Supervisor

Alex de Carvalho

Host lab

Département d'Études Cognitives (DEC), ENS

Location

Paris, France

I am pleased to invite applications for an M2 internship opportunity under my supervision at the LaPsyDÉ (CNRS/U Paris Cité). This internship will focus on language acquisition, with the specific topic tailored to align with the student's interests and the current research focus of my group. Our group investigates a wide range of questions in the field of language acquisition, and we are particularly interested in: decrypting the mechanisms children use to learn their native language; the role of executive function skills and socioeconomic status (SES) in oral and written language processing and acquisition; developing educational tools and strategies to fight social inequalities and to promote language acquisition and school success for all children. Key details: Location La Sorbonne, 46 rue Saint-Jacques, 75005 Paris; Duration from September 2025 until June 2026 (1 day per week in the first semester, full-time in the second semester); a background in cognitive science, linguistics, psychology, or a related field is preferred. To apply, send a CV, a brief cover letter, and contact information for one or two references to alex.de-carvalho@u-paris.fr.

Supervisor

Alex de Carvalho

Host lab

LaPsyDÉ (CNRS/Université Paris Cité)

Duration

September 2025 until June 2026 (approximately 9 months)

Location

Paris, France

The cochlea — the peripheral sensory organ for hearing — encodes sound into action potentials, which are then transmitted via the auditory nerve to the auditory brain. Biophysically realistic computational models of the cochlea and auditory nerve have been developed over several decades and are now well established. Unlike artificial neural networks, these models are grounded in biological realism, replicating the actual stages of sound processing and offering valuable insights into both normal auditory function and its dysfunctions. However, the use of these models has remained limited because very few simulate the auditory brainstem nuclei beyond the auditory nerve, and existing models do not reproduce field potentials such as EEG, which serve as critical biomarkers in diagnosing auditory disorders. The goal of this internship is threefold: implement state-of-the-art computational models that include upper components of the auditory system (e.g. the inferior colliculus); simulate field potentials based on the activity of individual neurons; and reproduce the altered field potentials observed in two mouse models developed by the team (one with auditory neuropathy and another in which deafness has been cured). Proficiency in programming (Matlab/Python) and signal processing is required, along with a strong interest in biology and sensory systems. This work would take place at the Institut de l'Audition, 63 rue de Charenton, 75012 Paris, and could start in mid-January/February 2026.

Supervisor

Boris Gourévitch, Nicolas Michalski

Host lab

Institut de l'Audition, Institut Pasteur

Duration

6 months

Location

Paris, France

We are looking for a motivated master's student (M1 or M2) to join the DreamTeam at the Institut du Cerveau to work on the SOMPARK project – a large, multi-centre observational study focused on sleep neurophysiology and clinical/cognitive profiles in patients with Parkinson's disease and REM Sleep Behavior Disorder (RBD), a known risk factor for developing Parkinson's or other neurodegenerative diseases. In this study, participants spend two nights and three days at the Pitié-Salpêtrière Hospital sleep clinic, undergoing polysomnography recordings and a battery of clinical and cognitive assessments. The internship will primarily involve: assisting with electrode setup for polysomnography recordings; acquiring data from attention and perception tasks with EEG; conducting preliminary data analysis. As the study includes a wide range of measurements, the exact focus of the analysis and final report can be tailored to the student's interests and skills. Requirements: conversational French (for interacting with participants); interest in sleep, fatigue, or clinical neuroscience; prior experience with programming is a plus. For further information, please contact: claudia.albero@icm-institute.org

Supervisor

Thomas Andrillon

Host lab

DreamTeam, Institut du Cerveau (Pitié-Salpêtrière Hospital)

Duration

2-6 months

Location

Paris, France

In western countries, people report holding fewer racist and sexist attitudes than a decade ago, yet discrimination and inequalities persist (Charlesworth and Banaji 2022). This issue can be examined at two interrelated levels: the systemic roots of racism and sexism, and the individual-level reproduction of inequalities through implicit associations between appearance or group membership and traits, beliefs or attitudes, which can bias behavior. At the individual level, experimental evidence suggests that people can exert some control over the expression of their implicit biases, yet the factors that facilitate or hinder this control are poorly understood. Our project aims at elucidating the cognitive mechanisms underlying implicit bias expression through the investigation of key contextual factors influencing bias control. We conduct experimental studies on first impressions of competence, using behavioral measures (choices and RTs) alongside computational modeling (drift-diffusion models) and physiological recordings (pupillometry and possibly EEG).

