Atypical Speech Production Research Articles

Atypical speech production of multisyllabic words and phrases by children with developmental dyslexia.

The prevalent "core phonological deficit" model of dyslexia proposes that the reading and spelling difficulties characterizing affected children stem from prior developmental difficulties in processing speech sound structure, for example, perceiving and identifying syllable stress patterns, syllables, rhymes and phonemes. Yet spoken word production appears normal. This suggests an unexpected disconnect between speech input and speech output processes. Here we investigated the output side of this disconnect from a speech rhythm perspective by measuring the speech amplitude envelope (AE) of multisyllabic spoken phrases. The speech AE contains crucial information regarding stress patterns, speech rate, tonal contrasts and intonational information. We created a novel computerized speech copying task in which participants copied aloud familiar spoken targets like "Aladdin." Seventy-five children with and without dyslexia were tested, some of whom were also receiving an oral intervention designed to enhance multi-syllabic processing. Similarity of the child's productions to the target AE was computed using correlation and mutual information metrics. Similarity of pitch contour, another acoustic cue to speech rhythm, was used for control analyses. Children with dyslexia were significantly worse at producing the multi-syllabic targets as indexed by both similarity metrics for computing the AE. However, children with dyslexia were not different from control children in producing pitch contours. Accordingly, the spoken production of multisyllabic phrases by children with dyslexia is atypical regarding the AE. Children with dyslexia may not appear to listeners to exhibit speech production difficulties because their pitch contours are intact. RESEARCH HIGHLIGHTS: Speech production of syllable stress patterns is atypical in children with dyslexia. Children with dyslexia are significantly worse at producing the amplitude envelope of multi-syllabic targets compared to both age-matched and reading-level-matched control children. No group differences were found for pitch contour production between children with dyslexia and age-matched control children. It may be difficult to detect speech output problems in dyslexia as pitch contours are relatively accurate.

Speech production and speech perception in children with speech sound disorder

ABSTRACT The goal of the present study was to investigate the relationship between speech production (SP) and speech perception in children with Speech Sound Disorder (SSD) from both typical speech production (TSP) and child’s own atypical speech production (ASP). Ten children with SSD were evaluated by a speech production task (naming test) and two speech perception tasks (phonological contrast identification) using the PERCEFAL instrument. In the identification task, the acoustic stimulus, from typical and child’s own atypical speech productions, was presented to each child, and they needed to choose the corresponding stimulus between two pictures displayed on the computer screen. The percentages of errors on speech production (SP) task, on the identification task from TSP and ASP were calculated for each child. ANOVA showed a significant difference between speech production and speech perception. Post hoc analysis demonstrated a greater mean of errors in the ASP compared to the mean of errors in the TSP and SP performances. Only SP and ASP performances presented a significant correlation (r = 0.65). Errors involving manner and manner+place were most recurring in SP. TSP seems to precede SP; that is, children with SSD perceive more speech from typical production than they are able to produce. The significant correlation between ASP and SP suggests that the evaluation of these skills accesses the children’s equivalent underlying phonological representation. A more accurate production, such as typical production, may have cues that help children with SSD to identify phonological contrasts, even if their underlying representations are overly broad or not established.

Perceptive-auditive and perceptive-visual judgment in the identification of gradient productions in fricatives.

To analyze which method of judgment, auditory- perceptual (PAJ) of audios or perceptual-visual judgment (PVJ) (ultrasound images), is more sensitive to detect gradual productions between the class of deaf coronal fricatives and check if there is a correlation between these forms of judgment. Audio and video files of language ultrasound (LUS) related to the production of the words "frog" and "key", of 11 children, between 6 and 12 years old, with atypical speech production, were selected from a bank data and edited for judgments. After instruction and prior training, 20 judges should choose, immediately upon presentation of the stimulus (auditory or visual), one of three options arranged on the computer screen. In PAJ the options were: correct, incorrect or gradient production, while in PVJ the options were images corresponding to the production of [s], [∫] or undifferentiated. The presentation time of the stimuli and the reaction time were automatically controlled by the PERCEVAL software. PVJ provided a higher percentage of identification of gradient stimuli and a shorter reaction time in performing the task compared to PAJ, both statistically significant. Spearman's correlation test did not show statistical significance between PAJ and PVJ responses, nor for reaction time. PVJ using US images proved to be the most sensitive method for detecting gradient production in the production of fricatives [s] and [∫], and can be used as a complementary method to PAJ in speech analysis.

