Speech

The auditory brainstem response reflects neural encoding of the acoustic characteristic of a speech syllable with remarkable precision. Some children with learning impairments demonstrate abnormalities in this preconscious measure of neural encoding especially in background noise.

10. DC cultures were generated as described (5, 22). Immature I-A kϩ DCs grown from C3H/HeN, C3H/HeJ, and CBA/J (Jackson Labs, Bar Harbor, ME) marrow suspensions were pulsed with HEL (1 to 3 mg/ml) with or without LPS (1 to 10 ng/ml, E. coli 0111.B4) for 1 to 3 hours, washed, and chased in HEL-and LPS-free media. Similar results were obtained with DCs from each strain. DC cultures contained supernatant from mGM-CSFexpressing J558L cells (from A. Lanzavecchia, Basel Institute, Basel, Switzerland). LPS was removed from HEL (Sigma) with Kuttsuclean adsorbent (Maruha Corporation, Ibaraki, Japan). Maturation stimuli included LPS, CD40L, tumor necrosis factor-␣, or replating (22). 11. DCs were fixed with 4% (wt/vol) paraformaldehyde and washed in phosphate-buffered saline. Cells were incubated with monoclonal antibodies (mAbs) on ice as described (23). The rat mAb C4H3 was provided by R. Germain (NIH, Bethesda, MD). C4H3 was purified on protein A-Sepharose and biotinylated. Cells were labeled with biotin-conjugated mAbs (Pharmingen) to I-E k (14-4-4S) or I-A k (10-2.16) and fluorescein

By varying the sound-absorption treatments in a simulated classroom, experimental results were compared with analytical and computer predictions of reverberation time. Analytical predictions were made with different absorption exponents, which are the result of different weighting procedures involving room surface areas and the sound-absorption coefficients. Sound scattering was found to influence measured reverberation times. With the amount of sound scattering provided, more accurate analytical predictions were obtained with absorption exponents that give reverberation times longer than those obtained with the Sabine absorption exponent, which consistently underpredicted reverberation times. However, none of the absorption exponents could be singled out as more adequate because of similar average accuracy. Computer predictions of reverberation time were accomplished with two commercially available ray-based programs, RAYNOISE 3.0 and ODEON 2.6, with specular and calibrated diffuse reflection procedures. Neither type of procedure, in either program, was more accurate than the best analytical predictions. With RAYNOISE, neither the specular nor the calibrated diffuse reflection procedure could be singled out as more adequate. For ODEON, the calibrated diffuse reflection procedure gave consistently more accurate predictions than its specular reflection procedure, with the best accuracy of the computer predictions. ͓S0001-4966͑00͒04710-X͔

The visual dimension has featured prominently in person perception in the last 25 years. Alone, this dimension cannot give the complete 'picture' of others because language and speech (i.e., the auditory dimension) are also highly informative. Social-cognition research investigates the role of auditory information for impression formation. In a series of experiments, we tested perceived competence, hirability, and socio-intellectual status of different targets based on their regional accents. Given identical content of statements in different conditions of an alleged job interview, regional German accents (Saxon, Bavarian, and Berlin) resulted in lower perceived competence and hirability than standard German, even though the Bavarian accent at the same time resulted in higher ratings of socio-intellectual status compared to other regional accents (Experiment 1). These findings were confirmed when using a broader population sample and a 'matched guise' technique (Experiment 2). Our findings indicate that regional accents, similar to faces, can be very powerful in creating differentiated pictures of individuals.

In recent years, there are many great successes in using deep architectures for unsupervised feature learning from data, especially for images and speech. In this paper, we introduce recent advanced deep learning models to classify two emotional categories (positive and negative) from EEG data. We train a deep belief network (DBN) with differential entropy features extracted from multichannel EEG as input. A hidden markov model (HMM) is integrated to accurately capture a more reliable emotional stage switching. We also compare the performance of the deep models to KNN, SVM and Graph regularized Extreme Learning Machine (GELM). The average accuracies of DBN-HMM, DBN, GELM, SVM, and KNN in our experiments are 87.62%, 86.91%, 85.67%, 84.08%, and 69.66%, respectively. Our experimental results show that the DBN and DBN-HMM models improve the accuracy of EEGbased emotion classification in comparison with the state-ofthe-art methods.