Supervisor

Julie Grèzes

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

6 months

Location

Paris, France (29 rue d'Ulm, 75005)

Emotions significantly influence decision-making, shaping how individuals perceive and respond to their environment. While there is consensus among emotion theorists that emotions and actions are interconnected, the precise nature of this relationship remains a subject of debate (Moors et al., 2017; Ledoux and Daw 2018). By integrating behavioural, physiological (e.g., skin conductance, cardiac activity), and possibly neural (EEG) measures, our project seeks to uncover the neurocognitive mechanisms underlying the influence of emotional displays on decision-making (see for example Sequestro et al. 2024).

Supervisor

Julie Grèzes

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

6 months

Location

Paris, France (29 rue d'Ulm, 75005)

With ageing, there is an increasing incidence of balance and mobility impairments, causing dramatic impact on health and quality of life. Despite extensive study, the laboratory and clinical measures of balance developed so far are poorly predictive of fall risk. Our previous work suggests this may be because such studies neglect older adults' inability to adapt postural control. In typical study designs, balance is assessed by exposing participants to repeated perturbations, and the first (often much stronger) response is discarded as the subsequent habituated response is analyzed. Yet in daily living, falling may occur after a single unexpected perturbation. The novelty in our approach is to study how participants adapt, from the first trial response to subsequent habituated responses. The goal of the project is to determine how young, healthy participants adapt sensorimotor control to alterations in context. Our modelling work shows that efficient balance requires controlling the Centre of Pressure (CoP) as a function of the position and speed of the Centre of Mass (CoM). When standing on solid ground, ankle proprioception provides a clear indication of CoM position; when the ground sways (such as on a boat), subjects must instead rely on visual and vestibular information. The goal is to measure how individual subjects adapt sensory integration to context. The project will take place at the Institut des Systèmes Intelligents et de Robotique (Sorbonne University). The applicant is expected to have a background in engineering and neuroscience, with previous experience in data analysis in Python. We are looking for a candidate interested in pursuing this project for a PhD on postural adaptation and fall risk in ageing.

Supervisor

Charlotte Le Mouel

Host lab

Institut des Systèmes Intelligents et de Robotique (Sorbonne University)

Duration

5 months

Location

Paris, France

We are recruiting two highly motivated Master 2 students for a 5–6-month internship in the Human and Artificial Perception group at the Institut de l'Audition. Our multidisciplinary team combines neurobiology, psychoacoustics and computational modeling to study how the brain represents and predicts complex sounds such as speech and music—and how these processes are altered in hearing impairment and other clinical populations. Project Themes (choose one): (1) Neurobiology of Music and Speech Perception — in vivo electrophysiology or neuroimaging (EEG/MEG) and psychophysics in healthy listeners; (2) Developing AI Models of Music Perception — design and train deep-learning or probabilistic models that predict melodic/harmonic expectations and compare model outputs with human data; (3) Sequence Predictions in Hearing Loss & Clinical Populations — behavioral and neurophysiological assessment of predictive coding in patients with cochlear implants or auditory neuropathy. Profile: currently enrolled in Master 2 (Neuroscience, Cognitive Science, Bioengineering, Computer Science, Physics); strong interest in auditory neuroscience, psychoacoustics or AI modeling; programming skills (Python, MATLAB or equivalent); excellent communication skills in English (French a plus). Location & Duration: Institut de l'Audition, Institut Pasteur (Paris 12e - Bastille); 5 to 6 months, starting September 2025 (flexible). To apply, send a single PDF (cover letter stating your preferred theme, CV, two references) to keith.doelling@pasteur.fr with subject line "M2 Internship – Human & Artificial Perception."