Analysis of acoustic cue production for atypical speech diagnosis

Acoustic cues to lexical distinctive features can be found by examining speech waveform and spectrogram measurements, and can provide a more detailed analysis of speech than is currently provided by methods for identifying atypical speech production. Applying this individual-feature-cue framework to the clinical diagnosis domain, this project analyzes the acoustic cue patterns of young patients with various diagnoses, such as autism spectrum disorder, dyslexia, and Specific Language Impairment. It is hypothesized that each group will produce distinctive acoustic cue patterns for a widely used non-word repetition task, C-TOPP. The speech recordings of 37 subjects were labeled with their acoustic cues; these were then aligned with the cues expected in a canonical production, and acoustic cue level discrepancies were identified. With mispronunciation patterns of each atypical group identified, a Support Vector Machine (SVM) model was developed that can computationally help with diagnosing a patient, given his/her speech recording for this task. Although there are limitations to the study's diagnosis model due to the small data set, the preliminary results show promising insights, yielding 75.7% prediction accuracy for the diagnostic group.

Design and Development of a Speech Intelligibility Test Based on Pseudowords in French: Why and How?

Purpose The current intelligibility tests performed on speakers with atypical speech production are limited by the ability of listeners to restore distorted sequences. This results in a measure that is overvalued when compared with the real articulatory performance. In this article, we present a new intelligibility test in order to neutralize the commonly encountered bias in traditional perception-based assessments. We present the construction of the acoustic-phonetic decoding task and its first test during a perceptual judgment test of intelligibility and during a result comparison with a global perceptual evaluation. Method We developed a very large pseudoword directory including about 90,000 forms that respect French phonotactic constraints. From this directory, we have created lists of pseudowords intended to be recorded for the constitution of the corpus. These lists are established due to an algorithm integrating predefined linguistic constraints and produced by 47 speakers (nine healthy and 38 patients). We then performed a perceptual judgment of intelligibility test with 20 listeners who transcribed these productions. Results At the end of the data processing stage, we obtained a Perceived Phonological Deviation (PPD) score for each speaker that reflects the average number of features altered per phoneme. We then compared the PPD score with a global intelligibility score derived from a global perceptual assessment of intelligibility and of the alteration severity. Conclusions The current findings confirm that a speech intelligibility test based on pseudowords in French achieves fine-grained PPD scores, which enables discrimination between patients and healthy speakers. Moreover, the PPD score is related to the global intelligibility score, especially in severity. Further studies are needed to better understand the discrimination power of this intelligibility test based on an acoustic-phonetic decoding task.

Differentiating Typical From Atypical Speech Production in 5-Year-Old Children With Cerebral Palsy: A Comparative Analysis.

Objective Early diagnosis of speech disorders in children with cerebral palsy (CP) is of critical importance. A key problem is differentiating those with borderline or mild speech motor deficits from those who are within an age-appropriate range of variability. We sought to quantify how well functional speech measures differentiated typically developing (TD) children from children with CP. Method We studied speech production in 45 children with CP (26 with clinical speech motor impairment [SMI] and 19 with no evidence of speech motor impairment [NSMI]) and in 29 TD children of the same age. Speech elicitation tasks were used. Intelligibility, speech rate, and intelligible words per minute were examined. Results All measures differentiated between all 3 groups of children with considerable precision based on area under the receiver operating characteristic curve (AUC) data. AUC was highest for overall intelligibility, which ranged from .88 to .99. Intelligible words per minute also yielded very strong AUCs, ranging from .81 to .99. In each of the receiver operating characteristic models, discrimination between groups was highest for children with speech motor impairment versus TD children. Data indicated that 90% of TD children had overall intelligibility above 87% at 5 years of age, but that no child was 100% intelligible. Furthermore, 90% children with SMI had intelligibility below 72%. Conclusion Findings suggest that functional speech measures differentiate very clearly between children with and without CP and that even children who do not show evidence of speech motor impairment have functional differences in their speech production ability relative to TD peers.

Automatic Screening to Detect 'At Risk' Child Speech Samples using a Clinical Group Verification framework.

Pediatric speech sound disorders (SSD) encompass a wide range of speech production deficits that can interfere with children's educational growth, social engagement and employment opportunities. Early detection of SSDs can facilitate timely intervention and minimize the potential for life-long adverse effects, but distinguishing between typical and atypical speech production in preschoolers is challenging due to developmental and individual variability in speech acquisition. In this study we apply Gaussian Mixture Models to speech samples from 3- to 6-year-old children, recorded by parents using an iOS app. Speech-language pathologists previously classified the samples as positive ('at risk' speech, warranting a referral for a speech-language evaluation) or negative ('no risk' speech). In a series of exploratory analyses, novel distance measures and group scoring techniques are developed which show good subject-level prediction accuracy. Our results provide evidence that it may be feasible to use Speech Processing and Speaker Verification techniques to model and screen speech samples from children for possible speech sound disorders.

