Semantic Predictability Facilitates Comprehension of Degraded Speech in a Graded Manner

Previous studies have shown that at moderate levels of spectral degradation, semantic predictability facilitates language comprehension. It is argued that when speech is degraded, listeners have narrowed expectations about the sentence endings; i.e., semantic prediction may be limited to only most highly predictable sentence completions. The main objectives of this study were to (i) examine whether listeners form narrowed expectations or whether they form predictions across a wide range of probable sentence endings, (ii) assess whether the facilitatory effect of semantic predictability is modulated by perceptual adaptation to degraded speech, and (iii) use and establish a sensitive metric for the measurement of language comprehension. For this, we created 360 German Subject-Verb-Object sentences that varied in semantic predictability of a sentence-final target word in a graded manner (high, medium, and low) and levels of spectral degradation (1, 4, 6, and 8 channels noise-vocoding). These sentences were presented auditorily to two groups: One group (n =48) performed a listening task in an unpredictable channel context in which the degraded speech levels were randomized, while the other group (n =50) performed the task in a predictable channel context in which the degraded speech levels were blocked. The results showed that at 4 channels noise-vocoding, response accuracy was higher in high-predictability sentences than in the medium-predictability sentences, which in turn was higher than in the low-predictability sentences. This suggests that, in contrast to the narrowed expectations view, comprehension of moderately degraded speech, ranging from low- to high- including medium-predictability sentences, is facilitated in a graded manner; listeners probabilistically preactivate upcoming words from a wide range of semantic space, not limiting only to highly probable sentence endings. Additionally, in both channel contexts, we did not observe learning effects; i.e., response accuracy did not increase over the course of experiment, and response accuracy was higher in the predictable than in the unpredictable channel context. We speculate from these observations that when there is no trial-by-trial variation of the levels of speech degradation, listeners adapt to speech quality at a long timescale; however, when there is a trial-by-trial variation of the high-level semantic feature (e.g., sentence predictability), listeners do not adapt to low-level perceptual property (e.g., speech quality) at a short timescale.

The Impact of Neurocognitive Skills on Recognition of Spectrally Degraded Sentences

Background Recent models theorize that neurocognitive resources are deployed differently during speech recognition depending on task demands, such as the severity of degradation of the signal or modality (auditory vs. audiovisual [AV]). This concept is particularly relevant to the adult cochlear implant (CI) population, considering the large amount of variability among CI users in their spectro-temporal processing abilities. However, disentangling the effects of individual differences in spectro-temporal processing and neurocognitive skills on speech recognition in clinical populations of adult CI users is challenging. Thus, this study investigated the relationship between neurocognitive functions and recognition of spectrally degraded speech in a group of young adult normal-hearing (NH) listeners. Purpose The aim of this study was to manipulate the degree of spectral degradation and modality of speech presented to young adult NH listeners to determine whether deployment of neurocognitive skills would be affected. Research Design Correlational study design. Study Sample Twenty-one NH college students. Data Collection and Analysis Participants listened to sentences in three spectral-degradation conditions: no degradation (clear sentences); moderate degradation (8-channel noise-vocoded); and high degradation (4-channel noise-vocoded). Thirty sentences were presented in an auditory-only (A-only) modality and an AV fashion. Visual assessments from The National Institute of Health Toolbox Cognitive Battery were completed to evaluate working memory, inhibition-concentration, cognitive flexibility, and processing speed. Analyses of variance compared speech recognition performance among spectral degradation condition and modality. Bivariate correlation analyses were performed among speech recognition performance and the neurocognitive skills in the various test conditions. Results Main effects on sentence recognition were found for degree of degradation (p = < 0.001) and modality (p = < 0.001). Inhibition-concentration skills moderately correlated (r = 0.45, p = 0.02) with recognition scores for sentences that were moderately degraded in the A-only condition. No correlations were found among neurocognitive scores and AV speech recognition scores. Conclusions Inhibition-concentration skills are deployed differentially during sentence recognition, depending on the level of signal degradation. Additional studies will be required to study these relations in actual clinical populations such as adult CI users.

Lexical Access in the Face of Degraded Speech: Adapting to Moment by Moment Uncertainty

Listeners often process speech in adverse conditions. One challenge is spectral degradation, where information is missing from the signal. Lexical competition dynamics change when processing degraded speech, but it is unclear why and how these changes occur. We ask if these changes are driven solely by the quality of the input from the auditory periphery, or if these changes are modulated by cognitive mechanisms. Across two experiments, we used the visual world paradigm to investigate changes in lexical processing. Listeners heard different levels of noise-vocoded speech (4- or 15-channel vocoding) and matched the auditory input to pictures of a target word and its phonological competitors. In Experiment 1 levels of vocoding were either blocked together consistently or randomly interleaved from trial-to-trial. Listeners in the blocked condition showed more differentiation between the two levels of vocoding; this suggests that some level of learning is in effect to adapt to the varying levels of uncertainty in the input. Exploratory analyses suggested that when less intelligible speech is processed there is a cost to switching processing modes. In Experiment 2 levels of vocoding were always randomly interleaved. A visual cue was added to inform listeners of the level of difficulty of the upcoming speech. This was enough to attenuate the effects of interleaving as well as the switch cost. These experiments support a role for central processing in dealing with degraded speech. Listeners may be actively forming expectations about the level of degradation they will encounter and altering the dynamics of lexical access.

Distinct sensitivity to spectrotemporal modulation supports brain asymmetry for speech and melody

Does brain asymmetry for speech and music emerge from acoustical cues or from domain-specific neural networks? We selectively filtered temporal or spectral modulations in sung speech stimuli for which verbal and melodic content was crossed and balanced. Perception of speech decreased only with degradation of temporal information, whereas perception of melodies decreased only with spectral degradation. Functional magnetic resonance imaging data showed that the neural decoding of speech and melodies depends on activity patterns in left and right auditory regions, respectively. This asymmetry is supported by specific sensitivity to spectrotemporal modulation rates within each region. Finally, the effects of degradation on perception were paralleled by their effects on neural classification. Our results suggest a match between acoustical properties of communicative signals and neural specializations adapted to that purpose.