Speech Scores Research Articles

Bimodal neglect: When electrically- and acoustically stimulated ears do not play nice with each other

This study describes speech perception in quiet, before and after implantation, in 54 adult cochlear implant (CI) patients with usable hearing in the ear contralateral to the CI. Patients had preoperative aided CNC word scores of at least 8% as well as stable unaided hearing in the unimplanted ear throughout the analysis period. Word identification scores in quiet were analyzed for the implanted ear only (CI), acoustic hearing ear only (HA), and the bimodal condition (CI + HA) to determine changes in performance over time and to characterize bimodal benefit (difference between bimodal score and best single-ear score). On average, patients experienced 6% bimodal benefit in quiet after implantation. Not surprisingly, speech scores in the implanted ear and in the binaural condition significantly improved post-implantation. However, scores in the HA decreased unexpectedly after implantation in almost one third of the patients despite unchanged PTA. Although simple explanations like hearing aid malfunction or ...

Comparing audiological outcomes between the Bonebridge and bone conduction hearing aid on a hard test band: Our experience in children with atresia and microtia

IntroductionTo compare the audiological results of Bone Conduction Hearing Aid (BCHA) on hard test band and Bonebridge (BB) implant among children with microtia and atresia. MethodsThis is a retrospective review of patients with microtia and atresia who underwent BB implant insertion from September 2014 to February 2017 in Starship Children's Hospital. Preoperative audiological testing using a powered BCHA (Oticon Medical Ponto Pro Power) on a hard test band was used to compare post-operative hearing assessments with the BB. ResultsTen microtia and atresia patients were treated with a BB of whom three were treated bilaterally The children were aged between 5 and 15 and all had moderate to moderately severe conductive hearing loss. For each ear tested and subsequently implanted, BB aided speech scores were equivalent to that obtained by a BCHA. The mean improvement of speech reception threshold level between unaided and BB was statistically significant (p > 0.0001). Subjective questionnaire data indicated that BB implanted patients were performing within the norms of overall listening, both in quiet and in noise. Aided Speech In Noise (SIN) testing values were found to range from 0.8–6.5 for BCHA and 0.2–1.2 for BB and the difference was not statistically significant with a p value of 0.143. ConclusionIn audiologic assessments BB performs comparably to BCHA among children with microtia and atresia.

A method for determining precise electrical hearing thresholds in cochlear implant users

Objective: A psychoacoustic procedure designed for the precise assessment of perceptional threshold (T-level) in cochlear implant (CI) users is presented. The impact of this procedure on speech perception was investigated. Design: Individual T-level measurements were obtained with the proposed procedure and three different speech processor fitting conditions were applied: (1) fitting familiar to the subject, T-levels unchanged, (2) T-level set to thresholds determined with the new procedure, (3) T-level set to thresholds determined with the new procedure, but T-level is decreased by 10 clinical units (CU). The impact of the different fitting conditions was measured by means of categorical loudness scaling (CLS) and speech perception tests in quiet and noise. Study sample: A prospective study at a tertiary referral university hospital. 18 experienced postlingually deafened cochlear implanted adult subjects. Results: Average sound-field thresholds obtained by CLS were lowest in condition using the new procedure yielding a larger dynamic range with significantly higher speech scores in quiet compared to those with a subject’s commonly used programme, and significantly improved in noise even after reducing T-levels by 10 CU. Conclusion: The precise determination of T-levels by means of the proposed procedure improved performance in several speech recognition tasks. Compared to the default behavioural setting, T-level increased on median by 9 CU. Average speech reception threshold in noise for soft speech levels (50 dB sound pressure level) decreased by 1 dB.