Prosodic traits in speech produced by children with autism spectrum disorders – Perceptual and acoustic measurements

Background Autism spectrum disorder has been associated with atypical voice characteristics and prosody. In the scientific literature, four different aspects of atypical speech production in autism spectrum disorder have been highlighted; voice quality together with the prosodic aspects pitch, duration and intensity. Studies of prosody in autism spectrum disorder have almost exclusively used perceptual methods. Recently, some studies have used acoustic analyses. In these studies, it has been pointed out that the acoustic differences found are not necessarily perceived as atypical by listeners, which is why it is important to let listeners evaluate perceptual correlates to acoustic findings. The aims of this study were to use both perceptual and acoustic analyses to study prosodic production in children with autism spectrum disorder and to examine if voice and speech characteristics could be used as clinical markers for autism spectrum disorder. Method Eleven children within normal range of intelligence diagnosed with autism spectrum disorder and 11 children with typical development participated. Every child was recorded telling a story elicited with the expression, reception and recall of narrative instrument. Excerpts of one minute were extracted from the audio files creating the material underlying the perceptual ratings and in the acoustic analysis. An evaluation procedure, partly based on a standardized voice evaluation procedure developed for clinical practice in Sweden, was designed for the perceptual judgments and analysis. To capture critical prosodic variables, aspects of prosody based on characteristic features of Swedish prosody, prosodic features known to cause Swedish children with language impairment particular problems and current research of prosodic impairments in children with autism, were used as rating variables. The acoustic analysis was based on the four variables fundamental frequency ( fo) average, fo range, fo variation and speech rate, together with the language production-related variable number of words per utterance. Results In the acoustic analysis, no differences were found with regards to fo-related variables or speech rate. However, the children in the autism spectrum disorder-group produced significantly more words per utterance than the typically developing children. The perceptual analysis showed no differences between the groups. Only three children with autism spectrum disorder were correctly identified as such. The narrative ability of these children, according to scores on the narrative assessment profile, was poorer than that of the other eight children. They were also more atypical in fluency and in speech rate. Given the small sample, the results should be interpreted with caution. Conclusions and implications The only difference in prosodic production discovered in the acoustic analysis, namely that children with autism spectrum disorder used more words per utterance than the children in the comparison group, was not detected in the perceptual assessment. This implies that it was not perceived as atypical by expert listeners. The results indicate difficulties in using voice and speech characteristics as markers of autism spectrum disorder in clinical settings. The correct identification of some of the children as having autism spectrum disorder or not also indicates that some children with autism spectrum disorder have a prosodic production sufficiently ‘atypical’ in combination with a limited ability to tell stories to be perceived.

Automated screening for speech disorders using acoustic landmark detection

Most children effortlessly learn how to coordinate movements for normal speech production. About one in twelve preschool-aged children, however, show delays in speech production capability that may put them at risk for academic and behavioral difficulties, if not identified and treated. Automated tools that can distinguish between children with and without speech and language impairments could serve as a useful clinical tool for early identification of speech related disorders in young children. In this study, we consider measures based on detecting sequences of acoustic landmarks characteristic of normal speech production over multisyllabic words and continuous speech samples. Ten normal adults, ten typical children and ten with speech disorders recorded twenty multisyllabic words and thirty-three sentences. Acoustic landmarks patterns within utterances and syllabic clusters are examined to characterize differences in landmark sequences between normal and disordered speech. Shannon’s Entropy and ROC analysis are used to evaluate the landmark patterns as potential diagnostic measures of a-typical speech production. We discuss these results and our future work toward developing a fully automated clinical screening tool.

Atypical delayed auditory feedback effect and Lombard effect on speech production in high-functioning adults with autism spectrum disorder.

Individuals with autism spectrum disorder (ASD) show impaired social interaction and communication, which may be related to their difficulties in speech production. To investigate the mechanisms of atypical speech production in this population, we examined feedback control by delaying the auditory feedback of their own speech, which degraded speech fluency. We also examined feedforward control by adding loud pink noise to the auditory feedback, which led to increased vocal effort in producing speech. The results of Japanese speakers show that, compared with neurotypical (NT) individuals, high-functioning adults with ASD (including Asperger’s disorder, autistic disorder, and pervasive developmental disorder not otherwise specified) were more affected by delayed auditory feedback but less affected by external noise. These findings indicate that, in contrast to NT individuals, those with ASD relied more on feedback control than on feedforward control in speech production, which is consistent with the hypothesis that this population exhibits attenuated Bayesian priors.

