Analysis Of Spontaneous Speech Research Articles

Language Impairment in Alzheimer's Disease-Robust and Explainable Evidence for AD-Related Deterioration of Spontaneous Speech Through Multilingual Machine Learning.

Alzheimer’s disease (AD) is a pervasive neurodegenerative disease that affects millions worldwide and is most prominently associated with broad cognitive decline, including language impairment. Picture description tasks are routinely used to monitor language impairment in AD. Due to the high amount of manual resources needed for an in-depth analysis of thereby-produced spontaneous speech, advanced natural language processing (NLP) combined with machine learning (ML) represents a promising opportunity. In this applied research field though, NLP and ML methodology do not necessarily ensure robust clinically actionable insights into cognitive language impairment in AD and additional precautions must be taken to ensure clinical-validity and generalizability of results. In this study, we add generalizability through multilingual feature statistics to computational approaches for the detection of language impairment in AD. We include 154 participants (78 healthy subjects, 76 patients with AD) from two different languages (106 English speaking and 47 French speaking). Each participant completed a picture description task, in addition to a battery of neuropsychological tests. Each response was recorded and manually transcribed. From this, task-specific, semantic, syntactic and paralinguistic features are extracted using NLP resources. Using inferential statistics, we determined language features, excluding task specific features, that are significant in both languages and therefore represent “generalizable” signs for cognitive language impairment in AD. In a second step, we evaluated all features as well as the generalizable ones for English, French and both languages in a binary discrimination ML scenario (AD vs. healthy) using a variety of classifiers. The generalizable language feature set outperforms the all language feature set in English, French and the multilingual scenarios. Semantic features are the most generalizable while paralinguistic features show no overlap between languages. The multilingual model shows an equal distribution of error in both English and French. By leveraging multilingual statistics combined with a theory-driven approach, we identify AD-related language impairment that generalizes beyond a single corpus or language to model language impairment as a clinically-relevant cognitive symptom. We find a primary impairment in semantics in addition to mild syntactic impairment, possibly confounded by additional impaired cognitive functions.

Социолингвистическое варьирование параметров устного спонтанного текста

The article examines the social factors influence on prosodic manifestation of syntactic and macrosyntactic units. The data collected during suprasegmental analysis of spontaneous speech are used, which are directly related to the syntactic level, namely: the average length of a phrase in words and in syntagmas and the average length of a syntagma in phonetic words, as well as parameters associated with the design of the whole text: the number of composition blocks in the text, their length in phrases, the length of the whole text in phrases and words. The results of statistical modeling of the influence of the factors "age", "type of education", "level of education" and "gender" clearly indicate that the factors "type of education" and "gender" significantly influence the variation of syntactic and macrosyntactic parameters of an oral text. The factor "type of education" significantly affects the number of phrases in the text and the length of the text block in phrases – both parameters are significantly higher for the speakers who specialize in Humanities compared to those who specialize in Sciences. The length of the syntagma depends on the speaker’ gender – syntagmas produced by male speakers are longer.

On adverbial clauses in spoken Lithuanian

The aim of the paper is to investigate adverbial clauses of time, cause, condition and concession in spontaneous private communication. The study explores semantic relations between the main and subordinate clauses, grammatical features and predominant conjunctions.The data for the research was collected from the morphologically annotated Corpus of Spoken Lithuanian, namely, its sub-corpus of spontaneous private speech which is used at home, at friends’ place, or which is produced by close friends.The analysis of spontaneous private communication shows that the finite adverbial clauses of time, cause, condition and concession are related to a set of conjunctions, but other indicators such as the use of verbal categories (especially tense, aspect and mood), contextual lexical markers as well as pragmatic inference also help to determine the semantic relationship between the main and the subordinate clause.In a spoken language, temporal clauses are usually combined with the conjunctions kai, kaip ‘when’, kol ‘while’, less frequently – with kada ‘when’; causal clauses are combined with the conjunction nes ‘because; since’, less frequently – with kad and kadangi ‘because’; conditional clauses are typically combined with the conjunction jeigu ‘if’, less frequently – with jei ‘if’, concessive clauses – with the conjunction nors ‘though’. The conjunctions kai ‘when’, kol ‘while’, kadangi ‘because’, jeigu and jei ‘if’ correlate with the particle tai that is very frequent in a spoken language, while the conjunction nors ‘though’ – with the contrastive conjunction bet ‘but’.In the natural language flow, the structure of adverbial sentences is modified: other sentential and discourse units can intervene between the main and the subordinate clauses, and the adverbial conjunction moves from the initial to the medial position.Traditional Lithuanian grammars emphasise that the position of adverbial clauses is undefined: they can appear before or after the main clause. However, the analysis of spontaneous speech shows that the position of a subordinate clause is influenced by the semantic relationship between the clauses. If a subordinate clause refers to a previous action or event, then it dominates in a preposition. Besides, the position of an adverbial clause is also influenced by correlative conjunctions: the main clause with the correlative particle tai dominates in the postposition.The research also revealed that Lithuanian adverbial clauses could function at the discourse level: in dialogues, the structure of a complex sentence is broken down and subordinate adverbial clauses can acquire additional – discourse – functions. Adverbial conjunctions, in their turn, can indicate relations with a previous discourse.

