Automated speech analysis tools for children's speech production: A systematic literature review.

PURPOSE: A systematic search and review of published studies was conducted on the use of automated speech analysis (ASA) tools for analysing and modifying speech of typically-developing children learning a foreign language and children with speech sound disorders to determine (i) types, attributes, and purposes of ASA tools being used; (ii) accuracy against human judgment; and (iii) performance as therapeutic tools. METHOD: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were applied. Across nine databases, 32 articles published between January 2007 and December 2016 met inclusion criteria: (i) focussed on children's speech; (ii) tools used for speech analysis or modification; and (iii) reporting quantitative data on accuracy. RESULT: Eighteen ASA tools were identified. These met the clinical threshold of 80% agreement with human judgment when used as predictors of intelligibility, impairment severity, or error category. Tool accuracy was typically <80% accuracy for words containing mispronunciations. ASA tools have been used effectively to improve to children's foreign language pronunciation. CONCLUSION: ASA tools show promise for automated analysis and modification of children's speech production within assessment and therapeutic applications. Further work is needed to train automated systems with larger samples of speech to increase accuracy for assessment and therapeutic feedback.

Expressive language skills in Chinese Singaporean preschoolers with nonsyndromic cleft lip and/or palate.

OBJECTIVE: The main objective of the present study was to examine THE EXPRESSIVE LANGUAGE SKILLS and obtain a prevalence estimate of expressive language IMPAIRMENT (not skills) in Chinese Singaporean preschoolers with nonsyndromic cleft lip and/or palate (CLP). METHODS: A group of 43 Chinese Singaporean preschoolers aged 3 to 6 years with a diagnosis of nonsyndromic CLP was assessed using the Singapore English Action Picture Test (SEAPT). The SEAPT is an English Language screening tool standardised on typically developing English-Mandarin Chinese Singaporean preschoolers that assesses expressive vocabulary and grammatical usage. A grammar and/or information score below the 20(th) percentile on the SEAPT is indicative of an expressive language impairment. In addition, the medical records of this cohort were examined retrospectively for documentation of surgical timings, audiological history, articulation and resonance. RESULTS: Based on the results of the SEAPT, 33% of the preschoolers with CLP were identified as having a-possible expressive language impairment. Hence, the likelihood that a child with CLP with normal cognitive functioning will have an expressive language impairment is between 3.9 to 12.7 times more likely than in the general population. There was no statistical significance when comparisons were made between dominant language groups or CLP groups on SEAPT measures of information and grammar content. Significantly more males than females were identified with language difficulties, relative to the sex ratio in the sample. No significance was found for the other participant variables. CONCLUSIONS: The findings of this study suggest that Chinese Singaporean preschoolers with CLP have more difficulty in the expressive use of grammar and vocabulary than their peers of typical development, with significantly more males affected than females. As language performance was not related to hearing, articulation or resonance; these early results suggest that a comprehensive investigation of cognition, literacy and family aggregation of communication disorders is urgently warranted to study other possible aetiologies for language impairment in children with CLP in Singapore.

Influence of continual biofeedback on jaw pursuit-tracking in healthy adults and in adults with apraxia plus aphasia.

Apraxia of speech (AOS) is a disorder of motor programming resulting from damage to premotor or anterior insula cortex. The authors used a pursuit visuomotor tracking task to test whether such a disorder interferes with development of motor programs or with modification of existing programs via integration of feedback. Healthy older adults (n = 15) and adults with AOS plus aphasia and nonverbal apraxia (n = 8) performed a jaw movement task with (a) continuous visual feedback of a target movement pattern and their jaw movement and (b) no feedback. Healthy speakers were more accurate and less variable with feedback, suggesting accurate development of a program and feedback integration. Apraxic individuals' performance accuracy and response to feedback suggested that the neurological damage impairs both development of new programs and efficient integration of feedback.

Age-related changes in motor control during articulator visuomotor tracking.