Supervisor

Keith Doelling

Host lab

Human and Artificial Perception group, Institut de l'Audition, Institut Pasteur

Duration

5–6 months

Location

Paris 12e - Bastille

The Institut Lyfe Research Center is dedicated to the study of the various factors (social, economic, cognitive and physiological) that underlie and influence human eating behavior. This internship is part of an interdisciplinary research project aimed at measuring the influence of gender stereotypes on food preferences. Recent research has shown that men consume significantly more meat, while women prefer vegetables, behaviors influenced by gendered social norms rather than actual nutritional needs. The project explores how certain cognitive mechanisms — notably the inherence heuristic (Cimpian & Salomon, 2014) and linguistic genericity (Berio & Musholt, 2023) — can rigidify food choices. The aim of this internship is to participate in the implementation of an experimental intervention in university restaurants in the Île-de-France region, in partnership with the CROUS, in order to promote a more varied, less gendered and more balanced diet. Mission: familiarization with the literature on gendered differences in eating behaviors and the cognitive/linguistic mechanisms involved in stereotype formation; participation in the implementation and monitoring of an intervention in several CROUS university restaurants; data collection and analysis (R / Python) based on behavioral measurements and questionnaires. Profile: Master 2 student in cognitive sciences or psychology, with solid skills in applied statistics (R, econometrics appreciated). Interest in gender issues, social psychology, nutrition and public policy appreciated. Remuneration: 4.35 euros/hour. Applications (M1 grades, CV and cover letter) to jlafraire@institutlyfe.com and adoracion.guzman-garcia@etu.ephe.psl.eu, subject line: M2 internship application - food behavior intervention.

Supervisor

Jérémie Lafraire; Adoración Guzmán-García

Host lab

Institut Lyfe Research Center (Écully, France)

Duration

6 months (September/October 2025 to June 2026)

Location

Paris, with travel to CROUS locations in Île-de-France

The aim of the internship is to better characterize the interactions between bodily signals and conscious perception: how does the neural processing of visceral signals, like heartbeats or the stomach's activity, influence our perception of the external world? The internship will include the development of a novel experimental paradigm as well as EEG data collection and data analysis. Programming skills will be required, with either previous experience or a willingness to develop such skills.

Supervisor

Catherine Tallon-Baudry

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

6 months

Location

Paris, France (29 rue d'Ulm, 75005)

Language represents possible alternatives. Modal information — information about what could be or must be the case as opposed to what the state of the world really is — is lexicalized in natural languages, using words such as could, should, might, possible, etc. Do modal concepts exist in the absence of language? Recent research suggests that preschoolers and great apes struggle with these concepts (cf. Leahy and Carey). The goal of the internship will be (1) to do a thorough literature review aiming to disentangle different properties of modals and/or task demands that may have interfered with preschoolers' and apes' capacity to represent possibilities, and (2) use a task with an implicit measure to prompt these concepts in infants and baboons (this is actually for two different M2 internships). The student working with infants will need a good command of French to communicate with the families. Both internships will take place in Marseille and can be conducted remotely for most parts (except testing).

Supervisor

Isabelle Dautriche

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

5 months

Location

Marseille, France

Tinnitus, often referred to as 'ringing in the ears,' is a disabling medical condition characterized by a subjective perception of sound without any external acoustic stimulus, affecting around 14% of the general population. Unlike auditory hallucinations, tinnitus sensations are generally of an unstructured acoustic nature, such as buzzing, hissing, or ringing. Analogous to phantom limb pain, subjective tinnitus is considered a phantom sound following hearing loss. However, hearing loss alone does not fully explain tinnitus perception: not everyone with hearing loss experiences tinnitus, while some individuals recover from tinnitus without noticeable changes in their hearing abilities, suggesting an interaction between the peripheral auditory system and central auditory processing. In a recent study, our team documented an attentional dysfunction linked to tinnitus in a large cohort that deserves to be further studied with EEG. It is a collaborative project involving ENT specialists and researchers of the Hearing Institute and the reconnect IHU. Internship objective: the student will be involved in an ongoing project aiming at advancing knowledge about the cognitive aspects and electrophysiological markers associated with tinnitus. An interest in attention modeling, EEG responses and statistics is welcome, but a background in these domains is not required. The internship will take place at the Institut de l'Audition (rue de Charenton, Paris 12) within the NeuroSpeech team.