Automated assessment of speech production and prediction of MCI in older adults

The population of older adults is growing dramatically and, with it comes increased prevalence of neurological disorders, including Alzheimer’s disease (AD). Though existing cognitive screening tests can aid early detection of cognitive decline, these methods are limited in their sensitivity and require trained administrators. The current study sought to determine whether it is possible to identify persons with mild cognitive impairment (MCI) using automated analysis of spontaneous speech. Participants completed a brief neuropsychological test battery and a spontaneous speech task. MCI was classified using established research criteria, and lexical-semantic features were calculated from spontaneous speech. Logistic regression analyses compared the predictive ability of a commonly-used cognitive screening instrument (the Modified Mini Mental Status Exam, 3MS) and speech indices for MCI classification. Testing against constant-only logistic regression models showed that both the 3MS [χ 2(1) = 6.18, p = .013; AIC = 41.46] and speech indices [χ 2(16) = 32.42, p = .009; AIC = 108.41] were able to predict MCI status. Follow-up testing revealed the full speech model better predicted MCI status than did 3MS (p = .049). In combination, the current findings suggest that spontaneous speech may have value as a potential screening measure for the identification of cognitive deficits, though confirmation is needed in larger, prospective studies.

Factor Analysis of Spontaneous Speech in Aphasia.

Purpose Spontaneous speech tasks are critically important for characterizing spoken language production deficits in aphasia and for assessing the impact of therapy. The utility of such tasks arises from the complex interaction of linguistic demands (word retrieval, sentence formulation, articulation). However, this complexity also makes spontaneous speech hugely variable and difficult to assess. The current study aimed to simplify the problem by identifying latent factors underlying performance in spontaneous speech in aphasia. The ecological validity of the factors was examined by examining how well the factor structures corresponded to traditionally defined aphasia subtypes. Method A factor analysis was conducted on 17 microlinguistic measures of narratives from 274 individuals with aphasia in AphasiaBank. The resulting factor scores were compared across aphasia subtypes. Supervised (linear discriminant analysis) and unsupervised (latent profile analysis) classification techniques were then conducted on the factor scores and the solutions compared to traditional aphasia subtypes. Results Six factors were identified. Two reflected aspects of fluency, one at the phrase level (Phrase Building) and one at the narrative level (Narrative Productivity). Two other factors reflected the accuracy of productions, one at the word level (Semantic Anomaly) and one at the utterance level (Grammatical Error). The other two factors reflected the complexity of sentence structures (Grammatical Complexity) and the use of repair behaviors (Repair), respectively. Linear discriminant analyses showed that only about two thirds of speakers were classified correctly and that misclassifications were similar to disagreements between clinical diagnoses. The most accurately diagnosed syndromes were the largest groups-Broca's and anomic aphasia. The latent profile analysis also generated profiles similar to Broca's and anomic aphasia but separated some subtypes according to severity. Conclusions The factor solution and the classification analyses reflected broad patterns of spontaneous speech performance in a large and representative sample of individuals with aphasia. However, such data-driven approaches present a simplified picture of aphasia patterns, much as traditional syndrome categories do. To ensure ecological validity, a hybrid approach is recommended, balancing population-level analyses with examination of performance at the level of theoretically specified subgroups or individuals. Supplemental Material https://doi.org/10.23641/asha.13232354.