The present study provides normative data on changes in visuomotor control of the oral-facial system across the lifespan. Control of the lower lip, jaw, and larynx (i.e., fundamental frequency) was examined using a nonspeech visuomotor tracking (VMT) task, where subjects move the articulator of interest to track a moving target on an oscilloscope screen. This task examines articulator motor control during movements that are similar to speech but that do not impose linguistic units or the demands of coordinating multiple structures. Accuracy and within- and between-subject variability in tracking performance were measured by cross correlation, gain ratio, phase shift, and target-tracker amplitude difference. Cross-correlation analyses indicated that performance of children (aged 8;2 to 17;0 [years;months]) and older adults (aged 45;1 to 84;3) is poorer than that of younger adults (aged 17;1 to 45;0). Accuracy of movement amplitude tended to increase during development and decline with aging, whereas age did not appear to influence accuracy of temporal parameters in lip and jaw tracking. In contrast, age tended to influence individual variability in temporal but not amplitude parameters. Differences were noted between articulators. The data complement previous studies that considered accuracy and variability of articulator movement during speech. The VMT method and the data provided may be applied to assessment of impairments in the motor speech system and to differential diagnosis of motor speech versus linguistically based disorders.

Response generalization in apraxia of speech treatments: taking another look.

The purpose of the present paper is to present a critical review and reanalysis of response generalization effects in studies of treatment efficacy in apraxia of speech (AOS). Response generalization takes two forms: generalization of treatment effects to untrained exemplars of trained behaviors and generalization to untrained (i.e., novel) behaviors. In the past, response generalization has not been extensive and typically has been restricted to untrained exemplars of trained behaviors. Reasons for these findings are discussed with reference to recent advances in our understanding of the nature of AOS and to theories of speech motor control and teaming. The discussion focuses on the influence of the theoretical basis used to develop hypotheses and select behaviors to test predictions, the complexity of the treatment task/s, and patient characteristics. Suggestions for future directions in treatment efficacy research are offered. Learner outcomes: (1) An understanding of the nature of AOS, based on recent experimental analyses. (2) An understanding of the efficacy of current treatments for AOS in the area of response generalization. (3) An understanding of how theories of speech motor control might be applied to develop hypotheses for testing and increase effectiveness in treatment studies. (4) An understanding of how theories of speech motor control might guide selection of behaviors to test treatment effects and response generalization.

Treatment and generalization of complex sentence production in agrammatism.

The present study applies single-subject experimental design to examine (a) the acquisition and generalization of complex sentence production in agrammatism using Linguistic Specific Treatment (LST) and (b) the utility of syntactic theory in guiding hypotheses of treatment effects. LST trains construction and production of complex sentence structures. Four sentence types were selected for study: object clefts and object-extracted matrix and embedded questions (which are noncanonical with wh-movement), and embedded actives (which are canonical with no overt movement). All sentences contain overt material in the complementizer phrase (CP) of the syntactic tree. Three of five participants (1, 2, and 3) demonstrated generalization from object cleft treatment to production of matrix questions. Thus, LST was effective in improving their ability to generate less complex sentences with wh-movement. Once production of object clefts and matrix questions was acquired, all 5 participants demonstrated generalization from treatment to improved production of embedded questions and/or embedded actives. This generalization involved improved ability to generate embedded clausal structure to form complex sentences but continuing inability to express overt material in CP. Finally, direct treatment for embedded questions did not result in accurate production of embedded actives or vice versa. There were no trends across participants toward improved production of morphosyntactic behaviors in narrative. Persons 1, 2, and 3 showed generalization to increased informativeness and efficiency of expression and were judged by independent listeners to improve in content, coherence, and fluency of spontaneous production. The remaining two participants showed no change or a decline in performance in narrative language production (4 and 5, respectively). These participants demonstrated more severe Broca's aphasia at pretesting compared to Persons 1, 2, and 3, with greater impairments in auditory comprehension, naming, and reading. Etiology and size of lesion did not appear to account for the different behavioral patterns. This study supports the use of LST, which applies syntactic theory to predict patterns of generalization, as an effective treatment approach.

