Research lines

Language comprehension involves more than simply integrating incoming information: we actively predict what comes next at multiple levels of the linguistic hierarchy, making processing and communication more efficient and fluid. During my PhD at the University of Padova under the supervision of Francesca Peressotti, I investigated how the brain generates lexical predictions (the anticipation of specific upcoming words) during sentence comprehension and the possible overlap with planning processes in word production. Using sentence-constrained comprehension and production paradigms and EEG recordings, I investigated time-frequency dynamics and event-related potentials related to anticipatory lexical processing and planning processes during word production, in both typical adults and adults with persistent developmental stuttering, aiming at establishing a possible contributing role of the language production system in generating predictions during comprehension.

During my postdoctoral work under the supervision of Francesca Peressotti and with the collaboration of Francesco Vespignani, I expanded my research on prediction to the audiovisual modality. First, I studied deaf individuals with cochlear implants as part of the PRIN 2017 project "The role of cochlear implantation and bimodal bilingualism in early deafness: a window into the neurofunctional mechanisms of human language" (local PI: Francesca Peressotti) funded by the Italian Ministry of University and Research. Subsequently, I investigated normal hearers under different listening effort conditions (speech in noise in high vs low signal-to-noise ratio) as part of my project "Predictive Brain in Audiovisual Speech Comprehension" funded by Fondazione CARIPARO. These studies investigate how prediction operates in populations with altered sensory systems and under different processing constraints.

Under this topic, I have also contributed as co-author to investigations of predictive processing in language and cognition, including a meta-analysis of neural activation in predictive processing across cognitive domains (Cristiano Costa and Cristina Scarpazza, University of Padova), speaker-specific phonological prediction in controlled experimental paradigms (Marco Sala, University of Padova), and viseme prediction in naturalistic audiovisual speech comprehension (Giorgio Piazza, University of Padova) as collaborator in the PRIN 2022 project "Perceiving and predicting multisensory speech: a window on the interplay between sensory and motor processes and brain representations", which I contributed to conceptualize and write. As part of the PRIN 2022, I'm also collaborating on decoding of phonological prediction in auditory and motor brain regions with the collaboration of Stefania Benetti, Francesco Pavani, Moritz Wurm and Giuseppe Rabini (CIMeC, University of Trento).

PUBLICATIONS:

• [in preparation] Gastaldon, S., Gheller, F., Piazza, G.*, Sala, M.*, Casalino, L.°, Battistutta, M.°, Vespignani, F., Peressotti, F. (in preparation). Multi-talker chatter noise intensity modulates audiovisual speech prediction

• [in preparation] Piazza, G., Vespignani, F., Gastaldon, S., Peressotti, F. (in preparation). Dynamic Hierarchies of Multimodal Prediction: How the Brain Predicts Visual Speech During Naturalistic Communication

• [preprint, under review] Gastaldon, S., Gheller, F., Bonfiglio, N., Brotto, D., Bottari, D., Trevisi, P., Martini, A., Vespignani, F.*, Peressotti, F.* (2026, v1). Chronic acoustic degradation via cochlear implants alters predictive processing of audiovisual speech. bioRxiv. DOI: https://doi.org/10.64898/2026.01.25.701504 - under review in Cortex 

• Sala, M., Vespignani, F., Gastaldon, S., Casalino, L., & Peressotti, F. (2025). In the Words of Others: ERP Evidence of Speaker‐Specific Phonological Prediction. Psychophysiology, 62(9), e70135. DOI: https://doi.org/10.1111/psyp.70135  

• Costa, C., Pezzetta, R., Masina, F., Lago, S., Gastaldon, S., Frangi, C. Genon, S., Arcara, G., Scarpazza, C. (2024). Comprehensive Investigation of Predictive Processing: A Cross- and Within-Cognitive Domains fMRI Meta-Analytic Approach. Human Brain Mapping 45(12), e26817. DOI: https://doi.org/10.1002/hbm.26817 

• Gastaldon, S.*, Bonfiglio, N.*, Vespignani, F., Peressotti, F. (2024). Predictive language processing: integrating comprehension and production, and what atypical populations can tell us. Frontiers in Psychology, 15, 1369177. DOI: https://doi.org/10.3389/fpsyg.2024.1369177 