Supervisor

Séverine Samson

Host lab

Institut de l'Audition (Hearing Institute), NeuroSpeech team

Duration

6 months

Location

Paris (Institut de l'Audition, rue de Charenton, Paris 12)

Psychology of human learning distinguishes implicit learning, resulting in unconscious knowledge, and explicit learning, a more effortful form of learning leading to conscious knowledge. Our research assesses whether these two routes to learning also exist in apes and monkeys, and how and when they appear during childhood. This internship will involve developing touchscreen-based experiments to probe implicit and explicit learning without relying on verbal responses, and data collection in children.

Supervisor

Raphaëlle Malassis

Host lab

LSCP & Babylab, ENS, Département d'Études Cognitives

Duration

5 months

Location

Paris, France (29 rue d'Ulm, 75005)

The project is aimed at investigating the diachronic development of phonological patterns over time, combining the analysis of diachronic linguistic corpora (e.g. literary text) with methods from psycholinguistics. The internship involves all (or a subset of) the following steps: establishing the corpus of relevant diachronic data; conducting a corpus analysis investigating the role of frequency effects on sound change (testing among others the Frequency Actuation Hypothesis); investigating the emergence of lexical restrictions; and investigating the implications of the results for phonological theories of the diachronic development of sound patterns. Some background in phonology is useful but not necessary. Basic knowledge and interest in the manipulation and analysis of written corpora are required.

Supervisor

Maria Giavazzi

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

2-6 months

Location

Paris, France (29 rue d'Ulm, 75005)

This project is led by Sho Tsuji and Alex Cristia at the LSCP, and will be co-supervised by May Abdou in the context of a PhD thesis. Our research focuses on uncovering the cognitive and linguistic mechanisms underlying early language acquisition and developing experimental tools to collect and analyse large-scale data. Language acquisition is deeply influenced by the social environment surrounding a child. Tech advances have provided a considerable amount of naturalistic, open-access data (audio & video recordings) that can give us insights into this dynamic context of language learning. We are developing a coding scheme to systematically extract the different types of linguistic information present in interactions with children, asking what kind of cues are communicated to them and whether these cues similarly manifest in both modalities of data, speech (audio) and behaviour (video). Who we're looking for: M2 students to annotate a chunk of English audio & video recordings, applying our coding scheme to a first batch of data. Theory-led analysis of a data subset is possible, with the opportunity to work on a Master's thesis or research paper. You will be integrated into a multidisciplinary team in Cognitive Science and Psycholinguistics (campus Panthéon). Requirements: basic experience with annotation tools such as ELAN; ensure the project aligns with your research interests and studies.

Supervisor

Sho Tsuji; Alex Cristia; May Abdou

Host lab

LSCP (Laboratoire de Sciences Cognitives et Psycholinguistique), ENS, Département d'Études Cognitives

Duration

4.5 months

Location

Paris, France (Campus Panthéon, ENS DEC, 29 rue d'Ulm, 75005)

Le travail de stage consistera à proposer et étudier des méthodes pour analyser et comprendre comment les archéologues étudient le matériel archéologique à travers ses activités perceptives et cognitives (regard, neuro, tactile). Ces méthodes s'appuieront sur des environnements de réalité virtuelle pour contrôler les contextes archéologiques et des instrumentations pour suivre l'activité de l'archéologue (eye-tracking, BCI, tactile-tracking). Le stage se déroulera au sein de l'équipe Seamless, à l'UMR IRISA, à Rennes, en collaboration avec un archéologue de l'Inrap (Institut National de Recherches en Archéologie Préventive). L'équipe Seamless (ex-Hybrid) a développé depuis plus de 10 ans une expertise dans le domaine du patrimoine culturel et de l'archéologie, notamment sur la production et l'utilisation d'environnements 3D interactifs, en réalité virtuelle et augmentée, appliquées à l'archéologie. Des travaux préliminaires ont déjà abordé la perception visuelle de l'archéologue avec une approche quantitative.