Agrammatic aphasic subjects' comprehension of subject and object extracted Wh questions.

This study attempted to replicate the unusual wh question comprehension pattern of Hickok and Avrutin's (1996) subjects who showed an expected subject/object extraction asymmetry for which NP questions, but not for who questions. We also examined comprehension of what and which one questions, which are similar to who and which NP questions, respectively, and we examined passivized wh questions in order to test predictions of Grodzinsky's (1995) restrictive theory of trace deletion, the Trace-Based Account (TBA). Results, using both a figurine manipulation task and a picture pointing task, showed that only one of four agrammatic (Broca's) aphasic subjects showed the pattern reported by Hickok and Avrutin and that this pattern extended to comprehension of what and which one questions. One of the subjects showed subject/object asymmetry for all wh questions tested, as would be predicted by the original trace deletion hypothesis (Grodzinsky, 1984), and two subjects showed neither pattern. None of our subjects demonstrated ability to comprehend passivized wh questions as predicted by the TBA. We discuss our findings in terms of the lack of homogeneity of wh question comprehension among individuals with agrammatic aphasia and we explore alternatives to the syntactic explanation for differences between who and which NP question comprehension advanced by Hickok and Avrutin.

The role of syntactic complexity in training wh-movement structures in agrammatic aphasia: optimal order for promoting generalization.

This study examined the postulate that training production of syntactically complex sentences results in generalization to less complex sentences that have processes in common with treated structures. Three agrammatic aphasic patients were trained to produce wh-movement structures, object clefts and/or object extracted who-questions, while generalization between these structures was tested. One NP-movement structure, passive sentences, also was tested for control purposes. Wh-movement occurs from the direct object position to specifier position in the complementizer phrase [SPEC, CP] for both wh-movement structures. In who-questions movement occurs in the matrix sentence, whereas, in object clefts movement occurs within an embedded relative clause, rendering them the most complex. Results showed robust generalization effects from object clefts to matrix who-question for 1 participant (D.L.); however, no generalization was noted from who-questions to object clefts for another (F.P.), and 1 participant (C.H.) showed acquisition of who-questions, but not object clefts, during the baseline condition without direct treatment. As expected, none of the participants showed improved production of passives. These findings supported those derived from our previous studies, indicating that generalization is enhanced not only when target structures are related along dimensions articulated by linguistic theory, but also when the direction of treatment is from more to less complex structures. The present findings also support proposals that projections of higher levels in the syntactic treatment are dependent on successful projection of lower levels. For our participants, training movement within CP in a lower (embedded) clause resulted in their ability to project to CP at higher levels.

Training and generalized production of wh- and NP-movement structures in agrammatic aphasia.

The present research examines production of "complex" sentences, which involve movement of noun phrases (NPs), in 2 agrammatic aphasic subjects. According to linguistic theory (Chomsky, 1991, 1993), such sentences are derived using one of two movement operations, either wh- or NP-movement, subsumed under the general rule "move alpha." In this experiment recovery of both wh- and NP-movement derived sentences was investigated using a treatment research paradigm. Subjects were sequentially trained to produce either wh-movement (i.e., who questions, object clefts) or NP-movement (i.e., passives, subject-raising structures) derived sentences. Throughout training, generalization to untrained sentences relying on both types of movement was tested. The influence of training on aspects of narrative discourse also was examined. Results showed generalization patterns constrained to type of movement. Training wh-movement structures resulted in generalized production of untrained wh-movement structures without influencing production of NP-movement structures. Similarly, training of NP-movement structures resulted in generalization only to other sentence types also relying on NP-movement. Aspects of sentence production in narrative contexts also was improved with treatment. These data indicate that movement to an argument (A) position as in NP-movement is distinct from movement to a non-argument (A-bar) position, required in wh-movement. The site where movement terminates in the s-structure of noncanonical sentences appears to influence sentence production. These findings show that linguistic properties of sentences influence sentence production breakdown and recovery in aphasia.