Phonological aspects in premature children aged 12 to 25 months

Objetivo: analisar os aspectos fonológicos empregados na fala de crianças prematuras. Método: estudo observacional e descritivo da produção de fala espontânea em onze crianças prematuras durante seis meses de avaliações. Foram acompanhadas entre os anos de 2014 a 2017, ambos os sexos, nascidas com idade gestacional entre 24 e 35 semanas, independente do peso, sendo avaliadas entre 12 e 25 meses de idade. Inicialmente, consultou o prontuário do bebê, em seguida procedeu-se a anamnese fonoaudiológica e análise de fala espontânea entre criança e mãe em situação lúdica, seguindo-se os preceitos éticos, com o consentimento e a assinatura dos responsáveis. Foram computados em tabelas os totais das produções verbais e quantificado o preenchimento segmental vocálico e consonantal em sílabas tônica e átonas. Resultados: dentre os participantes, Sujeito 2 apresentou-se defasado em relação às demais crianças quanto ao total de palavras e ao preenchimento consonantal e vocálico em todas as posições silábicas. Sujeitos 1, 8, 10 e 11 destacaram-se nos segmentos consonantais presentes na pré-tônica, tônica e pós-tônica, resultando em palavras mais inteligíveis. Conclusão: a sílaba tônica foi a detentora de maior preenchimento segmental. A distinção entre segmentos consonantal e vocálico favoreceu a produção de novas palavras, refletidas no aumento do vocabulário.

La Stanza nella Teoria della Lingua in Atto

This paper presents a corpus-based analysis of Italian spontaneous speech in the theoretical framework of Language into Act Theory (Cresti 2000; Cresti & Moneglia 2010; Moneglia & Raso 2014). A selection of IPIC database (Panunzi & Gregori 2012) has been studied to describe the syntactic behaviors of specific structures named stanzas. Stanzas are terminated sequences prosodically recognizable in the speech flow as chains of nuclei on the same prosodic level. This structure marks nearly 10% of the spoken communication and it is typical of monologues, stories, and life narrations. The analysis shows an additive and flat syntax mostly composed by chained clauses and juxtapositions, without a hierarchical order: stanzas are usually formed by main clauses one after the other or connected through coordinative conjunctions; subordination is not so common and mostly in the form of causal or relative subordinate clauses.

Computer-Assisted Speech Analysis in Mild Cognitive Impairment and Alzheimer's Disease: A Pilot Study from Shanghai, China.

Language dysfunction is a frequently reported symptom in Alzheimer's disease (AD). However, computer-assisted analysis of spontaneous speech in AD and mild cognitive impairment (MCI) is rarely used to date. To characterize the language impairment in AD and amnestic MCI (aMCI) with computer-based automatic analysis via the "Automatic Speech Recognition (ASR) software for cognitive impairment V1.3". A total of 64 subjects, including 20 AD patients, 20 aMCI patients, and 24 healthy controls were recruited. All subjects underwent neuropsychological tests, and spontaneous speech samples were recorded through the description of the "Cookie-Theft Picture" and then analyzed by the computerized software. Subsequently, we compared the speech parameters between the subjects and the controls. We identified seven spontaneous speech parameters (percentage of silence duration, average duration of phrasal segments, average duration of silence segments, number of speech segments, number of long pauses, ratio of hesitation/speech counts and ratio of short pause/speech counts) demonstrating significant differences between the three groups (p < 0.05). All seven speech parameters significantly correlated with cognitive performance, with average duration of silence segments demonstrating the best correlation to cognitive performance on stepwise multiple linear regression analysis. Computer-assisted automated analysis of speech/silence segments demonstrated the potential to reflect the intrinsic linguistic impairment associated with MCI and AD. It has a promising prospect in the early detection of AD and assessment of disease severity.

성별에 따른 구어 종결어미 ‘-잖아’의 화용 전략과 억양의 기능: 대학생 자유발화 분석을 통하여

성별에 따른 구어 종결어미 ‘-잖아’의 화용 전략과 억양의 기능: 대학생 자유발화 분석을 통하여

Automatic Assessment of Speech Impairment in Cantonese-speaking People with Aphasia.