• Lago, S., Pezzetta, R., Gastaldon, S., Peressotti, F., Arcara, G. (2023) Trial-by-trial fluctuations of pre-stimulus alpha power predict post-stimulus amplitude of language ERPs. Psychophysiology, e14388. DOI: https://doi.org/10.1111/psyp.14388 

• Gastaldon, S., Busan, P., Arcara, G., Peressotti, F. (2023) Inefficient speech-motor control affects predictive speech comprehension: atypical electrophysiological correlates in stuttering. Cerebral Cortex 33(11): 6834–6851. DOI: https://doi.org/10.1093/cercor/bhad004 

• Gastaldon, S., Arcara, G., Navarrete, E., Peressotti, F. (2020) Commonalities in alpha and beta neural desynchronizations during prediction in language comprehension and production. Cortex 133: 328-345. DOI: https://doi.org/10.1016/j.cortex.2020.09.026 

My investigation of predictive processing in adults who stutter sparked a broader interest in this neurodevelopmental fluency condition. In particular, my interest is in how the speech-motor deficit impacts aspects of speech perception to shed light on motor-auditory interactions. I have since consolidated existing collaborations and established new partnerships to expand the range of phenomena under investigation. In particular, a study I led in collaboration with Pierpaolo Busan, Nicola Molinaro, and Mikel Lizarazu opened the way to investigating cortical tracking of speech in this population, planting the seeds of ideas on cortical tracking that later developed into an independent project awarded with a Marie Skłodowska-Curie Actions Postdoctoral Fellowship (see below). 

I am currently involved as a collaborator in research led by Pierpaolo Busan (University of Trieste), also in collaboration with Mikel Lizarazu (BCBL), examining neural mechanisms underlying stuttering.

PUBLICATIONS

• Gastaldon, S., Busan, P., Molinaro, N., Lizarazu, M.  (2024). Cortical tracking of speech is reduced in adults who stutter when listening for speaking.  Journal of Speech, Language and Hearing Research, 67(11), 4339–4357. DOI: https://doi.org/10.1044/2024_JSLHR-24-00227  

• Gastaldon, S., Busan, P., Arcara, G., Peressotti, F. (2023) Inefficient speech-motor control affects predictive speech comprehension: atypical electrophysiological correlates in stuttering. Cerebral Cortex 33(11): 6834–6851. DOI: https://doi.org/10.1093/cercor/bhad004 

I approached the study of speech processing in deaf individuals with cochlear implants from the perspective of predictive processing, in particular in the audiovisual modality. While cochlear implants represent impressive advances in allowing to experience audition in profound deaf individuals, the signal encoded by the implant is profoundly different from that conveyed by an healthy cochlea. This research investigates how this chronic degradaton impacts prediction mechanisms and how visual speech cues (mouth movements) modulate these processes. This research line allowed me to collaborate with medical experts and pracitioners in otorhinolaryngology (Davide Brotto, Patrizia Trevisi and Alessandro Martini, Azienda Ospedale-Università Padova) and experts in cognition in deaf individuals (Davide Bottari, IMT Lucca).

To further contribute to this line of research, thanks to a Language & Linguistics Research Grant awarded by Gorilla™ Experiment Builder, I am currently developing an online study to investigate differences in predictive processing between normal hearing listeners and cochlear implant users under clear and noisy listening conditions in order to understand the role of prediction under effortful listening as a function of sensory experience.

PUBLICATIONS:

• [preprint, under review] Gastaldon, S., Gheller, F., Bonfiglio, N., Brotto, D., Bottari, D., Trevisi, P., Martini, A., Vespignani, F.*, Peressotti, F.* (2026, v1). Chronic acoustic degradation via cochlear implants alters predictive processing of audiovisual speech. bioRxiv. DOI: https://doi.org/10.64898/2026.01.25.701504 - under review in Cortex 

This line of research is part of my Marie Skłodowska-Curie Actions Postdoctoral Fellowship project (FlexyRhythm; DOI: https://doi.org/10.3030/101206531) 

Speech exhibits quasi-rhythmic features across languages, shaped by biomechanical, physical, and neural factors, limiting this variation to a specific syllabic rate, with a peak around 3-5 Hz. These semi-regularities enable the brain to align to the speech stream, generating temporal predictions that aid in parsing the auditory stream and identifying linguistic boundaries. Interestingly, the ability to synchronize speech production with an external rhythms predicts receptive speech abilities. However, current research focuses on isochronous (regular, periodic) rhythms, while everyday speech contains fine-grained irregularities in timing due to lexical, contextual, and individual factors. Such irregularities may be instead highly informative and preferred by the human language-ready brain.