The ultrastructure of cat fusimotor endings and their relationship to foci of sarcomere convergence in intrafusal fibres.

1. Six muscle spindle poles, five from experiments in which foci of sarcomere convergence had been observed during stimulation of fusimotor axons, were serially sectioned for light and electron microscopy. Every somatic motor terminal was studied in ultrathin sections at several levels.2. In all six poles static gamma axons, or presumed static gamma axons, supplying the static bag(2) fibre and/or chain fibres had no terminations on the dynamic bag(1) fibre. In five poles, the dynamic bag(1) fibre was selectively innervated by dynamic gamma or beta axons save in one case where a dynamic gamma axon also innervated one chain fibre.3. Seventy-seven motor endings were of four distinct ultrastructural types: ;m(a) plates' lay superficially on the surface of static bag(2) or chain fibres; ;m(b) plates' were deeply indented into dynamic bag(1) fibres; in ;m(c) plates', found on chain fibres only, the muscle surface was thrown into projecting fingers between which the axon terminals were embedded; one type ;m(d) plate' was found, fully indented into a long chain fibre. A few plates of intermediate form (m(ab)) were variants of m(a) and m(b) plates.4. The muscle membrane beneath both m(a) and m(b) plates was smooth, or had a few wide, shallow folds; m(c) plates usually had wide, shallow subjunctional folds; numerous deep, narrow folds were characteristic of the m(d) plate. The length of unmyelinated pre-terminal axon or the number of sole plate nuclei were not useful diagnostic features.5. Obvious foci of sarcomere convergence in the capsular sleeve region of dynamic bag(1) and static bag(2) fibres coincided with the location of motor plates. Additional contraction foci were observed in the extracapsular region of dynamic bag(1) fibres where there was no motor innervation; contraction occurs principally in the outer half of these fibres. No foci of contraction or motor plates were observed in the extracapsular region of static bag(2) fibres; contraction in these fibres is typically mid-polar.6. In some poles local contraction of chain fibres centred on the location of m(c) plates. In others, very localized contraction occurred distal to the sites of m(a) plates. Both m(a) and m(c) plates were never found on the same pole of a chain fibre.7. Dynamic gamma or beta axons end in m(b) plates, probably equivalent to p(2) plates. The concept of distinctly different p(1) and p(2) plates on dynamic bag(1) fibres, supplied by dynamic beta and gamma axons, respectively, is not supported by ultrastructural evidence.8. Some static gamma axons end in multiple m(a) plates which correspond with ;trail endings', or in single large m(a) plates, on static bag(2) or chain fibres. The m(c) plates are the terminations of other static gamma, or occasionally dynamic gamma, axons on chain fibres. Static beta axons probably end in m(d) plates on long chain fibres which may correspond with p(1) plates.9. It is proposed that there are two types of static gamma motoneurone, one terminating in m(a) plates and the other in m(c) plates, possibly directed preferentially towards static bag(2) fibres and chain fibres, respectively.

Quantitative study of the non-circularity of myelinated peripheral nerve fibres in the cat.

