New preprint out on multimodal statistical learning!

New preprint out together with Davide Crepaldi!

In this project we examined multimodal integration during statistical learning. While human perception is largely multimodal and requires integration across sensory modalities, it is less clear whether and how humans integrate across modalities during statistical learning.

Participants were simultaneously presented with auditory and visual syllable streams that contained hidden pseudotriplets. In a subsequent test task, we showed that participants were able to track different types of statistics.

First, they tracked unimodal transitional probablities, extracting pseudotriplets in the auditory and visual modalities. Second, they also integrated across modalities to compute cross-modal transitional probabilities and extract cross-modal triplets. These results suggests that statistical learning can compute statistics cross-modally, which sheds new light on the long-standing issue of whether SL is a unified, domain-general mechanism.

Curious? Check out the preprint here and let us know what you think!