FlexyRhythm addresses this gap by assessing flexible rhythm production (FRP) abilities, evaluating how individuals reproduce and synchronize to non-isochronous rhythms that are more representative of naturalistic speech, introducing this restricted variation. Then, the project will investigate how such rhythmic abilities relates to four main language states: statistical learning in an artificial language as proxy for language acquisition, native language processing, second language learning, and language processing in individuals with acquired speech-motor deficits following brain damage after stroke. This comprehensive approach connects FRP abilities across the "life of the speaking brain", from acquisition and learning to deterioration.

Using MEG and cortical tracking of speech (CTS) as a neural proxy for processing efficiency, FlexyRhythm will explore how FRP relates to the brain's ability to align with speech streams and encode linguistic information. Connectivity between motor and auditory brain regions will also be investigated, hypothesizing that individuals with superior FRP abilities exhibit more efficient neural connectivity. By employing state-of-the-art analytical techniques to link FRP abilities with neural measures of speech processing in continuous, naturalistic speech, FlexyRhythm pioneers a systematic evaluation of flexible rhythm abilities beyond conventional periodic entrainment.

This project will be deloped under the supervision of Nicola Molinaro (BCBL), in collaboration with other scientists from BCBL (Simona Mancini, Mikel Lizarazu) and other institutions (Vitória Piai, Donders Centre for Cognition, Radboud University as secondment supervisor; Sonja Kotz, Maastricht University as external advisor).

Preregistration work package 1a (FlexyRhythm WP1a): https://osf.io/btnzf/overview 

Dogs have co-evolved with humans for thousands of years, and research shows that some morphological and behavioral traits (e.g., visible sclera, snout muscles allowing facial expressiveness) have evolved to adapt to human communicative constraints and facilitate interspecific interaction. Strikingly, the auditory cortex of domestic dogs activates strongly in response to human voice, especially when exposed to dog-directed speech, a speech register that shares many acoustic properties with infant-directed speech. Additionally, dogs appear able to build a limited vocabulary, displaying the capacity to attach referents to speech sounds.

As the owner of a dog from a primitive breed (Shiba Inu), I developed an interest in understanding this interspecific communicative interplay. More specifically, I am interested in identifying which characteristics of human speech are most informative to dogs and in uncovering the neural mechanisms underlying their processing and comprehension of human speech. To explore these questions, I collaborate with Paolo Mongillo and Lieta Marinelli from the Applied Ethology Laboratory at the University of Padova, Italy.

Currently, at the BCBL, I am collaborating with Valeria Bevilacqua (visiting PhD student from the Applied Ethology Lab, University of Padova) and Antje Stoehr (BCBL Staff Scientist) to study the acoustic properties of dog-directed speech.

• Speech and language in acquired brain damage (stroke, tumor) (Giorgio Arcara, University of Padova & IRCCS San Camillo Hospital; Valentina Bambini, IUSS Pavia). Currently wrapping up a study on pragmatic competence in individuals with left- and right-hemisphere stroke.

• Linguistic alignment in multimodal large language models (Giulia Calignano and Alberto Testolin, University of Padova):

• EEGManyAnalysts Project (Katharina Paul and Jan Wacker, Universität Hamburg, Germany). EEGMANYANALYSTS is an innovative project that focuses on the sources of variance in data analysis routines in EEG research. Unlike previous studies that often explored within-subject effects, EEGMANYANALYSTS focuses on the analysis of inter-individual differences. In addition, the project aims at understanding how choices and characteristics of analysts impact the replicability of findings.

• Cortical tracking of speech and brain stimulation (Antonio Criscuolo and Sonja Kotz, Maastricht University)