The Impact of Activity-Based Oral Expression Course on Speech Self-Efficacy of Students

The objective of the present study was to examine the effect of the activity-based oral expression course on the speech self-efficacy of psychological counseling and guidance students. The study group included 80 freshmen students in the Psychological Counseling and Guidance Department in the Faculty of Education of a university located in western Turkey. Quantitative data were collected with the "Speech Self-Efficacy Scale" developed by Katranci and Melanlioglu (2013). Qualitative data were obtained with the descriptive and content analysis of student views. It was found that the difference between the group averages obtained with the paired group t-test that was conducted on the measurement variable of the mean student speech self-efficacy scores was statistically significant (t (158) = -3.335; p = 0.001 <0.05). The posttest speech self-efficacy scores (x̄ = 96.025) were higher than the pretest speech self-efficacy scores (x̄ = 89.213). Based on the qualitative research the results of the present research demonstrated that students considered the activity-based oral expression course beneficial for all sources of self-efficacy beliefs, including direct experiences, indirect experiences, verbal persuasion and emotional state.

Automatic Multimodal Speech Scoring for Using a New Speech Rehabilitation Technology

Automatic Multimodal Speech Scoring for Using a New Speech Rehabilitation Technology

Musicians do not benefit from differences in fundamental frequency when listening to speech in competing speech backgrounds

Recent studies disagree on whether musicians have an advantage over non-musicians in understanding speech in noise. However, it has been suggested that musicians may be able to use differences in fundamental frequency (F0) to better understand target speech in the presence of interfering talkers. Here we studied a relatively large (N = 60) cohort of young adults, equally divided between non-musicians and highly trained musicians, to test whether the musicians were better able to understand speech either in noise or in a two-talker competing speech masker. The target speech and competing speech were presented with either their natural F0 contours or on a monotone F0, and the F0 difference between the target and masker was systematically varied. As expected, speech intelligibility improved with increasing F0 difference between the target and the two-talker masker for both natural and monotone speech. However, no significant intelligibility advantage was observed for musicians over non-musicians in any condition. Although F0 discrimination was significantly better for musicians than for non-musicians, it was not correlated with speech scores. Overall, the results do not support the hypothesis that musical training leads to improved speech intelligibility in complex speech or noise backgrounds.

A hybrid intelligent system for the prediction of Parkinson's Disease progression using machine learning techniques

Parkinson's Disease (PD) is a progressive degenerative disease of the nervous system that affects movement control. Unified Parkinson's Disease Rating Scale (UPDRS) is the baseline assessment for PD. UPDRS is the most widely used standardized scale to assess parkinsonism. Discovering the relationship between speech signal properties and UPDRS scores is an important task in PD diagnosis. Supervised machine learning techniques have been extensively used in predicting PD through a set of datasets. However, the most methods developed by supervised methods do not support the incremental updates of data. In addition, the standard supervised techniques cannot be used in an incremental situation for disease prediction and therefore they require to recompute all the training data to build the prediction models. In this paper, we take the advantages of an incremental machine learning technique, Incremental support vector machine, to develop a new method for UPDRS prediction. We use Incremental support vector machine to predict Total-UPDRS and Motor-UPDRS. We also use Non-linear iterative partial least squares for data dimensionality reduction and self-organizing map for clustering task. To evaluate the method, we conduct several experiments with a PD dataset and present the results in comparison with the methods developed in the previous research. The prediction accuracies of method measured by MAE for the Total-UPDRS and Motor-UPDRS were obtained respectively MAE=0.4656 and MAE=0.4967. The results of experimental analysis demonstrated that the proposed method is effective in predicting UPDRS. The method has potential to be implemented as an intelligent system for PD prediction in healthcare.

Inner Speech's Relationship With Overt Speech in Poststroke Aphasia.