Constraints on the Transfer of Perceptual Learning in Accented Speech

The perception of speech sounds can be re-tuned through a mechanism of lexically driven perceptual learning after exposure to instances of atypical speech production. This study asked whether this re-tuning is sensitive to the position of the atypical sound within the word. We investigated perceptual learning using English voiced stop consonants, which are commonly devoiced in word-final position by Dutch learners of English. After exposure to a Dutch learner’s productions of devoiced stops in word-final position (but not in any other positions), British English (BE) listeners showed evidence of perceptual learning in a subsequent cross-modal priming task, where auditory primes with devoiced final stops (e.g., “seed”, pronounced [si:th]), facilitated recognition of visual targets with voiced final stops (e.g., SEED). In Experiment 1, this learning effect generalized to test pairs where the critical contrast was in word-initial position, e.g., auditory primes such as “town” facilitated recognition of visual targets like DOWN. Control listeners, who had not heard any stops by the speaker during exposure, showed no learning effects. The generalization to word-initial position did not occur when participants had also heard correctly voiced, word-initial stops during exposure (Experiment 2), and when the speaker was a native BE speaker who mimicked the word-final devoicing (Experiment 3). The readiness of the perceptual system to generalize a previously learned adjustment to other positions within the word thus appears to be modulated by distributional properties of the speech input, as well as by the perceived sociophonetic characteristics of the speaker. The results suggest that the transfer of pre-lexical perceptual adjustments that occur through lexically driven learning can be affected by a combination of acoustic, phonological, and sociophonetic factors.

Instrumental and perceptual phonetic analyses: The case for two-tier transcriptions

In this article, we discuss the relationship between instrumental and perceptual phonetic analyses. Using data drawn from typical and atypical speech production, we argue that the use of two-tier transcriptions, which can compare and contrast perceptual and instrumental information, is valuable both for our general understanding of the mechanisms of speech production and perception and also for assessment and intervention for individuals with atypical speech production. The central tenet of our case is that instrumental and perceptual analyses are not in competition to give a single ‘correct’ account of speech data. They take instead perspectives on complementary phonetic domains, which interlock in the speech chain to encompass production, transmission and perception.

Transcribing disordered speech: the segmental and prosodic layers

In this article we explore the speech layers of a transcription toolkit. We begin by discussing issues of transcription theory and practice, including the difference between broad and narrow phonetic transcription and the importance of narrow transcription with disordered speech, and the measurement of transcriber reliability. We also look at extending symbol sets and using instrumental approaches to deal with atypical speech production. Both segmental and prosodic transcription are dealt with in turn and illustrated with sample transcriptions. We note that, while several phenomena can be subsumed under the prosodic layer, less agreement exists on how to transcribe these than with segmental examples, and that little agreement yet exists on how to notate prosodic disorders. We conclude with illustrations from various adult and child cases.

The realization of affricates in a group of individuals with atypical speech production: a perceptual and instrumental study

The realization of affricates in a group of individuals with atypical speech production: a perceptual and instrumental study

The realization of affricates in a group of individuals with atypical speech production: a perceptual and instrumental study

A partir d'une etude menee sur trois locuteurs anglais presentant des realisations d'affriquees post-alveolaires atypiques, l'A. examine les differentes strategies mises en oeuvre lors de ces productions. Il cherche a identifier des patrons de realisations qui n'auraient pas encore ete mis en lumiere

Neuromotor Development and Language Processing in Developmental Dyspraxia: A Follow-Up Case Study

A longitudinal study of a child (MV) with developmental verbal dyspraxia was conducted to determine to what extent language development and motor performance in this clinical diagnosis followed a similar course of maturation. Patient MV was observed for two years from the age of 5 years and 6 months. Initially, this young patient exhibited unintelligible and atypical speech production (multiword utterances without consonants), delay in balance and coordination, and impairments in rhythmic tasks; but she was otherwise developing normally with no intellectual impairment or behavioral disorder. MRI scans showed moderately enlarged ventricles, a thin, incompletely myelinated corpus callosum and intact basal ganglia. Two years later, MV's performance was nearly normal only in comprehension aspects of language. In contrast, production aspects of language and speech and neuromotor development showed very little improvement after two years. These observations first suggest that development of receptive and expressive domains within language may be asynchronous, and that the progression of motor control of language appears to follow a parallel course to neuromotor development.

Characteristics of Atypical Speech currently not included in the Extensions to the IPA

The Extensions to the IPA for the transcription of disordered speech and voice quality (extIPA) were approved at the 1989 IPA Congress in Kiel (see Duckworth et al. 1990, Ball 1991). A transcription system for disordered speech needs to provide sufficient diversity of symbols and diacritics to account for the majority of observed phonetic variants. Creation of standard new symbols and diacritics for every observed variant is ultimately counterproductive. We submit that the following phenomena have been observed frequently enough and have sufficient diagnostic and therapeutic implications to warrant consideration as new extensions to the IPA. We do not imply, however, that these are the only atypical speech production phenomena that may, in the long run, be worthy of inclusion in a revised extIPA, and we look forward to further suggestions from clinicians and phoneticians.