On the Role of Neural Oscillations Across Timescales in Speech and Music Processing

Abstract

This mini review is aimed at a clinician-scientist seeking to understand the role of oscillations in neural processing and their functional relevance in speech and music perception. We present an overview of neural oscillations, methods used to study them, and their functional relevance with respect to music processing, aging, hearing loss, and disorders affecting speech and language. We first review the oscillatory frequency bands and their associations with speech and music processing. Next we describe commonly used metrics for quantifying neural oscillations, briefly touching upon the still-debated mechanisms underpinning oscillatory alignment. Following this, we highlight key findings from research on neural oscillations in speech and music perception, as well as contributions of this work to our understanding of disordered perception in clinical populations. Finally, we conclude with a look toward the future of oscillatory research in speech and music perception, including promising methods and potential avenues for future work. We note that the intention of this mini review is not to systematically review all literature on cortical tracking of speech and music. Rather, we seek to provide the clinician-scientist with foundational information that can be used to evaluate and design research studies targeting the functional role of oscillations in speech and music processing in typical and clinical populations.

Keywords: cortical entrainment; cortical tracking; electrophysiology; music processing; neural oscillations; neurogenic communication disorders; speech processing; speech tracking.

FIGURE 1

Different analysis metrics used to investigate oscillatory contributions to speech and music perception. (A) Cross correlation between stimulus envelope (lag time series) and EEG extracted for sentences. (B) Left panel shows time-frequency representation of inter-trial phase coherence for a 6 Hz modulated tone showing stronger values corresponding to the modulation frequency. Right panel shows single trial phase values collapsed across time. (C) Multivariate TRF modeling of multiband stimulus envelope to map a kernel function onto the EEG. The observed EEG, multivariate TRF predicted EEG (at a single electrode) and their correlation at each electrode is shown (adapted from Dial et al., 2021). (D) Time-frequency representation of cross frequency coupling showing phase amplitude coherence between the theta phase and beta power (above), and theta phase and gamma power (below) in adults who do (AWS) and do not stutter (AWNS) (Sengupta et al., 2019).