Journal of Speech Language and Hearing Research最新文献

Purpose: This study examined how mothers' question-asking behavior relates to their child's syntactic skills. One important aspect of maternal question-asking behavior is the use of complex questions when speaking with children. These questions can differ based on both their purpose and structure. The purpose may be to seek out information, to teach, or to get a simple yes/no response. Questions may even be rhetorical, with no answer intended at all. Structurally, questions can include a wh-word (who, what, when, where, why, and how) or not; however, these wh-questions are important because they elicit utterances from the child and support vocabulary development. Despite wh-questions eliciting a response from children, it remains unknown how these questions relate to children's syntactic skills.

Method: Thirty-four mother-child dyads participated in a 15-min seminaturalistic play session. Children were between the ages of 5 and 7 years (M = 6.26 years, SD = 1.04 years; 20 girls/14 boys). The Diagnostic Evaluation of Language Variation (DELV) assessment was used to measure syntactic skills in children. Using the Systematic Analysis of Language Transcripts, questions were categorized based on structure (wh-questions vs. non-wh-questions) and purpose (information-seeking, pedagogical, or yes/no and rhetorical questions). A repeated-measures analysis of covariance and a linear regression model were implemented to address the frequency of different questions asked by mothers, as well as what types of questions are most related to children's concurrent syntactic skills.

Results: When controlling for total maternal utterances, results revealed that non-wh-questions and rhetorical/yes and no questions were the most frequent types of questions produced by mothers, in terms of structure and purpose, respectively. However, wh-questions were predominantly information-seeking questions. This is important, as the use of information-seeking wh-questions was positively associated with children's syntactic skills, as measured by the DELV, and resulted in children producing longer utterances in response to these questions, as determined by child mean length of utterance in words.

Conclusion: Taken together, these findings suggest maternal use of wh-questions aids syntactic skills in children ages 5-7 years, likely because they require a more syntactically complex response on the child's behalf.

Supplemental material: https://doi.org/10.23641/asha.27276891.

Purpose: Prior research extensively documented challenges in recognizing verbal and nonverbal emotion among older individuals when compared with younger counterparts. However, the nature of these age-related changes remains unclear. The present study investigated how older and younger adults comprehend four basic emotions (i.e., anger, happiness, neutrality, and sadness) conveyed through verbal (semantic) and nonverbal (facial and prosodic) channels.

Method: A total of 73 older adults (43 women, M_age = 70.18 years) and 74 younger adults (37 women, M_age = 22.01 years) partook in a fixed-choice test for recognizing emotions presented visually via facial expressions or auditorily through prosody or semantics.

Results: The results confirmed age-related decline in recognizing emotions across all channels except for identifying happy facial expressions. Furthermore, the two age groups demonstrated both commonalities and disparities in their inclinations toward specific channels. While both groups displayed a shared dominance of visual facial cues over auditory emotional signals, older adults indicated a preference for semantics, whereas younger adults displayed a preference for prosody in auditory emotion perception. Notably, the dominance effects observed in older adults for visual and semantic cues were less pronounced for sadness and anger compared to other emotions. These challenges in emotion recognition and the shifts in channel preferences among older adults were correlated with their general cognitive capabilities.

Conclusion: Together, the findings underscore that age-related obstacles in perceiving emotions and alterations in channel dominance, which vary by emotional category, are significantly intertwined with overall cognitive functioning.

Supplemental material: https://doi.org/10.23641/asha.27307251.

Purpose: The aim of the study was to determine the ability of children with developmental language disorder (DLD) to infer word class and meaning from text and to document variations by word class (noun, verb, adjective) and modality (listening, reading). We also asked whether the children could integrate global cues across the entire passage as well as local cues from the immediate sentence frame to support inferences.

Method: Fourth graders with DLD (n = 28) and typical language development (TLD; n = 41) read and listened to expository texts and guessed the noun, verb, and adjective removed from each. Adults (n = 20) completed the task to establish a baseline of correct responses. We used latent semantic analysis (LSA) to determine the semantic fit of the responses to the texts and to determine whether global cues were more difficult for children with DLD than local cues.

