Ana Zappa

Language learning is often a social process, and social feedback may play a motivational role. We examined the neurophysiological correlates of word learning with feedback varying in reliability (accuracy) and social content. In a forced-choice task, participants learned to associate novel auditory words with known objects and received feedback. There were three types of feedback: Social Reliable (always providing accurate feedback regarding performance), Social Unreliable (providing random feedback: 50% correct and 50% incorrect feedback irrespectively of the performance), and Symbolic Reliable (always accurate feedback). Posttraining behavioral performance was better for words learned with social and symbolic reliable feedback. ERP amplitudes and pupil dilation showed differences as a function of feedback reliability and social content. In the reliable conditions, before feedback, stimulus-preceding negativity amplitude grew as learning progressed, likely due to the expectation of receiving positive feedback. During feedback, late positive complex amplitude for positive feedback diminished as learning progressed but not for negative feedback, which was likely consistently used for context updating. These effects were not observed for unreliable feedback, probably because its value was not used for updating information. Pupillometry results corroborated this showing greater dilation for negative versus positive feedback in reliable conditions. Finally, when feedback was social, processing was associated with more frontal activation and behavioral performance was closely correlated with both ERP and pupillometry results. Overall, our findings show differential processing of feedback depending on its informational and social content, advancing our understanding of how social and cognitive processes interact to shape word learning.

La cognition incarnée mise à l’épreuve : des preuves électrophysiologiques dans des environnements virtuels et classiques: Bien que la neuro-imagerie et les études comportementales aient montré un recrutement des systèmes sensorimoteurs lors du traitement sémantique, le rôle et le décours temporel de ces activations doivent encore être précisés. Notre objectif était donc d’observer les différents contextes dans lesquels l’activation motrice peut contribuer à la compréhension et à l’apprentissage du langage, en combinant l’électroencéphalographie (EEG) et des environnements interactifs et écologiques. Dans une première étude, les résultats en EEG ont indiqué un apprentissage rapide de mots en L2 ainsi qu’un effet de congruence de genre grammatical entre la L1 et la L2. Dans l’étude 2, associant un paradigme de compatibilité action-phrase (ACE) à l’EEG, nous avons observé une interférence provoquée par la compatibilité entre les processus moteurs et langagiers. Dans l’étude 3, les participants ont entendu des verbes d’action et ont effectué diverses actions dans un Cave automatic virtual environment (CAVE), pendant une tâche de Go-Nogo. Des analyses en temps-fréquence ont montré des activations motrices dans les deux conditions, et plus dans la condition Go, pendant le traitement des verbes d’action. Notre dernière étude est un registered report qui vise à déterminer les corrélats neuronaux de l’apprentissage incarné d’une deuxième langue en combinant l’EEG et la réalité virtuelle interactive. Le travail présenté dans cette thèse constitue un pas important vers une meilleure compréhension de la relation subtile entre les processus moteurs et sémantiques et ouvre de nouvelles perspectives pour prendre en compte la nature contextuelle de la façon dont nous comprenons et apprenons le langage.

The Music Ensemble dataset is a large-scale, cross-national database that provides detailed information about the musical, cognitive, personality, and demographic profiles of young adult musicians and nonmusicians. Data were collected from 1438 participants (aged 18–30) across thirty-five research sites in Europe, North America, South America, and Australia. Participants completed an in-person, in-lab battery of objective tests, including measures of verbal, visuospatial and musical short-term memory, executive functions (updating component), nonverbal reasoning, verbal comprehension, and music perception skills. The battery also included standardized and custom self-report questionnaires assessing music sophistication, music reward, personality traits, socioeconomic status, and demographic characteristics. Music Ensemble was preregistered, and the research protocol followed a standardized procedure across sites. The dataset also includes a large subsample of musicians and nonmusicians that are pair-matched for age, gender, and education (678 pairs). It enables well-powered investigations into the relationship between musical expertise and individual differences in cognition, personality, and demographic variables. It is also suitable for training in statistical and psychometric methods.

Social interaction can play a crucial role in how a second language (L2) is learned. In the current review, we examine theoretical frameworks and empirical studies demonstrating how social factors influence L2 learning, but we also identify gaps in the current literature. We propose using virtual reality (VR) as a methodology to fill these gaps with controlled, ecologically valid social simulations that can elucidate how social factors shape L2 learning.