Relatively preserved inner speech alongside poor overt speech has been documented in some persons with aphasia (PWA), but the relationship of overt speech with inner speech is still largely unclear, as few studies have directly investigated these factors. The present study investigates the relationship of relatively preserved inner speech in aphasia with selected measures of language and cognition. Thirty-eight persons with chronic aphasia (27 men, 11 women; average age 64.53 ± 13.29 years, time since stroke 8-111 months) were classified as having relatively preserved inner and overt speech (n = 21), relatively preserved inner speech with poor overt speech (n = 8), or not classified due to insufficient measurements of inner and/or overt speech (n = 9). Inner speech scores (by group) were correlated with selected measures of language and cognition from the Comprehensive Aphasia Test (Swinburn, Porter, & Al, 2004). The group with poor overt speech showed a significant relationship of inner speech with overt naming (r = .95, p < .01) and with mean length of utterance produced during a written picture description (r = .96, p < .01). Correlations between inner speech and language and cognition factors were not significant for the group with relatively good overt speech. As in previous research, we show that relatively preserved inner speech is found alongside otherwise severe production deficits in PWA. PWA with poor overt speech may rely more on preserved inner speech for overt picture naming (perhaps due to shared resources with verbal working memory) and for written picture description (perhaps due to reliance on inner speech due to perceived task difficulty). Assessments of inner speech may be useful as a standard component of aphasia screening, and therapy focused on improving and using inner speech may prove clinically worthwhile. https://doi.org/10.23641/asha.5303542.

Use of Automated Scoring in Spoken Language Assessments for Test Takers With Speech Impairments

This study is an evaluation of the performance of automated speech scoring for speakers with documented or suspected speech impairments. Given that the use of automated scoring of open‐ended spoken responses is relatively nascent and there is little research to date that includes test takers with disabilities, this small exploratory study focuses on one type of scoring technology, automatic speech scoring (the SpeechRaterSM automated scoring engine); one type of assessment, spontaneous spoken English by nonnative adults (six TOEFL iBT® test speaking items per test taker); and one category of disability, speech impairments. The results show discrepancies between human and SpeechRater scores for speakers with documented speech or hearing impairments who receive accommodations and for speakers whose responses were deferred to the scoring leader by human raters because the responses exhibited signs of a speech impairment. SpeechRater scores for these studied groups tended to be higher than the human scores. Based on a smaller subsample, the word error rate was higher for these groups relative to the control group, suggesting that the automatic speech recognition system contributed to the discrepancies between SpeechRater and human scores.

Abstract: Diagnosis, Treatment, and LONG-TERM Speech Outcomes of Occult Submucous Cleft Palate-Associated Velopharyngeal Insufficiency

INTRODUCTION: Overt submucous cleft palate (SMCP) is characterized by Calnan’s triad of bifid uvula, hard palate notch and zona pellucida. Because occult SMCP (OSMCP) lacks this triad, diagnosis is challenging and treatment (if indicated by the presence of velopharyngeal insufficiency [VPI]) is delayed. This study discusses diagnosis, treatment, and long-term outcomes of OSMCP. METHODS: A retrospective chart review was performed for all patients presenting to the Cleft-Craniofacial Center of Children’s Hospital of Pittsburgh who were surgically treated for SMCP between September 2004 to September 2015. Patients with any one of Calnan’s triad were excluded from analysis. The primary outcome was the requirement of secondary speech surgery due to persistent VPI. To evaluate long-term speech outcomes, postoperative follow-up was required to be >3 years for study inclusion. Speech was evaluated over time using Pittsburgh Weighted Speech Scores (PWSS). RESULTS: 317 SMCP patients underwent surgical treatment for VPI, 41 patients were diagnosed with OSMCP (mean age at surgery 6.7 years). Furlow palatoplasty was performed for all patients (mean follow-up: 5.5 years). Diagnostic techniques for OSMCP: 1) intraoral physical examination demonstrating vaulted V-shaped palatal elevation with gag or phonation in 87.8% (36/41) of patients; 2) MRI demonstrating sagittally oriented levator palatini muscles in 19.5% (8/41) of patients; and 3) videofluroscopy to evaluate palatal kinetics in 65.8% (27/41) of patients. Secondary speech surgery was performed on 39.0% (16/41) of patients. 55.6% of patients (5/9) with VCFS underwent secondary surgery (p>0.05). Patients undergoing secondary surgery had a significantly lower change in PWSS (postoperative PWSS minus preoperative PWSS) compared to those who did not (4.4 and 10.0; p=0.022). Using videofluroscopy, patients who required secondary speech surgery had lateral wall motion with 53.6% closure compared to 74.3% closure for those who did not (p=0.11). No oronasal fistulae developed. Posterior pharyngeal flap (PPF) was performed in 87.5% of patients as the secondary speech surgery. Two of these patients (14.2%) developed obstructive sleep apnea (OSA) and required PPF takedown. CONCLUSION: Patients with VPI and no overt signs of SMCP often have delayed diagnosis and incorrect clinical management. Accurate diagnosis of OSMCP is imperative to guide appropriate surgical treatment. Furlow palatoplasty is an important procedure for patients with OSMCP-associated VPI. This shows promising outcomes in patients with OSCMP and limits the risk of PPF-related OSA.