Aphasia is a common type of acquired language impairment resulting from dysfunction in specific brain regions. Analysis of narrative spontaneous speech, e.g., story-telling, is an essential component of standardized clinical assessment on people with aphasia (PWA). Subjective assessment by trained speech-language pathologists (SLP) have many limitations in efficiency, effectiveness and practicality. This paper describes a fully automated system for speech assessment of Cantonese-speaking PWA. A deep neural network (DNN) based automatic speech recognition (ASR) system is developed for aphasic speech by multi-task training with both in-domain and out-of-domain speech data. Story-level embedding and Siamese network are applied to derive robust text features, which can be used to quantify the difference between aphasic speech and unimpaired one. The proposed text features are combined with conventional acoustic features to cover different aspects of speech and language impairment in PWA. Experimental results show a high correlation between predicted scores and subject assessment scores. The best correlation value achieved with ASR-generated transcription is .827, as compared with .844 achieved with manual transcription. The Siamese network significantly outperforms story-level embedding in generating text features for automatic assessment.

The evolution of creaky voice use in read speech by native-French and native-English speakers in tandem: a pilot study

Creaky voice is a non-modal (non-neutral) speech feature occurring both at the segmental and suprasegmental levels of languages. It also performs paralinguistic, stylistic, sociolinguistic and extra-linguistic functions. It has been reported as prevalent in (especially American) English but it is rarely described in French. During tandem exchanges between native-French speakers and native-English speakers, does voice quality found in each of the 2 languages spoken evolve in the course of the interactions? In this pilot study of read speech by 9 French/English tandem pairs (undergraduate Francophone and Anglophone students paired up), how do the number and duration of creaky occurrences vary according to: the language spoken (French or English), the language status (L1 or L2), the speaker, and the segmental or suprasegmental contexts? Creaky occurrences were manually annotated. Their frequency and duration were calculated for the 9 native French female speakers and the 9 native English speakers (eight women, one man). They were found to be more frequent among the Anglophones (in both languages) than their Francophone counterparts (transfer of creaky voice production from English to French for Anglophones), and more frequent when the Francophones spoke in English rather than French (after 3 months of interaction), with a high degree of inter-speaker variability. Moreover, the relative durations of creaky occurrences were similar in both languages spoken by the Francophones, while for the Anglophones, they were shorter in French than in English. The scope of creaky voice can be a segment, the end of stress groups, or whole intonational units (with longer occurrences). In French, glottalisation may also occur at a hiatus, especially when produced by Anglophones. This pilot study is a first step in the research of how speech interaction affects voice quality. The results obtained from read speech will have to be compared to the analysis of spontaneous speech produced by these same speakers.

The function of face gestures and head movements in spontaneous humorous communication

This paper assesses the results of a multimodal analysis of humorous instances found in a collection of 14 interviews from The Late Show with Stephen Colbert. The interviews have been imported and annotated in ELAN for head movements and face expressions. Only the interviewees’ remarks are studied in order to ensure the analysis of spontaneous speech, avoiding mostly pre-scripted host’s comments and questions. The results demonstrate that there are no gestures consistently and recurrently paired with humour. Therefore, none of the multimodal cues found in the sample can be considered markers of humour. Furthermore, the functions of gestures identified are consistent with the uses reported in previous multimodal studies of humorous and non-humorous communication found in the literature.

Native speech plasticity in the German-English late bilingual Stefanie Graf: A longitudinal study over four decades

The purpose of this study was to expose the trajectory of native speech plasticity in the context of late bilingualism through analysis of spontaneous speech of Stefanie Graf (SG) over four decades.With regard to segmental variables, results showed a significant lowering of F2 in /l/, suggesting darkening of the German lateral under the influence of English as a second language (L2). F2 significantly increased in /i/, indicating a more front pronunciation, as predicted due to English L2 acquisition. There was also a significant decrease in F1 of /l/, as well as of /i/, but a significant increase in the F1 frequency of /a/, suggesting widening of the vertical vowel space, potentially due to increasing age. Regarding prosody, there was a significant decrease in pitch level and narrowing of pitch span over time, as expected with increasing age, and surprisingly an increase in average maximum f0 over time, possibly as a result of English L2 acquisition of prosody.These findings suggest that native speech remains plastic post adolescence throughout adulthood, here proposed to be evidenced through a result of acquisition of L2 counterparts in late bilingualism, and that such changes are intertwined with natural biological speech developments over the lifespan.

