Oscillation-based models of speech perception postulate a cortical computational principle by which decoding is performed within a window structure derived by a segmentation process. Segmentation of syllable-size chunks is realized by a theta oscillator. We provide evidence for an analogous role of a delta oscillator in the segmentation of phrase-sized chunks. We recorded Magnetoencephalography (MEG) in humans, while participants performed a target identification task. Random-digit strings, with phrase-long chunks of two digits, were presented at chunk rates of 1.8 Hz or 2.6 Hz, inside or outside the delta frequency band (defined here to be 0.5 -2 Hz). Strong periodicities were elicited by chunk rates inside of delta in superior, middle temporal areas and speech-motor integration areas. Periodicities were diminished or absent for chunk rates outside delta, in line with behavioral performance. Our findings show that prosodic chunking of phrase-sized acoustic segments is correlated with acoustic-driven delta oscillations, expressing anatomically specific patterns of neuronal periodicities.
Acoustically driven cortical delta oscillations underpin prosodic chunking
Figure 8. Correspondence between behavioral data and electrophysiological data. Shown here are the 3rd order GMMs for Correct responses in the left hemisphere. The probability densities (with the ∫p(x)dx = 1) are displayed. The title of each panel shows three measures: (i) [dprime σ] –the behavioral performance; (ii) [Bias σ] –the average of the absolute difference between the mean of the prominent Gaussian component of the GMM and the driving acoustic chunk rate, and variance across the ROIs; and (iii) [P σ] --the average P-value (i.e. percentage of datapoints inside the frequency range of interest with respect to the total number of datapoints) and variance across the ROIs. Note the tightness of the PDFs in the 1.8Hz chunking condition compared to the pseudo-uniform shape of the PDFs in the 2.6Hz chunking condition, and the correlation between the decrease in dprime and the increase in Bias and the decrease in P. (the figure is taken from Rimmele, Poeppel, Ghitza, 2021, and the legend has been adapted from this reference)
Max Planck Institute for Empirical Aesthetics