This paper presents the deep neural networks to classification of children with voice impairments from speech signals. In the analysis of speech signals, 6,373 static acoustic features are extracted from many kinds of low-level-descriptors and functionals. To reduce the variability of extracted features, two-dimensional normalizations are applied to smooth the interspeaker and inter-feature mismatch using the feature warping approach. Then, the feature selection is used to explore the discriminative and low-dimensional representation based on techniques of principal component analysis and linear discriminant analysis. In such representation, the robust features are obtained by eliminating noise features via subspace projection. Finally, the deep neural networks are adopted to classify the children with voice impairments. We conclude that deep neural networks with the proposed feature normalization and selection can significantly contribute to the robustness of recognition in practical application scenarios. We have achieved an UAR of 60.9% for the four-way diagnosis classification on the development set. This is a relative improvement of 16.2% to the official baseline by using our single system. I.

This study investigated changes in oral health-related quality of life (OHRQoL) in partially edentulous patients treated with removable dental prostheses (RDPs), fixed dental prostheses (FDPs), fixed-removable (combined) restorations (COMBs), and implant-supported fixed prostheses (ISFPs). A total of 200 patients (30 to 50 years old) were enrolled: 45 received RDPs, 32 received FDPs, 66 received COMBs, and 57 received ISFPs. OHRQoL was measured using the shortened version of the Oral Health Impact Profile (OHIP-14) before treatment and 6 weeks and 6 months after treatment. Treatment groups were sex-neutral; however, significant differences were found relative to age and Kennedy classification. A general linear model was used to explore the interaction of age and Kennedy classification with treatment modality. Pretreatment analysis revealed that the psychologic discomfort domain showed the greatest negative impact on OHRQoL, while functional limitation had the smallest effect. Within...

We argue for a compositional semantics grounded in a strongly typed ontology that reflects our commonsense view of the world and the way we talk about it. Assuming such a structure we show that the semantics of various natural language phenomena may become nearly trivial.

The anatomy, physiology, acoustics and perception of speech: essential elements in analysis of the evolution of human speech Inferences on the evolution ofhuman speech based on anatomical data muat take into account its physiology, acoustics and perception. Human sperch is generated by the supralaryngeal vocal tract (SVT) acting as an acoustic filter on noise sources generated by turbulent airflow and quasi-periodic phonation generated by the activity of the larynx. The formant frequencies, which arc major determinants of phonetic quality, are the frequencies at which relative energy maxima will pass through the SVT filter. Neither the articulator) gestures of the tongue nor their acoustic consequences can be fractionattd into oral and pharyngeal cavity components. Moreover, the acoustic cues that specify individual consonants and vowels are "encoded", i.e., melded together. Formant frequency encoding makes human speech a vehicle for rapid vocal communication. Non-human primates lack the anatomy that enables modern humans to produce sounds that enhance this process, as well as the neural mechanisms necessary for the voluntary control of speech articulation. The specific claims ofDuchin [ 1990) arr discussed.

Disponível em: http://www.redalyc.org/src/inicio/ArtPdfRed.jsp?iCve=169320515017 ... Redalyc Sistema de Informação Científica Rede de Revistas Científicas da América Latina o Caribe, a Espanha e Portugal ... Nogueira Gregio, Fabiana; Gama-Rossi, Aglael; Madureira, ...