Results: The DLD group was 24% less accurate than the TLD group. In both diagnostic groups, accuracy varied by word class (nouns > adjectives > verbs) but not modality (reading = listening). Word class errors were rare, and errors of semantic fit were frequent. LSA cosines were higher for correct responses relative to the passage as a whole than the immediate sentence frame, suggesting that both groups mined the more extensive information in the global cues to support inferences. Compared to the TLD group, the DLD group tended to make "worse" errors: repeating words from the sentence frame or coming up with no response at all. Accuracy in the DLD group, but not the TLD group, was related to vocabulary knowledge. When the two groups were collapsed, scores on verbal short-term/working memory and sustained attention also predicted performance, but weaknesses in these aspects of executive function on the part of individuals with DLD did not fully explain the difference between the performance of the DLD and TLD groups.

Conclusions: Whether listening or reading, fourth graders with DLD are less able to infer word meaning from texts than their age-mates. The problem reflects, in part, deficits in executive function and lexical semantic knowledge.

Purpose: The purpose of the present study was to use a large swallowing database to explore and compare the swallow-physiology impairment profiles of five dysphagia-associated diagnoses: chronic obstructive pulmonary disease (COPD), dementia, head and neck cancer (HNC), Parkinson's disease (PD), and stroke.

Method: A total of 8,190 patients across five diagnoses were extracted from a de-identified swallowing database, that is, the Modified Barium Swallow Impairment Profile Swallowing Data Registry, for the present exploratory cross-sectional analysis. To identify the impairment profiles of the five diagnoses, we fit 18 partial proportional odds models, one for each of the 17 Modified Barium Swallow Impairment Profile components and the Penetration-Aspiration Scale, with impairment score as the dependent variable and diagnoses, age, sex, and race as the independent variables with interactions between age and diagnoses and between PD and dementia (in effect creating a PD with dementia [PDwDem] group). For components with > 5% missingness, we applied inverse probability weighting to correct for bias.

Results: PD and COPD did not significantly differ on 13 of the 18 outcome variables (all ps > .02). Dementia, stroke, and PDwDem all showed worse impairments than COPD or PD on five of six oral components (all ps < .007). HNC had worse impairment than all diagnoses except PDwDem for nine of 10 pharyngeal components (all ps < .006). Stroke and HNC had worse penetration/aspiration than all other diagnoses (all ps < .003).

Conclusions: The present results show that there are both common and differing impairment profiles among these five diagnoses. These commonalities and differences in profiles provide a basis for the generation of hypotheses about the nature and severity of dysphagia in these populations. These results are also likely highly generalizable given the size and representativeness of the data set.

Supplemental material: https://doi.org/10.23641/asha.27478245.

Purpose: Speech recognition in noise is challenging for listeners and appears to require support from executive functions to focus attention on rapidly unfolding target speech, track misunderstanding, and sustain attention. The current study was designed to test the hypothesis that lower executive function abilities explain poorer speech recognition in noise, including among older participants with hearing loss who often exhibit diminished speech recognition in noise and cognitive abilities.

Method: A cross-sectional sample of 400 younger-to-older adult participants (19 to < 90 years of age) from the community-based Medical University of South CarolinaLongitudinal Cohort Study of Age-related Hearing Loss were administered tasks with executive control demands to assess individual variability in a card-sorting measure of set-shifting/performance monitoring, a dichotic listening measure of selective attention/working memory, sustained attention, and processing speed. Key word recognition in the high- and low-context speech perception-in-noise (SPIN) tests provided measures of speech recognition in noise. The SPIN scores were adjusted for audibility using the Articulation Index to characterize the impact of varied hearing sensitivity unrelated to reduced audibility on cognitive and speech recognition associations.

Results: Set-shifting, dichotic listening, and processing speed each explained unique and significant variance in audibility-adjusted, low-context SPIN scores (ps < .001), including after controlling for age, pure-tone threshold average (PTA), sex, and education level. The dichotic listening and processing speed effect sizes were significantly diminished when controlling for PTA, indicating that participants with poorer hearing sensitivity were also likely to have lower executive function and lower audibility-adjusted speech recognition.