Combining human and automated scores for the improved assessment of non-native speech

In this study, we propose an efficient way to combine human and automated scoring to increase the reliability and validity of a system used to assess spoken responses in the context of an international English language assessment. A set of filtering systems are used to automatically identify various classes of spoken responses that are difficult to score with an automated scoring system, for example, due to a high level of noise or imperfections in components of the overall system. Finally, these flagged responses are then routed to and scored by human raters. The vast majority of responses are not flagged by the filtering system and receive scores by the automated scoring system, resulting in a hybrid scoring approach. The overall hybrid speech scoring system presented here is comprised of multiple subprocesses, including the recording of spoken responses, transcription generation based on an automated speech recognizer, linguistic feature generation, filtering of problematic responses, automated score generation, human rater scoring, and final score combination. We evaluate this scoring approach with pilot data from a novel international English proficiency assessment. It achieves a substantial improvement in scoring performance and score validity with a limited amount of human scoring and most responses scored automatically; the correlation between the baseline system (baseline filtering with imputation) with human raters’ scores is 0.72, and using an extended filtering model, the performance improves to 0.82. The improvement can be attributed in part to the extended filtering model itself that identified more classes of non-scorable responses, and in part to the combination of machine and human scores in our hybrid system.

The Effect of Furlow Palatoplasty Timing on Speech Outcomes in Submucous Cleft Palate.

Because some patients with submucous cleft palate (SMCP) are asymptomatic, surgical treatment is conventionally delayed until hypernasal resonance is identified during speech production. We aim to identify whether speech outcomes after repair of a SMCP is influenced by age of repair. We retrospectively studied nonsyndromic children with SMCP. Speech results, before and after any surgical treatment or physical management of the palate were compared using the Pittsburgh Weighted Speech Scoring system. Furlow palatoplasty was performed on 40 nonsyndromic patients with SMCP, and 26 patients were not surgically treated. Total composite speech scores improved significantly among children repaired between 3 and 4 years of age (P = 0.02), but not older than 4 years (P = 0.63). Twelve (86%) of 14 patients repaired who are older than 4 years had borderline or incompetent speech (composite Pittsburgh Weighted Speech Scoring ≥3) compared with 2 (29%) of 7 repaired between 3 and 4 years of age (P = 0.0068), despite worse prerepair scores in the latter group. Resonance improved in children repaired who are older than 4 years, but articulation errors persisted to a greater degree than those treated before 4 years of age (P = 0.01.) CONCLUSIONS: Submucous cleft palate repair before 4 years of age appears associated with lower ultimate rates of borderline or incompetent speech. Speech of patients repaired at or after 4 years of age seems to be characterized by persistent misarticulation. These findings highlight the importance of timely diagnosis and management.