Supervisor

Ronan Gaugne

Host lab

Équipe Seamless, UMR IRISA, Rennes

Duration

6 mois

Location

Rennes, France

This project is led by Sho Tsuji and Alex Cristia at the LSCP and co-supervised by May Abdou in the context of a PhD thesis. Our research focuses on uncovering the cognitive and linguistic mechanisms underlying early language acquisition and developing experimental tools to collect and analyse large-scale audio data. Using naturalistic, open-access audio recordings, we are developing a coding scheme to systematically extract the different types of linguistic information present in interactions with children. Who we're looking for: M1/M2 students to annotate a chunk of English audio-recordings, applying our coding scheme to a first batch of data. Depending on progress and prior knowledge, analysis of a data subset is possible. You will integrate a multidisciplinary team in Cognitive Science and Psycholinguistics (campus Panthéon). Requirements: basic experience with annotation tools such as ELAN. No prior knowledge of Language Acquisition or Cognition is necessary.

Supervisor

Sho Tsuji; May Abdou

Host lab

LSCP (Laboratoire de Sciences Cognitives et Psycholinguistique), ENS DEC

Duration

1 month

Location

Paris, France (ENS, DEC, 29 rue d'Ulm, 75005)

The availability of a linguistic contrast between speech sounds (e.g. between [d] and [t]) in a given language is often modulated by the position (e.g. within the word) in which the sounds occur. One such example is the case of lexically stressed positions, in which contrast is more likely to be available than elsewhere in the word. This internship project is part of a larger research project on lexical stress, and it investigates the relative contribution of attentional and phonetic facts in shaping this distribution of phonological contrasts. The internship involves the programming (PsychoPy) of an already planned experiment, (online) testing of participants and data analysis (R).

Supervisor

Maria Giavazzi

Host lab

Département d'Études Cognitives (DEC), ENS, PSL University

Duration

2-5 months

Location

Paris

Animals emit signals at different rates. The call rate of vocal sequences may vary across species, and also within a species, when the rate may be indicative of varying degrees of emergency, proximity, quality of food, certainty of the information, etc. The goal of this internship is to review the literature on the topic to document precisely how call rate is used in the animal kingdom, and to reveal the phylogenetic history of this feature-meaning mapping. Suggested format: mini-internship. Learning/Tasks (depending on time): systematic literature review (manual or automated), construction of the database (species, baseline call rates, observed effect of varying call rates, reference), qualitative and quantitative analysis, reconstruction of the likely evolutionary history, writing up.

Supervisor

Ambre Salis and Emmanuel Chemla

Host lab

LSCP (Département d'Études Cognitives, ENS)

Duration

5 months

Location

Paris

Nous recherchons un.e étudiant.e motivé.e pour nous aider à l'acquisition de données dans le cadre de notre projet "Corels : La co-représentation est-elle boostée par l'expertise en langue des signes ?". Le but de ce projet est d'identifier si les compétences développées par le traitement d'informations visuelles et gestuelles lors de l'acquisition seconde d'une langue des signes (ex : compréhension des signes et marqueurs non manuels, anticipations d'actions, adaptabilité et sensibilité aux détails) facilitent la mise en place de mécanismes de co-représentation d'action — définie ici comme la capacité à se représenter les actions d'autrui, indispensable pour partager l'environnement social. En intégrant ce projet, vous participerez aux recrutements des participants, aux passations expérimentales et à l'analyse des données. Un travail bibliographique pourra être demandé sur la plasticité générée par l'acquisition de la langue des signes. Si le projet mène à une publication, votre contribution sera mentionnée.

Supervisor

Dr. Morgan Beaurenaut, Dr. Charlotte Hauser, Dr. Justine Mertz

Host lab

Laboratoire des Interactions Cognition Action Émotion, Université Paris Nanterre; Laboratoire Structures Formelles du Langage (CNRS, Université Paris 8); Institut Jean Nicod (ENS Paris)

Duration

42 jours à 7h/jour (~2 mois, flexible), de février à juin

Location

Paris (Université Paris Nanterre)

