Purpose: The aim of this study was to decode intended and overt speech from neuromagnetic signals while the participants performed spontaneous overt speech tasks without cues or prompts (stimuli).
Method: Magnetoencephalography (MEG), a noninvasive neuroimaging technique, was used to collect neural signals from seven healthy adult English speakers performing spontaneous, overt speech tasks. The participants randomly spoke the words yes or no at a self-paced rate without cues. Two machine learning models, namely, linear discriminant analysis (LDA) and one-dimensional convolutional neural network (1D CNN), were employed to classify the two words from the recorded MEG signals.
Results: LDA and 1D CNN achieved average decoding accuracies of 79.02% and 90.40%, respectively, in decoding overt speech, significantly surpassing the chance level (50%). The accuracy for decoding intended speech was 67.19% using 1D CNN.
Conclusions: This study showcases the possibility of decoding spontaneous overt and intended speech directly from neural signals in the absence of perceptual interference. We believe that these findings make a steady step toward the future spontaneous speech-based brain-computer interface.
Purpose: We investigated speech and nonspeech auditory processing of temporal and spectral cues in people who do and do not stutter. We also asked whether self-reported stuttering severity was predicted by performance on the auditory processing measures.
Method: People who stutter (n = 23) and people who do not stutter (n = 28) completed a series of four auditory processing tasks online. These tasks consisted of speech and nonspeech stimuli differing in spectral or temporal cues. We then used independent-samples t-tests to assess differences in phonetic categorization slopes between groups and linear mixed-effects models to test differences in nonspeech auditory processing between stuttering and nonstuttering groups, and stuttering severity as a function of performance on all auditory processing tasks.
Results: We found statistically significant differences between people who do and do not stutter in phonetic categorization of a continuum differing in a temporal cue and in discrimination of nonspeech stimuli differing in a spectral cue. A significant proportion of variance in self-reported stuttering severity was predicted by performance on the auditory processing measures.
Conclusions: Taken together, these results suggest that people who stutter process both speech and nonspeech auditory information differently than people who do not stutter and may point to subtle differences in auditory processing that could contribute to stuttering. We also note that these patterns could be the consequence of listening to one's own speech, rather than the cause of production differences.
Purpose: Previous studies have suggested that inspirations during speech pauses are influenced by the length of adjacent utterances, owing to respiratory motor planning and physiological recovery processes. The goal of this study was to examine how attention to respiratory sensations may influence these processes in aging speakers with dyspnea, by measuring the effect of sensory monitoring on the relationship between utterance length and the occurrence of inspirations, as well as on functional voice and respiratory measures.
Method: Seventeen adults aged 50 years and older with complaints of voicing-related dyspnea completed a repeated-measures protocol consisting of a 2-week baseline phase and a 4-week sensory monitoring phase. Audiovisual recordings of semistructured speech and self-report questionnaires were collected at study onset, after the baseline phase, and after the sensory monitoring phase. Repeated-measures logistic regressions were conducted to examine changes in the relationship between utterance length and the occurrence of inspirations in adjacent pauses, and repeated-measures analyses of variance were used to investigate any changes in functional voice and respiratory measures.
Results: Planning and recovery processes appeared to remain constant across the baseline phase. From postbaseline to postsensory monitoring timepoints, a strengthening of the relationship between the presence of an inspiration during a speech pause and the length of the subsequent-but not preceding-utterance was noted. Significant improvements were noted in voice-related handicap from study onset to postsensory monitoring, but no changes were reported in respiratory comfort during speech.
Conclusions: Results suggest that respiratory planning processes, that is, the ability to plan breath intakes based on the length of upcoming utterances, may be modifiable behaviorally through targeted sensory monitoring. Further studies are warranted to validate the proposed role of respiratory sensation awareness in achieving skilled temporal coordination between voicing and breathing.
Purpose: The study examines the effect of hearing loss and hearing aid (HA) amplification on the conversational dynamics between hearing-impaired (HI) and normal-hearing (NH) interlocutors. Combining data from the current and a prior study, we explore how the speech levels of both interlocutors correlate and relate to HI interlocutors' degree of hearing loss.
Method: Sixteen pairs of younger NH and elderly HI interlocutors conversed in quiet, with the HI interlocutor either unaided or wearing HAs. We analyzed the effect of hearing status and HA amplification on the conversational dynamics, including turn-taking times (floor-transfer offsets), utterance lengths, and speech levels. Furthermore, we conducted an in-depth analysis of the speech levels using combined data sets from the current and previously published data by Petersen, MacDonald, and Sørensen (2022).
Results: Unaided HI interlocutors were slower and more variable at timing their turns, but wearing HAs reduced the differences between the HI and NH interlocutors. Conversations were less interactive, and pairs were slower at solving the conversational tasks when the HI interlocutor was unaided. Both interlocutors spoke louder when the HI interlocutor was unaided. The speech level of the NH interlocutors was related to that of the HI interlocutors, with the HI speech levels also correlating with their own degree of hearing loss.
Conclusions: Despite typically being unchallenging for HI individuals, one-on-one conversations in quiet were impacted by the HI interlocutor not wearing HAs. Additionally, combining data sets revealed that NH interlocutors adjusted their speech level to match that of HI interlocutors.
Purpose: Response times (RTs) are commonly used in studying language acquisition. However, previous research utilizing RT in the context of language has largely overlooked the intra-individual variability (IIV) of RTs, which could hold significant information about the processes underlying language acquisition.
Method: We explored the association between language abilities and RT variability in visuomotor tasks using two data sets from previously published studies. The participants were 7- to 10-year-old children (n = 77).
Results: Our results suggest that increased variability in RTs is associated with weaker language abilities. Specifically, this within-participant variability in visuomotor RTs, especially the proportion of unusually slow responses, predicted language abilities better than mean RTs, a factor often linked to language skills in past research.
Conclusions: Based on our findings, we introduce the IIV hypothesis in explaining individual differences in language development. According to our hypothesis, inconsistency in the timing of cognitive processes, reflected by increased IIV in RTs, degrades learning different aspects of language, and results in individual differences in language abilities. Future studies should further examine the relationship between IIV and language abilities, and test the extent to which the possible relationship is causal.
Purpose: This study examined the language and nonverbal inhibitory control skills of Italian monolingual and bilingual typically developing (TD) preschoolers with Italian as their second language and of age-matched monolingual and bilingual peers with developmental language disorder (DLD).
Method: Four groups of preschoolers were enrolled: 30 TD Italian monolinguals, 24 TD bilinguals, 19 Italian monolinguals with DLD, and 19 bilinguals with DLD. All children were assessed in Italian on vocabulary, receptive morphosyntax, and morphological markers for DLD in the Italian language (i.e., third-person verb inflections, definite articles, third-person direct-object clitic pronouns, simple prepositions) and nonverbal inhibitory control skills. Group performance was compared using a series of one-way analyses of variance.
Results: Monolingual and bilingual children with DLD achieved significantly lower performance in all language measures compared to both TD monolingual and bilingual children. However, TD bilinguals, although comprehensively showing better language skills than monolinguals with DLD, achieved a performance closer to that of monolinguals with DLD but significantly higher than that of bilinguals with DLD. Both TD monolinguals and bilinguals showed better results than both DLD groups in inhibitory control tasks, particularly in the interference suppression task.
Conclusions: This study provides a picture of language and inhibitory control characteristics of children with various language profiles and adds to the literature on potential markers of DLD among bilingual children. These results suggest that the assessment of nonlinguistic markers, which are associated with language impairment, could be a useful approach to better specify the diagnosis of DLD and reduce cases of misdiagnosis in the context of bilingualism.