The Spanish Verbs Estar (To Be) and Ser (To Be) in Child-Directed Speech

Abstract The verbs ser and estar have been a subject of great debate in the literature, mainly because the adjectives that are combined with each copula are not in complementary distribution. A cognitive linguistics approach proposes that estar allows for a comparison of the entity referred to by the utterance’s subject and that very same entity that goes through a temporal change; on the other hand, ser allows for a comparison among entities of different type (Delbecque, 1997). I provide an analysis of spontaneous child-directed speech from a longitudinal database and find variation sets that may allow children to detect the differences between ser and estar. In child-directed speech, the entities referred to by the subject of a sentence with estar are always entities that undergo a perceptible change within an activity of daily life, while the entities referred to with ser never undergo a change.

Evidence for Incomplete Neutralization in Chilean Spanish

Background/Aims: In Chilean Spanish, syllable- and word-final /s/ are frequently weakened to an [h]-like segment or completely deleted. In word-final position, /s/ serves as the plural morpheme, so its deletion renders a site for potential neutralization with singular items. Chilean scholars have previously described differences in the vowel preceding weakened or deleted /s/ distinguishing it from non-/s/-final words, but this putative incomplete neutralization has not yet been acoustically verified, nor have its conditioning factors been explored. The primary purpose of this study was to assess via phonetic analysis of spontaneous speech whether neutralization of final vowels in singular words and plural words in Chilean Spanish is indeed incomplete, as hypothesized by scholars during the 20th century. Additionally, these vowels were also compared to the vowels of monomorphemic /s/-final words in order to ensure that the attested singular-versus-plural differences were not simply indicative of closed syllable laxing processes. Methods: Vowels were extracted from the spontaneous speech of 20 Chilean Spanish speakers and acoustically analyzed via VoiceSauce. Results: The results revealed that final /a/ vowels of plural words were found to be breathier than singular vowels but less breathy than the final vowels of monomorphemic words, and that plural /o/ was significantly fronted. They also demonstrated increased breathiness on /e/ vowels closed by /s/, regardless of morphological status. Conclusion: These results provide the first account of incomplete neutralization of plural vowel correlates in spontaneous speech in Chilean Spanish, and they offer evidence for closed syllable processes in this particular dialect, in alignment with an exemplar-theoretic approach.

Gender-based Intonation Politeness Strategies: An Analysis of Spontaneous Speech in University Student

Gender-based Intonation Politeness Strategies: An Analysis of Spontaneous Speech in University Student

Automatic quantitative analysis of spontaneous aphasic speech

Spontaneous speech analysis plays an important role in the study and treatment of aphasia, but can be difficult to perform manually due to the time consuming nature of speech transcription and coding. Techniques in automatic speech recognition and assessment can potentially alleviate this problem by allowing clinicians to quickly process large amount of speech data. However, automatic analysis of spontaneous aphasic speech has been relatively under-explored in the engineering literature, partly due to the limited amount of available data and difficulties associated with aphasic speech processing. In this work, we perform one of the first large-scale quantitative analysis of spontaneous aphasic speech based on automatic speech recognition (ASR) output. We describe our acoustic modeling method that sets a new recognition benchmark on AphasiaBank, a large-scale aphasic speech corpus. We propose a set of clinically-relevant quantitative measures that are shown to be highly robust to automatic transcription errors. Finally, we demonstrate that these measures can be used to accurately predict the revised Western Aphasia Battery (WAB-R) Aphasia Quotient (AQ) without the need for manual transcripts. The results and techniques presented in our work will help advance the state-of-the-art in aphasic speech processing and make ASR-based technology for aphasia treatment more feasible in real-world clinical applications.

Language in individuals with left hemisphere tumors: Is spontaneous speech analysis comparable to formal testing?