Conclusions: Poor set-shifting/performance monitoring, slow processing speed, and poor selective attention/working memory appeared to partially explain difficulties with speech recognition in noise after accounting for audibility. These results are consistent with the premise that distinct executive functions support speech recognition in noise.

Purpose: Several studies have demonstrated that dynamic assessment can be a less biased, valid approach for the identification of language disorder among diverse school-age children. However, all prior studies have included a relatively small number of participants, which is generally not adequate for psychometric research. This is the first large-scale study to (a) examine whether a dynamic assessment of narrative language yields indifferent outcomes regardless of several demographic variables including age, race/ethnicity, multilingualism, or gender; (b) examine the sensitivity and specificity of the dynamic assessment of language among a large sample of students with and without language disorder; and (c) identify specific cut-points by grade to provide clinically useful data.

Method: Participants included 634 diverse first- through fifth-grade students with and without language learning disorder. Students were confirmed as having a language disorder using a triangulation technique involving several sources of data. A dynamic assessment of narrative language, which took approximately 10 min, was administered to all students.

Results: Results indicated that the dynamic assessment had excellent (> 90%) sensitivity and specificity and that modifiability scores were not meaningfully different across any of the demographic variables.

Conclusions: The dynamic assessment of narrative language accurately identified language disorder across all student demographic groups. These findings suggest that dynamic assessment may provide less biased classification than traditional, static forms of assessment.

Purpose: This study describes the development of verb inflectional morphology in an urban dialect of Palestinian Arabic (PA) spoken in northern Israel, specifically in the city of Haifa, and explores the effect of language typology on acquisition.

Method: We analyzed naturalistic longitudinal speech samples from one monolingual Arabic-speaking girl aged 1;11-2;3 during spontaneous interactions with family members.

Results: Initially, truncated forms ("bare stems") were common but disappeared by the end of the study. By age 1;11, the girl was in the proto-morphological stage, displaying clear three-member mini-paradigms. Affixation complexity gradually increased, with adjacent and obligatory suffixes acquired before distant and optional prefixes. The early acquisition of indicative prefixes (b-, m-) preceded the later emergence of complex proclitics (e.g., volitive d-, progressive ʕam), suggesting gradual, systematic morphological acquisition.

Conclusions: We propose three principles for the development of PA verb inflection: (a) Adjacency: Affixes adjacent to the base are acquired first. (b) R-salience: Suffixes are acquired earlier than prefixes. (c) Obligatoriness: Obligatory morphemes precede optional ones. These principles predict the girl's morphological development and reflect sensitivity to PA's richly inflecting typology. This study highlights the need for detailed descriptive research that is essential for understanding language acquisition processes and informing assessment tools, intervention programs, and educational curricula for PA-speaking children.

Purpose: The purpose of this study was to investigate cortical tracking of speech (CTS) in adults who stutter (AWS) compared to typically fluent adults (TFAs) to test the involvement of the speech-motor network in tracking rhythmic speech information.

Method: Participants' electroencephalogram was recorded while they simply listened to sentences (listening only) or completed them by naming a picture (listening for speaking), thus manipulating the upcoming involvement of speech production. We analyzed speech-brain coherence and brain connectivity during listening.

Results: During the listening-for-speaking task, AWS exhibited reduced CTS in the 3- to 5-Hz range (theta), corresponding to the syllabic rhythm. The effect was localized in the left inferior parietal and right pre/supplementary motor regions. Connectivity analyses revealed that TFAs had stronger information transfer in the theta range in both tasks in fronto-temporo-parietal regions. When considering the whole sample of participants, increased connectivity from the right superior temporal cortex to the left sensorimotor cortex was correlated with faster naming times in the listening-for-speaking task.

Conclusions: Atypical speech-motor functioning in stuttering impacts speech perception, especially in situations requiring articulatory alertness. The involvement of frontal and (pre)motor regions in CTS in TFAs is highlighted. Further investigation is needed into speech perception in individuals with speech-motor deficits, especially when smooth transitioning between listening and speaking is required, such as in real-life conversational settings.

