Zhao Xu1,2 · Qinhui Yao1,2 · Yuwei Jiang1,2

1 Institute of Science and Technology for Brain‑Inspired Intelligence, Fudan University, Shanghai 200433, China 

2 Key Laboratory of Computational Neuroscience and Brain‑Inspired Intelligence, Ministry of Education, Fudan University, Shanghai 200433, China

Abstract

Hierarchical auditory processing is crucial for human language perception. However, the spatiotemporal characteristics of frequency-specific neural dynamics and the representation of abstract auditory properties, such as sequential structures and probability attributes, during auditory processing remain insufficiently understood. We used magnetoencephalography and a hierarchical auditory sequence paradigm to investigate these characteristics. Our findings delineated the bidirectional connections underlying auditory processing across multiple timescales, as reflected by frequency-specific directed information flow, and revealed temporal alternations in the direction of low-frequency oscillatory activity. Using multivariate decoders applied to low-frequency signals, we decoded tone- and sequence-level regularities along cortical pathways in temporal order, demonstrating the progressive representation of hierarchical auditory information. Furthermore, frequency-specific decoding of neural activity in temporal and memory-associated regions disentangled the sequential and probabilistic properties integrated during abstract auditory processing. Collectively, these results advance our understanding of the mechanisms underlying information sharing and the integration of abstract auditory information during perception.

Keywords

Magnetoencephalography; Auditory sequence; Decoding; Frequency-specific dynamics; Hierarchical processing

[SpringerLink]