Hearing Impairment and Perceived Clarity of Predictable Speech

Abstract

The precision of stimulus-driven information is less critical for comprehension when accurate knowledge-based predictions of the upcoming stimulus can be generated. A recent study in listeners without hearing impairment (HI) has shown that form-and meaning-based predictability independently and cumulatively enhance perceived clarity of degraded speech. In the present study, we investigated whether formand meaning-based predictability enhanced the perceptual clarity of degraded speech for individuals with moderate to severe sensorineural HI, a group for whom such enhancement may be particularly important. Design: Spoken sentences with high or low semantic coherence were degraded by noise-vocoding and preceded by matching or nonmatching text primes. Matching text primes allowed generation of form-based predictions while semantic coherence allowed generation of meaningbased predictions. Results: The results showed that both form-and meaning-based predictions make degraded speech seem clearer to individuals with HI. The benefit of form-based predictions was seen across levels of speech quality and was greater for individuals with HI in the present study than for individuals without HI in our previous study. However, for individuals with HI, the benefit of meaning-based predictions was only apparent when the speech was slightly degraded. When it was more severely degraded, the benefit of meaning-based predictions was only seen when matching text primes preceded the degraded speech. The benefit in terms of perceptual clarity of meaning-based predictions was positively related to verbal fluency but not working memory performance. Conclusions: Taken together, these results demonstrate that, for individuals with HI, form-based predictability has a robust effect on perceptual clarity that is greater than the effect previously shown for individuals without HI. However, when speech quality is moderately or severely degraded, meaningbased predictability is contingent on form-based predictability. Further, the ability to mobilize the lexicon seems to contribute to the strength of meaning-based predictions. Whereas individuals without HI may be able to devote explicit working memory capacity for storing meaning-based predictions, individuals with HI may already be using all available explicit capacity to process the degraded speech and thus become reliant on explicit skills such as their verbal fluency to generate useful meaning-based predictions.

The Influence of Form- and Meaning-Based Predictions on Cortical Speech Processing Under Challenging Listening Conditions: A MEG Study

“…Although Nieuwland et al (2018) were unable to replicate the N400 effect demonstrated in the study by DeLong et al (2005), Nieuwland (2019) suggested that pre-activation of form is apparent in earlier brain responses. This hypothesis is in line with previous results showing that the perceived clarity of speech was greater when contingent on form-based predictions rather than meaning-based predictions, especially under adverse listening conditions (see Signoret et al, 2018;Signoret and Rudner, 2019). The difference in speech processing between form-and meaning-based predictions might then be observed on early neural activity, such as in the Mismatch Negativity (MMN) amplitudes.…”

“…Even though it has been observed in several studies that predictions can be generated at multiple levels (e.g., phonological, semantic) during language comprehension (for a review, see Kuperberg and Jaeger, 2016), it remains unclear whether deviations from expectations at different levels have a different impact on speech processing (for a review, see Nieuwland, 2019). Numerous studies have shown that predictions about the form (i.e., phonology) and the meaning (i.e., semantics) of speech increase both its intelligibility (e.g., Miller et al, 1951;Davis and Johnsrude, 2007;Zekveld et al, 2011Zekveld et al, , 2013 and its perceptual clarity (Wild et al, 2012;Signoret et al, 2018;Signoret and Rudner, 2019). This facilitative effect could explain the enhanced perception of a speech event for which we already have knowledge stored in long-term memory -a phenomenon that leads to improved speech detection at a phonological level (see the "speech detection effect" in Signoret et al, 2011), better speech recognition at a lexical level (see the "word detection effect" in Signoret et al, 2011), and facilitated speech categorization at a semantic level (Daltrozzo et al, 2011;Rönnberg et al, 2019).…”

“…This facilitative effect could explain the enhanced perception of a speech event for which we already have knowledge stored in long-term memory -a phenomenon that leads to improved speech detection at a phonological level (see the "speech detection effect" in Signoret et al, 2011), better speech recognition at a lexical level (see the "word detection effect" in Signoret et al, 2011), and facilitated speech categorization at a semantic level (Daltrozzo et al, 2011;Rönnberg et al, 2019). Additionally, predictions about form and meaning have been shown to have an additive and independent facilitative effect on speech perception in that the meaning can still enhance the perceptual clarity of degraded speech even when total reliance on the form is possible (Signoret et al, 2018;Signoret and Rudner, 2019), suggesting that predictions about the form and the meaning could have different kinds of impact on neural speech processing.…”

“…Numerous studies have shown that predictions about the form (i.e., phonology) and the meaning (i.e., semantics) of speech increase both its intelligibility (e.g., Miller et al, 1951 ; Davis and Johnsrude, 2007 ; Zekveld et al, 2011 , 2013 ) and its perceptual clarity ( Wild et al, 2012 ; Signoret et al, 2018 ; Signoret and Rudner, 2019 ). This facilitative effect could explain the enhanced perception of a speech event for which we already have knowledge stored in long-term memory – a phenomenon that leads to improved speech detection at a phonological level (see the “speech detection effect” in Signoret et al, 2011 ), better speech recognition at a lexical level (see the “word detection effect” in Signoret et al, 2011 ), and facilitated speech categorization at a semantic level ( Daltrozzo et al, 2011 ; Rönnberg et al, 2019 ).…”

Visual Rhyme Judgment in Adults With Mild-to-Severe Hearing Loss

“…To achieve understanding of a spoken message, the incoming speech signal has to be mapped onto knowledge representations stored in long-term memory (Rönnberg et al, 2013, 2019). Phonological representations of the sounds of words, which are dissociable from their lexical representations (Cutler, 2008), are fundamental to this process (Marslen-Wilson, 1987; Luce and Pisoni, 1998), and their importance increases when contextual or lexical information is impoverished (Mattys et al, 2005; Signoret and Rudner, 2019). When conditions for speech perception are optimal (Mattys et al, 2012), the process of matching the speech signal to phonological representations stored in long-term memory is seamless.…”