Dear Users

This Thursday (4-5 pm) in YNiC open plan, there will be a presentation on a "work in progress" MEG project.

This project is being carried out by Rebecca Millman and Philip Quinlan. The title of the talk is "The temporal envelope of speech is represented on multiple time scales".

Everyone is welcome to attend.

Best wishes Rebecca

Abstract: "The aim of this study was to determine the auditory cortical mechanisms that form the basis of representing the temporal envelope of speech in humans. The Asymmetric Sampling in Time (AST) model [e.g., Poeppel, D. (2003). The analysis of speech in different temporal integration windows: cerebral lateralization as “asymmetric sampling in time”, Speech Commun. 41: 245-255] proposes that speech perception involves multiple representations of the speech signal on at least two time scales. The AST model posits that the representation of speech is asymmetrical in the time domain as the left auditory areas preferentially integrate information from short (~20-40 ms) temporal windows whilst the right hemisphere homologues preferentially extract information from long (~150-250/300 ms) integration windows. Poeppel and colleagues, (e.g., Poeppel, 2003), suggest that temporal integration is reflected as oscillatory neuronal activity in at least two different frequency bands (theta, gamma). The AST model (e.g., Poeppel, 2003) hypothesises that the hemispheric lateralisation of speech processing results from speech driving gamma activity in the left hemisphere and theta activity in the right hemisphere. Magnetoencephalography (MEG) was used to determine whether changes in power in the canonical electrophysiological frequency bands (delta, theta, alpha, beta, gamma and high gamma) are used to represent speech processing. The speech stimuli were noise-vocoded single words. The attentional state (non-attend vs. attend) of the participants was manipulated. MEG beamformers were used to filter the data into the canonical frequency bands and localise the brain areas involved in processing the speech stimuli on the basis of these frequency bands. The results suggest that 1) the speech temporal envelope is represented on multiple time scales, 2) these time scales are commensurate with the canonical frequency bands corresponding to delta, theta and gamma activity, and 3) the attentional state of the participant modulates the spatiotemporal representation of the speech temporal envelope."