The goal of this internship is to pilot wearable devices to record children's language input and social interactions with the people around them. Over the past decade, wearable devices have been a game changer in developmental researchers' access to naturalistic recordings of children's learning environments. Insights gained from this line of research have unveiled the remarkable diversity of early language environments across communities and raised exciting questions about how environmental input shapes both language-general and language-specific learning pathways. While most work with wearable devices in fieldwork settings has focused on audio recordings, our team aims to incorporate multimodal recordings into fieldwork, including data streams such as video, acceleration, and heart rate. During this internship, you will: pilot various wearable devices in a lab setting; collaborate in designing and conducting experiments where caregivers of young children interact naturally; evaluate the feasibility and performance of different wearable devices for capturing high-quality, multimodal data. We welcome students with an interest in language acquisition and developmental research. A background in (psycho-)linguistics is not mandatory, but candidates should possess basic technical skills to handle and evaluate wearable devices, and a working knowledge of French, necessary for interacting with infants and their caregivers.

Supervisor

Sho Tsuji, Alejandrina Cristia

Host lab

Département d'Études Cognitives (DEC), École Normale Supérieure (ENS)

Duration

2-5 months

Location

Paris

One factor that contributes to children's differences in school performance is their own beliefs about intelligence. These beliefs are likely informed by children's early socialisation experiences, which vary across families and children. For example, some children may receive more frequent messages from their parents emphasizing their inherent intelligence or cleverness than others. Such differential reinforcement can contribute to shaping children's beliefs about their intelligence and, subsequently, impact their cognition, literacy, and educational achievement. This project aims to (a) review the empirical literature on the association between parents' perceptions of children's differences in intelligence, how parents express these perceptions in speech (e.g., in praise, criticism), and whether there is a link between parents' perceptions and children's beliefs about their own intelligence. A follow-up project (b) will use secondary data analyses of E-Risk to test, in an original empirical study using twin difference models, whether mothers' perceptions of their children's intelligence affect children's cognition, literacy, and educational achievement.

Supervisor

Sophie von Stumm

Host lab

ENS (École Normale Supérieure), Département d'Études Cognitives

Duration

2-3 months

Location

Paris

Background: The human brain can process single words during different tasks and via different modalities. We can hear the word 'apple' and repeat it out loud, or see the word written on a page and read it, or think of its concept and name it out loud. These three tasks (word repetition, reading and naming) require different computations at early stages, but share others at higher, amodal levels. Decades of research in neuropsychology have produced a detailed description of the various information-processing stages during each of these tasks, based on examination of patients with brain damage. Neuroimaging studies have isolated cortical regions putatively associated with several of these processing stages, but often target a specific component of the model in a specific modality. A similar limitation persists in computational work. In this project, we will develop a new experimental design to study the neural basis of these three tasks jointly, using fMRI and intracranial data. Throughout the project you will work closely with students developing deep neural models that simulate the three tasks jointly; predictions from the models will be tested against fMRI data. Main Goal: create a global map of the processing pathways in the brain for single-word processing and identify its amodal parts. We will test the hypothesis that the unified single-word processing model can be mapped onto the human language network, such that each processing component can be mapped onto dedicated brain regions.

Supervisor

Yair Lakretz, Emmanuel Chemla

Host lab

Département d'études cognitives (DEC), École Normale Supérieure (ENS)

Duration

6 months

Location

Paris

Incenteev is a French company dedicated to improving team performance and motivation. We offer innovative coaching and gamification solutions designed to stimulate engagement, collaboration, and productivity in the workplace. Over the past two years, we have consolidated studies on cognitive sciences and economic behaviors to understand what makes people happy at work. We have built a framework with 4 drivers managers can use to improve employee fulfillment: Transactional (sense of fairness), Individual (sense of autonomy), Mission (sense of purpose), and Social (sense of belonging). We are currently promoting this in a book titled "TIMS." We seek to partner with a university department by supervising a student's thesis. We envision this as a dynamic exchange of expertise, co-developing and conducting empirical studies exploring psychological and cognitive factors influencing employee well-being and motivation. We are open to doctoral theses examining how cognitive sciences, behavioral economics, or management practices influence workplace satisfaction. Your department gains access to our practical insights and real-world data; we benefit from your academic expertise and research methodologies.