Rationale and objectives: Considering the controversy in the literature regarding several aspects of temporomandibular dysfunction (TMD) in elderly populations and the absence of reliable data on elderly Brazilians in this field, this study consisted of an evaluation of TMD prevalence and the self-perception of oral health among institutionalised and community-dwelling elderly in Sã o José dos Campos, Brazil. Methods: Two hundred and fifteen community-dwelling and 185 institutionalised elderly people were evaluated by the Helkimo anamnestic (Ai) and clinical dysfunction (Di) indices and answered a questionnaire using the Geriatric Oral Health Assessment Index (GOHAI). Results: The major prevalence of TMD symptoms was for the Ai0 (symptom-free) group (69.5%), while the major prevalence of clinical signs was for the DiI (mild) group (56%). Women presented a higher AiII classification than men (v 2 test, p = 0.049). Community-dwelling elderly presented a significantly lower Ai0 classification than the institutionalised ones (Two ratios equality test, p < 0.001). There was no relationship between the institutionalised status and the clinical dysfunction index for Di0 and DiIII classification (Two ratios equality test, p = 0.194 and 0.535 respectively). The institutionalised elderly presented greater (One-way ANOVA = 0.005) self-perception of oral health (33.45) than did the community-dwelling group (32.66). There were only weak Pearson's correlations among the anamnestic ()33.0%) or clinical ()14.7%) findings by the TMD and GOHAI indices. Symptom-free (Ai0) institutionalised elderly presented better scores in all GOHAI dimensions and elderly representing an absence of clinical TMD signs (Di0) presented higher GOHAI physical dimension scores in both groups. Conclusions: The prevalence of TMD symptoms among this sample of elderly individuals was relatively low, self-perception of oral health was reasonable and a weak, inverse correlation was found between TMD signs and symptoms and elderly self-perception of oral health measured by the GOHAI index.

In this paper an operational model for the automatic generation of lifelike gestures of an anthropomorphic virtual agent is described. The biologically motivated approach to controlling the movements of a highly articulated figure provides a transformation of spatiotemporal gesture specifications into an analog representation of the movement from which the animations are directly rendered. To this end, knowledge-based computer animation techniques are combined with appropriate methods for trajectory formation and articulated figure animation.

In this paper, the details of processing algorithms used in a training program with language learning-impaired children (LLI's) are described. The training program utilized computer games, speech/language training exercises, books-ontape and educational CD-ROM's. Speech tracks in these materials were processed using these algorithms. During a four week training period, recognition of both processed and normal speech in these children continually increased to near age-appropriate levels. We conclude that this form of processed speech is subject to profound perceptual learning effects and exhibits widespread generalization to normal speech. This form of learning and generalization contributes to the rehabilitation of temporal processing deficits and language comprehension in this subject population.

There have been contradictory ®ndings in the fragile X (fraX) literature about possible neurocognitive and psychological symptoms due to the fraX premutation (pM). The purpose of the present study was to investigate the relationship between CGG repeat length and neurobehavioral functioning in carriers of the fraX pM. Eighty-®ve female carriers of the pM with allele sizes ranging from 59±166 were administered a comprehensive IQ test (WAIS-III) and completed a questionnaire designed to measure psychopathology (Symptom Checklist (SCL)-90-R). No relationship between allele size and cognition was identi®ed. A signi®cant negative relationship between allele size and age was found, as well as a positive relationship between allele size and depression. Followup analyses separating small and large allele sizes (below and above 100 CGG repeats) indicated that individuals with larger allele sizes scored signi®cantly higher on the Interpersonal Sensitivity and Depression subscales of a self-administered checklist assessing psychological symptoms (SCL-90-R). Despite the limitation of few individuals with high CGG repeat lengths, our ®ndings suggest that females with larger premutated alleles (! 100 repeats) display some clinical manifestations of fraX syndrome.

We used event-related brain potentials (ERPs) to study effects of selective attention on the processing of attended and unattended spoken syllables and letters. Participants were presented with syllables randomly occurring in the left or right ear and spoken by different voices and with a concurrent foveal stream of consonant letters written in darker or lighter fonts. During auditory phonological (A P) and non-phonological tasks, they responded to syllables in a designated ear starting with a vowel and spoken by female voices, respectively. These syllables occurred infrequently among standard syllables starting with a consonant and spoken by male voices. During visual phonological and non-phonological tasks, they responded to consonant letters with names starting with a vowel and to letters written in dark fonts, respectively. These letters occurred infrequently among standard letters with names starting with a consonant and written in light fonts. To examine genuine effects of attention and task on ERPs not overlapped by ERPs associated with target processing or deviance detection, these effects were studied only in ERPs to auditory and visual standards. During selective listening to syllables in a designated ear, ERPs to the attended syllables were negatively displaced during both phonological and non-phonological auditory tasks. Selective attention to letters elicited an early negative displacement and a subsequent positive displacement (Pd) of ERPs to attended letters being larger during the visual phonological than non-phonological task suggesting a higher demand for attention during the visual phonological task. Active suppression of unattended speech during the A P and non-phonological tasks and during the visual phonological tasks was suggested by a rejection positivity (RP) to unattended syllables. We also found evidence for suppression of the processing of task-irrelevant visual stimuli in visual ERPs during auditory tasks involving left-ear syllables.