The effects of early auditory-based intervention on adult bilateral cochlear implant outcomes

Objectives: The goal of this exploratory study was to determine the types of improvement that sequentially implanted auditory-verbal and auditory-oral adults with prelingual and childhood hearing loss received in bilateral listening conditions, compared to their best unilateral listening condition.Methods: Five auditory-verbal adults and five auditory-oral adults were recruited for this study. Participants were seated in the center of a 6-loudspeaker array. BKB-SIN sentences were presented from 0° azimuth, while multi-talker babble was presented from various loudspeakers. BKB-SIN scores in bilateral and the best unilateral listening conditions were compared to determine the amount of improvement gained.Results: As a group, the participants had improved speech understanding scores in the bilateral listening condition. Although not statistically significant, the auditory-verbal group tended to have greater speech understanding with greater levels of competing background noise, compared to the auditory-oral participants.Discussion: Bilateral cochlear implantation provides individuals with prelingual and childhood hearing loss with improved speech understanding in noise. A higher emphasis on auditory development during the critical language development years may add to increased speech understanding in adulthood. However, other demographic factors such as age or device characteristics must also be considered.Conclusion: Although both auditory-verbal and auditory-oral approaches emphasize spoken language development, they emphasize auditory development to different degrees. This may affect cochlear implant (CI) outcomes. Further consideration should be made in future auditory research to determine whether these differences contribute to performance outcomes. Additional investigation with a larger participant pool, controlled for effects of age and CI devices and processing strategies, would be necessary to determine whether language learning approaches are associated with different levels of speech understanding performance.

Cochlear implantation in patients with otosclerosis of the otic capsule

ObjectiveTo evaluate outcomes of cochlear implantation of patients with otosclerosis of the otic capsule. Study designA retrospective case series of 6 patients (7 ears). Patients6 patients (7 ears), 5 patients with severe to profound sensorineural hearing loss; 1 patient with mild to profound sensorineural hearing loss, with radiologic evidence of otosclerosis. All patients were adult males, with or without history of stapes surgery. InterventionCochlear implantation of 7 ears. 5 patients with severe to profound sensorineural hearing loss received the Nucleus Contour Advance peri-modiolar electrode array with binaural implantation performed in one patient. One patient with mild to profound sensorineural hearing loss received a Cochlear® Nucleus Hybrid L24 device. MethodsPreoperative temporal bone CT, audiometric and speech perception testing scores were reviewed, confirming presence of otosclerosis of the cochlea as well as cochlear implant candidacy. Speech perception testing included CNC words, HINT sentences and AZ Bio scores to measure hearing outcomes post implantation. ResultsAll recipients of the contour advance device had a significant improvement in hearing at both 3 and 6month follow up.The hybrid device recipient experienced loss of residual hearing in the implanted ear without improvement at 3months and mild improvement at 6months. ConclusionCochlear implantation has proven to be effective in the treatment of patients with sensorineural hearing loss, including those with otosclerosis of the cochlea.Hybrid candidacy in the setting of otosclerosis of the cochlea may require consideration of alternative electrode devices, most likely a peri-modiolar device.

One-year quality of life and functional outcomes of transoral robotic surgery for carcinoma of unknown primary.

The purpose of this study was to assess the quality of life (QOL) outcomes in patients who underwent transoral robotic surgery (TORS) for carcinoma of unknown primary (CUP). Twenty-nine patients with CUP were administered the Head and Neck Cancer Inventory (HNCI) prospectively to evaluate QOL preoperatively and postoperatively. There was statistically and clinically significant deterioration in 2 of 4 domains (eating and social disruption) in the immediate aftermath of TORS. There was statistically and clinically meaningful deterioration in all 4 domains at 3 months relative to baseline, and there remained statistically and clinically meaningful deterioration in 2 of 4 domains at 6 months compared to baseline. These 2 domains remained statistically and clinically worse than baseline at 12 months. Speech and aesthetics domain scores were comparable at baseline and 12 months. Patients reported difficulties with eating and social disruption over the course of 1 year, but problems with speech and appearance abated.