Supervisor

Lila Haesebrouck

Host lab

Incenteev

Duration

One semester (to be defined)

Location

Paris

Le son émis par une source réelle est normalement perçu comme « externalisé », c'est-à-dire comme provenant de l'environnement autour de l'auditeur. Par contre, les sources reproduites au casque sont en général perçues comme « internalisées », le son semblant alors provenir de l'intérieur de la tête. La reproduction audio binaurale permet d'externaliser des sources sonores simulées au casque. Reproduire des sources externalisées est crucial pour le réalisme de toute application d'audio spatial et par extension de réalité virtuelle ou augmentée. Le but est ici de mieux comprendre et caractériser l'externalisation des sons reproduits au casque, en mettant en évidence les indices acoustiques sous-jacents à notre perception, mais aussi les aspects cognitifs qui seront étudiés en variant le contexte de la reproduction sonore. Il sera important de rester vigilant sur la différenciation de l'externalisation et de la distance perçue. Type de travail : expérimentation avec mesures de réponses impulsionnelles binaurales de salles réelles (MatLab) et/ou simulations en acoustique des salles (Catt-Acoustic), analyse et traitement de signal audio (MatLab), mise en place de tests d'écoute contrôlés, analyses statistiques (R), publication. La poursuite en thèse est possible, à discuter avec l'étudiant/e pour identifier des sources de financement.

Supervisor

Mathieu Lavandier

Host lab

UMR CNRS 8248 - Équipe Audition (D. Pressnitzer), Département d'Études Cognitives (DEC); ENTPE (Université de Lyon), LTDS

Duration

4-6 mois

Location

Paris (DEC-ENS) / Vaulx-en-Velin (ENTPE)

Our research focuses on the neuronal basis of auditory cognition, exploring how contextual factors influence auditory processing at the cortical level. We use ferrets as a key model to bridge our understanding of human cognition. We explore how ferrets process complex sounds, including speech, how attention shapes perception, and how early auditory experiences shape cortical development. By leveraging electrophysiology and functional ultrasound neuroimaging, we aim to uncover parallels between ferret and human auditory processing, with applications to attention, learning, and memory. Interns will (not exclusively) engage in neural data quantitative analysis, experimental work, and advanced decoding techniques. Internship can be conducted in French or English.

Supervisor

Yves Boubenec

Host lab

ENS, Département d'Études Cognitives (DEC)

Duration

6 months

Location

Paris

We are seeking a highly skilled M2 student with a strong background in analytic philosophy and a serious interest in environmental philosophy and sciences for an internship on the topic of "Perceiving Environmental Unnaturalness." Internship objective: The primary aim is to map the dual concepts of naturalness and unnaturalness as they are discussed in both philosophical and scientific literature. In this project, we adopt a specific interpretation of naturalness: the set of phenomenal and processual expectations associated with a particular natural environment. We are particularly interested in the perceptual, cognitive, and emotional dimensions that emerge when these expectations are not met, which we define as unnaturalness. The task involves connecting this conceptual mapping to an analysis of how perception (in a multimodal way), knowledge, and affect contribute to judgments of unnaturalness. This conceptual clarification will serve as a foundation for designing experimental tests, and the project also aims to link phenomena deemed unnatural to foundational concepts in environmental law and ethics.

Supervisor

Christian Lorenzi (LSP) and Sacha Bourgeois-Gironde (IJN)

Host lab

Laboratoire des Systèmes Perceptifs (LSP), Audition team; Institut Jean Nicod (IJN)

Duration

6 months

Location

Paris

During the first year of life, infants develop word segmentation capacities and start storing segmented word candidates in a 'protolexicon'. Previously, we showed that their sensitivity to statistical information leads them to include in this protolexicon frequently occurring non-word strings in addition to real words. Indeed, we found that 11-month-old French-learning infants recognize frequently occurring strings regardless of whether they are words (e.g., gateau 'cake', chaussure 'shoe') or non-words (e.g., dans la 'in the', va faire 'will do'), while they do not recognize disyllabic non-word strings that rarely occur in their input. In this study, we will experimentally test the prediction that infants initially consider all items in their protolexicon to be real words that should be mapped onto a meaning. The student should have an interest in language acquisition but a background in (psycho-)linguistics is not required. A working knowledge of French is necessary to interact with infants and their parents.