Supplemental material: https://doi.org/10.23641/asha.27234885.

Purpose: A relevant aspect of listening is the effort required during speech processing, which can be assessed by pupillometry. Here, we assessed the pupil dilation response of normal-hearing (NH) and hard of hearing (HH) individuals during listening to clear sentences and masked or degraded sentences. We combined this assessment with functional magnetic resonance imaging (fMRI) to investigate the neural correlates of the pupil dilation response.

Method: Seventeen NH participants (M_age = 46 years) were compared to 17 HH participants (M_age = 45 years) who were individually matched in age and educational level. Participants repeated sentences that were presented clearly, that were distorted, or that were masked. The sentence intelligibility level of masked and distorted sentences was 50% correct. Silent baseline trials were presented as well. Performance measures, pupil dilation responses, and fMRI data were acquired.

Results: HH individuals had overall poorer speech reception than the NH participants, but not for noise-vocoded speech. In addition, an interaction effect was observed with smaller pupil dilation responses in HH than in NH listeners for the degraded speech conditions. Hearing impairment was associated with higher activation across conditions in the left superior temporal gyrus, as compared to the silent baseline. However, the region of interest analysis indicated lower activation during degraded speech relative to clear speech in bilateral frontal regions and the insular cortex, for HH compared to NH listeners. Hearing impairment was also associated with a weaker relation between the pupil response and activation in the right inferior frontal gyrus. Overall, degraded speech evoked higher frontal activation than clear speech.

Conclusion: Brain areas associated with attentional and cognitive-control processes may be increasingly recruited when speech is degraded and are related to the pupil dilation response, but this relationship is weaker in HH listeners.

Supplemental material: https://doi.org/10.23641/asha.27162135.

Purpose: This work introduces updated transcripts, disfluency annotations, and word timings for FluencyBank, which we refer to as FluencyBank Timestamped. This data set will enable the thorough analysis of how speech processing models (such as speech recognition and disfluency detection models) perform when evaluated with typical speech versus speech from people who stutter (PWS).

Method: We update the FluencyBank data set, which includes audio recordings from adults who stutter, to explore the robustness of speech processing models. Our update (semi-automated with manual review) includes new transcripts with timestamps and disfluency labels corresponding to each token in the transcript. Our disfluency labels capture typical disfluencies (filled pauses, repetitions, revisions, and partial words), and we explore how speech model performance compares for Switchboard (typical speech) and FluencyBank Timestamped. We present benchmarks for three speech tasks: intended speech recognition, text-based disfluency detection, and audio-based disfluency detection. For the first task, we evaluate how well Whisper performs for intended speech recognition (i.e., transcribing speech without disfluencies). For the next tasks, we evaluate how well a Bidirectional Embedding Representations from Transformers (BERT) text-based model and a Whisper audio-based model perform for disfluency detection. We select these models, BERT and Whisper, as they have shown high accuracies on a broad range of tasks in their language and audio domains, respectively.

Results: For the transcription task, we calculate an intended speech word error rate (isWER) between the model's output and the speaker's intended speech (i.e., speech without disfluencies). We find isWER is comparable between Switchboard and FluencyBank Timestamped, but that Whisper transcribes filled pauses and partial words at higher rates in the latter data set. Within FluencyBank Timestamped, isWER increases with stuttering severity. For the disfluency detection tasks, we find the models detect filled pauses, revisions, and partial words relatively well in FluencyBank Timestamped, but performance drops substantially for repetitions because the models are unable to generalize to the different types of repetitions (e.g., multiple repetitions and sound repetitions) from PWS. We hope that FluencyBank Timestamped will allow researchers to explore closing performance gaps between typical speech and speech from PWS.

Conclusions: Our analysis shows that there are gaps in speech recognition and disfluency detection performance between typical speech and speech from PWS. We hope that FluencyBank Timestamped will contribute to more advancements in training robust speech processing models.