This study investigates the cerebellar contribution to language processing by developing a biologically constrained cerebellar artificial neural network (CANN) model. The CANN, incorporating a recently identified cerebello-cerebellar recurrent pathway, demonstrates the emergence of both word prediction and syntactic recognition within a single circuit. The recurrent pathway proves crucial for both functions, a finding robust across various CANN variants with additional biological constraints. This suggests a unified cerebellar computational basis for prediction and grammar-like rule extraction from sequences, fundamental to cerebellar motor and cognitive functions.
Publisher
Nature Communications
Published On
Jan 31, 2024
Authors
Keiko Ohmae, Shogo Ohmae
Tags
cerebellum
language processing
artificial neural network
word prediction
syntactic recognition
biological constraints
cognition
Related Publications
Explore these studies to deepen your understanding of the subject.