This study investigated the relation between linguistic and spatial working memory (WM) resources and language comprehension for signed compared to spoken language. Sign languages are both linguistic and visual-spatial, and therefore provide a unique window on modality-specific versus modality-independent contributions of WM resources to language processing. Deaf users of American Sign Language (ASL), hearing monolingual English speakers, and hearing ASL-English bilinguals completed several spatial and linguistic serial recall tasks. Additionally, their comprehension of spatial and non-spatial information in ASL and spoken English narratives was assessed. Results from the linguistic serial recall tasks revealed that the often reported advantage for speakers on linguistic short-term memory tasks does not extend to complex WM tasks with a serial recall component. For English, linguistic WM predicted retention of non-spatial information, and both linguistic and spatial WM predicted retention of spatial information. For ASL, spatial WM predicted retention of spatial (but not non-spatial) information, and linguistic WM did not predict retention of either spatial or non-spatial information. Overall, our findings argue against strong assumptions of independent domain-specific subsystems for the storage and processing of linguistic and spatial information and furthermore suggest a less important role for serial encoding in signed than spoken language comprehension.

This study tested the relationship between prelinguistic vocalization and expressive vocabulary 1 year later in young children with mild to moderate developmental delays. Three vocalization variables were tested: rate of all vocalization, rate of vocalizations with consonants, and rate of vocalizations used interactively. The 58 toddlers in the study were 17–34 months old, not sensory impaired, and had Bayley Mental Development Indices (Bayley, 1969; Bayley, 1993) from 35–85. In addition, the children had fewer than 3 words in their expressive vocabularies and during classroom observation each showed at least one instance of intentional prelinguistic communication before testing. Selected sections of the Communication and Symbolic Behavior Scales procedures (CSBS; Wetherby & Prizant, 1993) were administered at the beginning and at the end of the study. The vocal measures were obtained in the initial CSBS session. One measure of expressive vocabulary was obtained in the CSBS session ...

Children with developmental disabilities often converse less frequently than their developmentally matched peers. This low conversational participation can cause problems for the children’s future language and discourse development. The purpose of this experimental study was to test the hypothesis that adult topic-continuing wh-questions would elicit topic continuations in children with relatively low language ability, but not in children with relatively high language ability. Twenty-three children with developmental delays interacted with an adult who conducted two play sessions. In each session, the adult used a different interaction style. The two styles differed in the adult’s use of topic-continuing wh-questions. Results indicate that adult use of topic-continuing wh-questions supported the use of child continuations in children at all language levels. The type of continuations (single word versus multiword) that were elicited depended on the language level of the children. Cli...

Access the most recent version at doi: 1996 169: 148-159 The British Journal of Psychiatry PK McGuire, DA Silbersweig, I Wright, RM Murray, RS Frackowiak and CD Frith in schizophrenia: relationship to auditory verbal hallucinations The neural correlates of inner speech and auditory verbal imagery References http://bjp.rcpsych.org/cgi/content/abstract/169/2/148#otherarticles Article cited in: permissions Reprints/ permissions@rcpsych.ac.uk to To obtain reprints or permission to reproduce material from this paper, please write to this article at You can respond http://bjp.rcpsych.org/cgi/eletter-submit/169/2/148

The frequency of verbal perseverations of two children with severe visual impairment was reduced using differential reinforcement of appropriate speech coupled with extinction of perseverative utterances. A reversal design in which baseline and intervention were alternated in an A–B–A–B sequence was employed to show the functional relationship between the target behaviors and the treatment procedure. There was a marked increase in appropriate utterances and a correlated decrease in perseverative ones. These findings suggest that this singular feature of the language of certain children with visual impairment is, under natural conditions, probably maintained by the attention provided by their caregivers.

