Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks

EEG evidence for statistical learning in sleeping newborns

There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.

Poster C98 in Poster Session C, Wednesday, October 25, 10:15 am - 12:00 pm CEST, Espace Vieux-Port

Lucas Benjamin¹, Ana Fló¹, Marie Palu¹, Shruti Naik¹, Lucia Melloni^2,3, Ghislaine Dehaene-Lambertz¹; ¹Cognitive Neuroimaging Unit, CNRS ERL 9003, INSERM U992, Paris-Saclay, France, ²Department of Neuroscience, Max Planck Institute for Empirical Aesthetics, Frankfurt am Main, Hesse, Germany, ³Department of Neurology, NYU Grossman School of Medicine, New York City, New York, USA

Since speech is a continuous stream with no systematic boundaries between words, how do pre-verbal infants manage to discover words? A proposed solution is that they might use the transitional probability between adjacent syllables, which drops at word boundaries to segment continuous sequences into word-like units. In a series of experiments, we used EEG to test this capacity and its limitation in sleeping neonates (one to three days old). For that, we developed and used EEG markers to assess statistical learning capacities in non-responsive population such as neonates. Finally, to explore the developmental trajectory of such cognitive capacity, we compared infants’ initial competences with adult behavioral performances in the same tasks. Our first result was to confirm that statistical learning is automatic enough to be efficient even in sleeping newborns. Using neural entrainment methods, we could follow the learning curve through time. It gave us an estimation of the neonates’ learning dynamic in such task. We also tested after learning memory representations of the pseudowords in multiple conditions. We showed that neonates specifically retained the first syllable of the extracted pseudowords. In a second experiment, we increased the difficulty of the segmentation task by investigating quadri-syllabic pseudowords instead of the usual tri-syllabic pseudowords always used in the literature. Crucially, the transition probability information was kept unchanged from our previous design. We revealed that despite successful tracking of transition probabilities in such sequence, neither neonates nor adults were able to successfully segment it. Strikingly, adding subtle prosodic cues such as subliminal pauses enabled both adults and infants to recover their segmentation capacities. These results showed that successfully tracking transition probability and segmenting pseudowords from a continuous sequence are two related but different processes. Thus, from birth on and despite immaturity, infants’ brains are equipped with adult-like tools, allowing them to extract coherent word-like units from auditory streams, based on the combination of statistical and auditory parsing cues.

Topic Areas: Language Development/Acquisition, Speech Perception