1. One hind limb of each of four cats was either chronically de-efferentated, or chronically de-afferentated, and perfused with buffered glutaraldehyde fixative. Up to three different muscle nerves were dissected from each limb, post-fixed in osmium tetroxide and embedded in Epon. Ultrathin transverse sections were mounted on Formvar-coated single-hole specimen grids so that all the fibres in each nerve could be examined individually by electron microscopy.2. Non-circularity was expressed as the ratio (ø): [Formula: see text] The degree of non-circularity of all the afferent axons, or all the efferent axons, in each muscle nerve was determined. The proportion of fibres cut through the paranodal region, or through the Schwann cell nucleus, was as expected for group I afferent and for alpha and gamma efferent fibres, but hardly any typical paranodal sections of group II or III afferent fibres were encountered which suggests that their paranodal arrangement differs from that of other groups. In a quantitative comparison of noncircularity in different functional groups, fibres cut through paranodes, Schwann cell nuclei or Schmidt-Lanterman clefts were rejected.3. All the gamma efferent fibres in one nerve were studied in a series of sections cut at 25 mum intervals. The degree of non-circularity was found to be relatively constant along the internode of most fibres when the values at paranodes, Schwann cell nuclei or Schmidt-Lanterman clefts were ignored.4. The value of ø varied widely from 1.0 (circular) to 0.5 or less from fibre to fibre within every functional group. However, the mean value of ø was less for gamma axons (0.68) than for alpha axons (0.78), and less for group III axons (0.79) than for axons in groups I and II (both 0.84). When the results for all the nerves were aggregated, these differences were statistically very highly significant, as was the difference in ø between group I and alpha fibres. If values of ø < 0.5 were rejected, the difference between the mean ø for group III and group II was then of doubtful significance whereas that between alpha and gamma fibres was still very highly significant.5. The external perimeter (S) of a non-circular fibre differs from pi times the diameter of a circle just enclosing the fibre (D). It is shown that S = 0.95 pi D for group I and II fibres, S = 0.90 piD for alpha and group III fibres, and S = 0.85 piD for gamma fibres.6. The myelin period, or interperiod repeat distance, varied from 14.1 to 15.6 nm in different cats, implying radial shrinkage of the myelin sheath from 15 to 23%. The myelin period in a particular cat was the same for several nerves, and the same for fibres in different functional groups.7. The possibility that repetitive firing of axons during fixation contributed to the varying degree of non-circularity is considered but rejected as unlikely.8. It is deduced that about 10% radial shrinkage of the myelin sheath, but little or no osmotic shrinkage of the axon, occurred during fixation and rinsing. Further radial shrinkage of about 8% in all components of the fibre probably occurred as a result of subsequent histological processing. It is concluded that the non-circularity of all axons, and the greater non-circularity of small axons, is unlikely to have been due to histological processing.9. It is concluded that axons are non-circular in vivo. The hypothesis that non-circularity allows axons to accommodate swelling during repetitive activity is discussed. Suggestions are made as to why gamma axons may be more non-circular than alpha or group III axons in an anaesthetized cat immediately prior to fixation, and why alpha axons may be more non-circular than axons in groups I and II.

Demonstration of choline acetyltransferase activity in the caotid body of the cat.

1. The distribution of choline acetyltransferase in the carotid body of the cat has been investigated with the electron microscope to determine sites of enzymic activity. This is of relevance to the possible role of acetylcholine as a transmitter in the carotid body.2. Tissues were fixed for short periods and incubated by the method of Kasa, Mann & Hebb, for the fine structural localization of choline acetyltransferase.3. The enzyme was found in the cytoplasm of the type I cells and seemed to be associated with vesicles. No enzyme was found in the large nerve endings synapsing with the type I cell.4. Whole carotid bodies were assayed for their choline acetyltransferase activity and significant amounts were found.5. It is concluded that acetylcholine may be a transmitter in the carotid body and that it is synthesized in type I cells. A possible mode of initiation of chemoreceptor afferent impulses is suggested.

The fine structural localization of cholinesterases in the carotid body of the cat.

1. The distribution of cholinesterases in the carotid body of the cat has been investigated with the electron microscope to obtain a clearer picture of the localization of the enzyme. This is of relevance to the possible role of acetylcholine as a transmitter in the carotid body.2. Tissues were fixed for short periods and incubated by the method of Karnovsky for the fine localization of cholinesterases.3. The enzyme was found in two main sites. Butyrylcholinesterase was present on the type II cell membrane while acetylcholinesterase was found in groups of peripheral axons. No intracellular enzyme was found.4. It is concluded that acetylcholine may be a transmitter in the carotid body and that it may be partly destroyed by enzymic hydrolysis. A possible explanation for some of the anomalous pharmacological findings is suggested.