ABSTRACTBackground: The relationship between spontaneous speech and formal language testing in people with brain tumors (gliomas) has been rarely studied. In clinical practice, formal testing is typically used, while spontaneous speech is less often evaluated quantitatively. However, spontaneous speech is quicker to sample and may be less prone to test/retest effects, making it a potential candidate for assessing language impairments when there is restricted time or when the patient is unable to undertake prolonged testing.Aim: To assess whether quantitative spontaneous speech analysis and formal testing detect comparable language impairments in people with gliomas. Specifically, we addressed (a) whether both measures detected comparable language impairments in our patient sample; and (b) which language levels, assessment times, and spontaneous speech variables were more often impaired in this subject group.Method: Five people with left perisylvian gliomas performed a spontaneous speech task and a formal language assessment. Tests were administered before surgery, within a week after surgery, and seven months after surgery. Performance on spontaneous speech was compared with that of 15 healthy speakers.Results: Language impairments were detected more often with both measures than with either measure independently. Lexical–semantic impairments were more common than phonological and grammatical impairments, and performance was equally impaired across assessment time points. Incomplete sentences and phonological paraphasias were the most common error types.Conclusions: In our sample both spontaneous speech analysis and formal testing detected comparable language impairments. Currently, we suggest that formal testing remains overall the better option, except for cases in which there are restrictions on testing time or the patient is too tired to undergo formal testing. In these cases, spontaneous speech may provide a viable alternative, particularly if automated analysis of spontaneous speech becomes more readily available in the future. These results await replication in a bigger sample and/or other populations.

A Speech Recognition-based Solution for the Automatic Detection of Mild Cognitive Impairment from Spontaneous Speech.

Background: Even today the reliable diagnosis of the prodromal stages of Alzheimer’s disease (AD) remains a great challenge. Our research focuses on the earliest detectable indicators of cognitive de-cline in mild cognitive impairment (MCI). Since the presence of language impairment has been reported even in the mild stage of AD, the aim of this study is to develop a sensitive neuropsychological screening method which is based on the analysis of spontaneous speech production during performing a memory task. In the future, this can form the basis of an Internet-based interactive screening software for the recognition of MCI.Methods: Participants were 38 healthy controls and 48 clinically diagnosed MCI patients. The provoked spontaneous speech by asking the patients to recall the content of 2 short black and white films (one direct, one delayed), and by answering one question. Acoustic parameters (hesitation ratio, speech tempo, length and number of silent and filled pauses, length of utterance) were extracted from the recorded speech sig-nals, first manually (using the Praat software), and then automatically, with an automatic speech recogni-tion (ASR) based tool. First, the extracted parameters were statistically analyzed. Then we applied machine learning algorithms to see whether the MCI and the control group can be discriminated automatically based on the acoustic features.Results: The statistical analysis showed significant differences for most of the acoustic parameters (speech tempo, articulation rate, silent pause, hesitation ratio, length of utterance, pause-per-utterance ratio). The most significant differences between the two groups were found in the speech tempo in the delayed recall task, and in the number of pauses for the question-answering task. The fully automated version of the analysis process – that is, using the ASR-based features in combination with machine learning - was able to separate the two classes with an F1-score of 78.8%.Conclusion: The temporal analysis of spontaneous speech can be exploited in implementing a new, auto-matic detection-based tool for screening MCI for the community.

Modeling consonant-context effects in a large database of spontaneous speech recordings.

Given recent interest in the analysis of naturally produced spontaneous speech, a large database of speech samples from the Canadian Maritimes was collected, processed, and analyzed with the primary aim of examining vowel-inherent spectral change in formant trajectories. Although it takes few resources to collect a large sample of audio recordings, the analysis of spontaneous speech introduces a number of difficulties compared to that of laboratory citation speech: Surrounding consonants may have a large influence on vowel formant frequencies and the distribution of consonant contexts is highly unbalanced. To overcome these problems, a statistical procedure inspired by that of Broad and Clermont [(2014). J. Phon. 47, 47-80] was developed to estimate the magnitude of both onset and coda effects on vowel formant frequencies. Estimates of vowel target formant frequencies and the parameters associated with consonant-context effects were allowed to vary freely across the duration of the vocalic portion of a syllable which facilitated the examination of vowel-inherent spectral change. Thirty-five hours of recorded speech samples from 223 speakers were automatically segmented and formant-frequency values were measured for all stressed vowels in the database. Consonant effects were accounted for to produce context-normalized vowel formant frequencies that varied across time.