Supervisor

Sharon Peperkamp

Host lab

Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP); Language and its acquisition team

Location

Paris

Words are a central building block of language, and segmenting words out of continuous speech is a challenging step in early language acquisition. In spoken language there is no acoustic equivalent to the spaces that separate words in written language: spoken words are not separated by pauses. Infants start developing word segmentation abilities during the first year of life, relying on their sensitivity to both phonological and syllable co-occurrence cues to word boundaries. One phonological cue is vowel harmony: in many languages (e.g. Turkish, Hungarian, Somali), vowels within a word must share some articulatory property. Adults speaking a language with vowel harmony are known to rely on this phenomenon as a word boundary cue, and 7-month-old English-learning infants are also sensitive to vowel harmony as a cue to word boundaries, despite English having no vowel harmony. This project will extend that research by pitting vowel harmony against consonant harmony. Whereas vowel harmony is widespread, consonant harmony is cross-linguistically rare, yet consonants are more important than vowels in lexical processing. This study will test 7-month-old French-learning infants on both vowel and consonant harmony, to examine which is preferentially used as a cue for word segmentation. Depending on the duration, the student will be involved in several or all stages of the project, from experimental design and stimulus preparation to data gathering (in the ENS babylab) and analysis. A background in (psycho-)linguistics is not required, but a working knowledge of French is necessary to interact with infants and their parents.

Supervisor

Sharon Peperkamp

Host lab

Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP); Language and its acquisition team

Location

Paris

Can reading performance of dyslexic readers be improved with auditory rhythmic training? A key ingredient for reading is the mapping of graphemes (letters) to phonemes (sounds). The internship explores whether rhythmic brain activity misalignment causes developmental dyslexia. Research indicates a sampling deficit at 30 Hz in dyslexic adults may impair phonological processing. The RnDys project — a collaborative initiative involving speech therapists, the Iologo® training creators, and the Hearing Institute — tests whether auditory rhythmic training can realign auditory and brain rhythms to enhance reading performance in dyslexic children.

Supervisor

Sophie Bouton

Host lab

Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP); Cognitive development and pathology team

Duration

3 to 6 months

Location

Paris

Despite a number of studies, it remains unclear whether male and female human infants (0-6 months old) manifest different visual preferences or perceptual abilities. The goal of this project is to determine with greater confidence whether this is the case or not, and at what age such differences appear. In particular, we are interested in whether male and female infants show different spontaneous preferences for human faces vs. inanimate objects, as suggested by various studies (e.g., Lewis et al. 1966; Connellan et al. 2000), and whether they show differences in certain cognitive abilities, such as mental rotation (Enge et al. 2023). One part of this project (suitable for a Master internship or a PhD) is to carry out analyses of data from the YOUth cohort, a cohort of about 2000 babies from the Netherlands who underwent eyetracking and EEG experiments (face vs. object viewing, facial expression discrimination, gaze following, attention orientation) at 5 months, 10 months and 3 years old. Another part is to carry out a series of experiments testing sex differences in newborns using EEG in our babylab at the Port-Royal maternity hospital (suitable for a PhD only).

Supervisor

Franck Ramus

Host lab

Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP); Cognitive development and pathology team

Location

Paris

Testing hypotheses about learning mechanisms and cognitive disorders requires computational models. Reading models traditionally fall into two categories: minimalist connectionist approaches with crude results, and sophisticated symbolic/connectionist models relying on ad hoc specifications. The BRAID family of reading models overcomes these limitations. The Bayesian BRAID-Acq model simulates the second learning phase when children decode words independently. The first objective adapts this model to the initial phase, simulating grapheme-phoneme correspondence learning and comparing teaching methods, with predictions testable against behavioral data from beginning readers. The second objective simulates cognitive causes of dyslexia by modeling phonological, visual, and visual-attentional deficits, studying their consequences on learning and reading performance, ultimately generating testable predictions to distinguish cognitive theories of dyslexia. Candidate profile: strong background in programming, simulation, mathematical modeling, and experimental psychology. Python proficiency is essential. Knowledge of probability is advantageous.

Supervisor

Franck Ramus, Julien Diard, Sylviane Valdois

Host lab

Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP); Cognitive development and pathology team

Location

Paris