Studying building blocks of statistical learning: Automatic computation of auditory predictability

The project investigates statistical learning (SL) and computation of auditory predictability (CAP) in infants and adults. The aim is to determine whether CAP may be an underlying neural mechanism of SL.

Infants are sensitive to regularities in the auditory input, and pick up on them from brief and passive exposure. This ability contributes to various aspects of language development, for example speech segmentation and grammar learning. SL of regularities in the input is thus an important ability for language development.

The adult brain quickly picks up on how predictable the auditory environment is. In complex auditory environments, this helps to make sense of overlapping sounds by decoding which sounds come from the same or different sources. CAP thus contributes to general auditory perception.

This project will run for three years and use electroencephalography (EEG) to investigate CAP. In adults, the CAP response will be examined with more language-like regularities and stimuli than has been done previously, and linked to performance in SL tasks. In infants, the scope and limitations of CAP as previously studied in adults will be investigated, and CAP variations in response will be linked to performance in SL tasks.

The project has the potential to establish whether a general-purpose neural mechanism (CAP) potentially underlies an ability linked to language development (SL), thus contributing to the understanding of the human language capacity.