Approaches to Automated Scoring of Speaking for K–12 English Language Proficiency Assessments

AbstractThis report is the fifth in a series concerning English language proficiency (ELP) assessments for English learners (ELs) in kindergarten through 12th grade in the United States. The series, produced by Educational Testing Service (ETS), is intended to provide theory‐ and evidence‐based principles and recommendations for improving next‐generationELPassessment systems, policies, and practices and to stimulate discussion on better serving K–12ELstudents. The first report articulated a high‐level conceptualization of next‐generationELPassessment systems (Hauck, Wolf, &amp; Mislevy, 2016). The second report addressed accessibility issues in the context ofELPassessments forELsandELswith disabilities (Guzman‐Orth, Laitusis, Thurlow, &amp; Christensen, 2016). The third report focused on critical policy and research issues of summativeELPassessments that state use for accountability purposes (Wolf, Guzman‐Orth, &amp; Hauck, 2016). The fourth report dealt with one of the major uses ofELPassessments—the initial identification and classification of ELs (Lopez, Pooler, &amp; Linquanti, 2016). The present report discusses approaches to using automated scoring technology for the evaluation of student‐spoken responses on K–12ELPassessments. As many states have begun to use computer‐basedELPassessments, there is a growing interest in automated scoring of spoken responses to increase the efficiency of scoring. This report delineates major areas to consider in using automated speech scoring for K–12ELPassessments (i.e., assessment construct and task design, scoring and score reporting, and artificial intelligence (AI) model development and test delivery) and makes recommendations for states on how to evaluate these considerations and determine a path forward.

Daily Sound Awareness of CI Users

Daily Sound Awareness of CI Users

Individual variation in the perception of different types of speech degradation

Both environmental noise and talker-related variation (e.g., accented speech) can create adverse listening conditions for speech communication. Individuals recruit additional cognitive, linguistic, or perceptual resources when faced with such challenges, and they vary in their ability to understand degraded speech. However, it is unclear whether listeners employ the same additional resources when encountering different types of challenging listening conditions. In the present study, we compare individuals’ ability on a variety of skills —including vocabulary, selective attention, rhythm perception, and working memory—with transcription accuracy (i.e., intelligibility scores) of speech degraded by the addition of speech-shaped noise or multi-talker babble and/or talker variation (i.e., a non-native speaker). Initial analyses show that intelligibility scores across degradations of the same class (i.e., either environmental or talker-related) significantly correlate, but correlations of intelligibility score...

Subjective evaluation of speech privacy at consulting rooms in hospitals: Relationship between feeling evoked by overhearing speech and word intelligibility score

Many previous studies and related standards on speech privacy have adopted speech intelligibility as a subjective measure under the assumption of a strong correlation between them. Meanwhile, some studies have attempted to directly evaluate speech privacy. However, the methods used in these studies required evaluators to clearly understand what speech privacy is, while the concept of speech privacy is not widely known to ordinary people using public spaces. In the present study, questionnaire-based surveys and a listening test were conducted to clarify the relationship between speech intelligibility and speech privacy using a more suitable method for directly evaluating speech privacy. The respondents and participants were instructed to imagine a specific situation at a consulting room in a hospital because the required speech privacy depends on the situation. The questionnaire before the listening test showed that a feeling of dissatisfaction was the strongest feeling evoked by overhearing speech. Therefore, we used this feeling as a term for evaluating speech privacy performance that most people can easily accept and understand. The threshold of the feeling of dissatisfaction obtained from the listening test was compared with word intelligibility scores from our previous study. The comparison showed that the threshold of the feeling of dissatisfaction corresponds to a word intelligibility score of around 10–20% for participants with a typical sensitivity to personal information leakage through conversation. In addition, the ratio of participants who felt dissatisfaction increased with increasing background noise level under equal-intelligibility conditions.

Performance of Automated Speech Scoring on Different Low- to Medium-Entropy Item Types for Low-Proficiency English Learners

This report presents an overview of the SpeechRaterSM automated scoring engine model building and evaluation process for several item types with a focus on a low-English-proficiency test-taker population. We discuss each stage of speech scoring, including automatic speech recognition, filtering models for nonscorable responses, and scoring model building and evaluation and compare how the performance at each step differs between different item types. We conclude by discussing the effect of item type on automated scoring performance. We also give recommendations about what considerations should be taken into account when developing tests for low-proficiency English speakers to obtain reliable scores from an automatic scoring engine.