Functional integration across brain regions improves speech perception under adverse listening conditions

Abstract

Speech perception is supported by both acoustic signal decomposition and semantic context. This study, using event-related functional magnetic resonance imaging, investigated the neural basis of this interaction with two speech manipulations, one acoustic (spectral degradation) and the other cognitive (semantic predictability). High compared with low predictability resulted in the greatest improvement in comprehension at an intermediate level of degradation, and this was associated with increased activity in the left angular gyrus, the medial and left lateral prefrontal cortices, and the posterior cingulate gyrus. Functional connectivity between these regions was also increased, particularly with respect to the left angular gyrus. In contrast, activity in both superior temporal sulci and the left inferior frontal gyrus correlated with the amount of spectral detail in the speech signal, regardless of predictability. These results demonstrate that increasing functional connectivity between high-order cortical areas, remote from the auditory cortex, facilitates speech comprehension when the clarity of speech is reduced.

Figure 1.

a, Illustration of the basic experimental design. Predictability (i.e., the inner semantic coherence of a sentence or how well one key word predicts the others) and intelligibility of the signal (through noise vocoding and thereby reducing the spectral detail) are varied orthogonally. b, Behavioral pretesting results unequivocally identified predictability to be most effective on speech comprehension at intermediate signal quality of eight-band noise-vocoded speech. ***p < 0.0001 [data based on n = 18 (replotted from Obleser, Alba-Ferrara, and Scott, unpublished observation)]. Error bars indicate SEM.

Figure 2.

Main effect of intelligibilty. Results from an F-contrast for the main effect of intelligibility (number of bands) is shown, yielding extensive bilateral clusters in temporal lobes, with peak voxels (contrast bar graphs for all 7 conditions are shown below) in the left and right STS that are not modulated by predictability. In the left hemisphere, the cluster also incorporates the inferior frontal cortex. The posterior parietal activations show the reverse pattern (decreasing activation with increasing intelligibility). All activations are p < 0.005; the cluster extent is >30 voxels. Error bars indicate SEM. lo, Low; hi, high; rot., rotated.

Figure 3.

Overview over differential activation for high- versus low-predictability sentences at intermediate (8-band) signal quality. The top and middle panels of activation overlays display the two conditions separately (in red and blue, respectively) compared with spectrally rotated speech. The bottom panels of brain overlays show the direct comparison of both. Bar graphs show activations in cluster peak voxels of direct comparison. BA 39, BA 8, and BA 30 (left panels) and BA 9 and BA 47 (right panels) all exhibit strongest activation through high-predictability eight-band speech. All activations are p < 0.005; the cluster extent is >30 voxels. Error bars indicate SEM. lo, Low; hi, high; rot., rotated; hi-pred, high predictability; lo-pred, low predictability.

Figure 4.

a, Overview over the changes in correlation between brain regions for the high-predictability eight-band condition (red) and the low-predictability eight-band condition (blue). Numbers in clusters indicate BAs of cluster peak voxels. All arrows shown indicate an increase in the positive correlation for the high-predictability condition, whereas no brain areas are more strongly interlinked for low predictability. ⁺p < 0.10; *p < 0.05; **p < 0.01; ***p < 0.001. The dashed blue lines indicate nonsignificant correlations. For details of changes in correlation (Δr), see Table 2. b, Scatterplots with linear regression slopes exemplifying the increase in the positive correlation between the left posterior temporal/inferior parietal cortex (angular gyrus, BA 39; x-axis) and left dorsolateral prefrontal cortex (superior frontal gyrus, BA 8; y-axis). Values plotted are condition-specific first eigenvector time series from the cluster identified in random-effects group analysis averaged across subjects (see Materials and Methods).