Learning regularities from different sensory modalities: Evidence for an a-modal learning mechanism

Abstract

How do we implicitly learn regularities from different modalities? Implicit learning refers to people’s ability to learn regularities in the environment automatically, unintentionally and without awareness. One central question is whether learning occurs separately within modalities (e.g., vision, audition) or whether learning is a-modal. Previous studies largely support separable modality-specific learning mechanisms. We explored this question using two prominent learning paradigms: statistical learning (SL), referring to learning of statistical regularities between elements such as transitional probabilities, and artificial grammar learning (AGL), referring to learning the underlying grammar of a set of exemplars. In Experiment 1 participants were familiarized with a structured stream of elements composed of reoccurring triplets, and subsequently were tested on their familiarity of these triplets. Learning occurred for visual, auditory and audiovisual triplets. Moreover, learning of the audiovisual triplets was significantly larger compared to learning of unimodal triplets. Similarly, using an AGL task, in Experiment 2 participants were familiarized with a set of sequences all adhering to a grammar. Participants could subsequently classify novel sequences as grammatical or not for both multimodal and unimodal sequences, with learning superior for multimodal compared to unimodal learning. In Experiment 3 we used a modified AGL task offering an alternative explanation for previous studies indicating modality-specific learning mechanisms. Two main findings are noted: (a) Implicit learning is amodal, learning both unimodal and multimodal regularities, and (b) multimodal is higher compared to unimodal learning. Together, these findings support the existence of an a-modal learning mechanism(s) sensitive to multimodal information.