Multimodal cues are crucial for real-life language comprehension. While gaze-following is a well-established mechanism in early language acquisition, its role in adult language comprehension - especially under challenging conditions - remains relatively unexplored. Here, we used a novel immersive and embodied virtual reality (VR) paradigm to examine whether referential gaze facilitates comprehension of unfamiliar spoken words in a noisy environment. Participants interacted with virtual agents whose gaze behavior varied in referential informativeness. Explicit feedback ensured that gaze-guided comprehension outcomes could be disentangled from informational content. Behavioral and eye-tracking data showed that participants identified spoken input more accurately when the teacher’s gaze reliably signaled the intended referent, and when they actively followed it. This facilitative effect of gaze emerged early and persisted across tasks. These findings highlight the enduring role of social gaze mechanisms in adult language comprehension, particularly in ambiguous or noisy contexts. They have important implications for a multimodal and situated language processing perspective, showing that gaze-following plays a meaningful role in real-world comprehension.

The multilab registered report by Grassi et al. (2025) represents the largest collaborative effort to date to estimate differences in short-term memory between musicians and nonmusicians, testing 1,200 participants across 33 units in 15 countries within a rigorously preregistered, open-science framework. Beyond providing precise effect-size estimates, the project exposed substantial diversity in how experts interpret the same dataset, particularly regarding the causal status and practical significance of musicians’ cognitive advantages. Here, we present six independent commentaries, each authored by a subset of researchers of the original team, that articulate these contrasting perspectives. Lima and Schellenberg argue that cross-sectional advantages are best explained by preexisting differences rather than training. Román-Caballero et al. emphasize small but reliable far-transfer effects of musical training on domain-general cognition. Zappa et al. call for greater caution in policy claims about music as a cognitive intervention, and Roncaglia et al. situate musical training alongside other forms of expertise (e.g., chess, physical exercise, bilingualism) as one of several routes to cognitive enhancement. Slevc highlights how coordinated multilab projects can help generate specific and testable predictions in a field that often lacks them, and Grassi and Talamini reflect on the broader methodological value of multilab initiatives for building a more accountable and replicable cognitive science. Together, these commentaries showcase productive theoretical pluralism and outline key directions for future research on musical training, cognition, and large-scale collaborative methods.

The Music Ensemble dataset is a large-scale, cross-national database that provides detailed information about the musical, cognitive, personality, and demographic profiles of musicians and nonmusicians. Data were collected from 1438 participants (aged 18–30) across thirty-five research sites in Europe, North America, South America, and Australia. Participants completed an in-person, in-lab battery of objective tests, including measures of verbal, visuospatial and musical short-term memory, executive functions (updating component), nonverbal reasoning, verbal comprehension, and music perception skills. The battery also included standardized and custom self-report questionnaires assessing music sophistication, music reward, personality traits, socioeconomic status, and demographic characteristics. Music Ensemble was preregistered, and the research protocol followed a standardized procedure across sites. The dataset also includes a large subsample of musicians and nonmusicians that are pair-matched for age, gender, and education (678 pairs). It enables well-powered investigations into the relationship between musical expertise and individual differences in cognition, personality, and demographic variables. It is also suitable for training in statistical and psychometric methods.

Language learning is often a social process, and social feedback may play a motivational role. We examined the neurophysiological correlates of word learning with feedback varying in reliability and social content. Participants associated novel auditory words with objects and received social (video clips) or symbolic (static images) feedback. In a forced-choice task, participants learned to associate novel auditory words with known objects and received feedback that was either Social Reliable (correct), Social Unreliable (random), or Symbolic Reliable (correct). Post-training behavioral performance was better for words learned with social and symbolic reliable feedback. Stimulus-preceding negativity (SPN) and late positive complex (LPC) ERP amplitudes, as well as pupil dilation, showed differences as a function of feedback reliability and social content. In the reliable conditions, before feedback, SPN amplitude grew as learning progressed, likely due to the expectation of receiving positive feedback. During feedback, LPC amplitude for positive feedback diminished as learning progressed but not for negative feedback, which was likely consistently used for context updating. These effects were not observed for unreliable feedback, probably because its value was not used for updating information. Pupillometry results corroborated these findings, showing greater dilation for negative vs positive feedback in reliable conditions. Finally, when feedback was social, processing was associated with more frontal activation and behavioral performance was closely correlated with both ERP and pupillometry results. Overall, our findings show differential processing of feedback depending on its informational and social content, advancing our understanding of how social and cognitive processes interact to shape word learning.