Three experiments were carried out to examine the cues that are used in learning to read and spell new words. In a reading task (Experiment 1), even preschoolers who could not read simple real words were able to benefit from print-sound relationships that were based on letter names. They found it easier to learn that the made-up word TM was pronounced as "team" (name condition) than that TM was pronounced as "tame" (sound condition) or as "wide" (visual condition). The letter-name strategy persisted among college students (Experiment 2). In a spelling task (Experiment 3), prereaders and novice readers again did better in the name condition than in the sound condition. The ability to use relationships based on letter sounds emerged later than the ability to use relationships based on letter names. However, sound-based relationships were used to a greater extent in spelling than in reading.

Children with the 22q11.2 deletion syndrome (22q11DS) are at an increased risk of psychiatric problems from pre-adolescence; little is known, however, about behavioral problems at a preschool age and the relationship between speech and behavior in this group. Parents of 90 children (aged 1.42-5.99 years) with 22q11DS filled out the Child Behavior Checklist, documenting behaviors including speech problems. Their profiles were compared with those of a comparison group consisting of 33 children with nonsyndromic orofacial clefts without 22q11DS, since both children with 22q11DS and children with clefts are expected to have speech problems. In the 22q11DS group, data on intelligence was acquired by means of formal tests. Parents of children with 22q11DS reported significantly higher mean scores on withdrawn behavior, affective problems and pervasive developmental problems compared to children with nonsyndromic clefts. Approximately 30% of children with 22q11DS had a score above the 97th percentile on at least one of the behavior subscales, indicating psychopathology. In children with 22q11DS, the reported behavioral problems were not associated with speech problems. Behavioral problems were found in 30% of young children with 22q11DS and were unlikely to be caused by speech problems. Within the 22q11DS group, behavioral problems were not related to the degree of cognitive impairment. This shows that many children with 22q11DS, known to be at an increased risk of psychiatric problems from pre-adolescence, already show behavioral problems before the age of 6 years.

Event-related potentials (ERPs) were recorded in cross-modal word fragment priming (CMWP) to address the function of pitch for the identification of spoken words. In CMWP fragments of spoken words (e.g., re taken from Regal [Engl. shelves]) are immediately followed by visual targets. Together with reduced reaction times (RTs), an ERP deflection named P350 has been found to be reduced for targets, which match the primes (e.g., in the prime -target pair re -REGAL) as compared to unrelated targets (e.g., re -WIRBEL [Engl. burble]). The P350 has been related to facilitated lexical identification [Friedrich, Kotz, Friederici and Gunter (in press), ERPs reflect lexical identification in word fragment priming, JOCN]. In the present study, we presented syllable primes with different pitch contours. One version of each prime bore a stressed pitch contour (e.g., re_1), the other an unstressed pitch contour (e.g., re_2). Primes were combined with targets being either stressed on the first syllable (e.g., REgel [Engl. rule]) or on the second syllable (e.g., reGAL [Engl. shelves]). We found a reduced amplitude of the P350 and slightly faster reactions for targets with a stress pattern that matched the pitch of the primes (e.g., re_1 -REgel) as compared to targets with a stress pattern that did not match the pitch of the primes (e.g., re_1 -reGAL). The present study replicates the P350 effect with different material, and indicates that pitch is used for lexical identification in spoken word recognition. D 2004 Elsevier B.V. All rights reserved.

A recent innovation in televised election debates is a continuous response measure (commonly referred to as the &quot;worm&quot;) that allows viewers to track the response of a sample of undecided voters in real-time. A potential danger of presenting such data is that it may prevent people from making independent evaluations. We report an experiment with 150 participants in which we manipulated the worm and superimposed it on a live broadcast of a UK election debate. The majority of viewers were unaware that the worm had been manipulated, and yet we were able to influence their perception of who won the debate, their choice of preferred prime minister, and their voting intentions. We argue that there is an urgent need to reconsider the simultaneous broadcast of average response data with televised election debates.

The vast majority of cochlear implant recipients realize significant improvement in speech perception. However, there continue to be a small group that does not realize such a benefit. In an effort to identify possible predictors for this, we have compared pre-and postimplant audiologic data using Hearing In Noise Test (HINT), City University of New York (CUNY), or Central Institute for the Deaf (CID) scores for 445 consecutive English-speaking adult patients followed for a minimum of 1 year postimplantation in two distinct groups, poor versus excellent performers.

The function of perseverative speech for an adult man who had been diagnosed with autism and mental retardation was examined. Results showed that verbal perseverations were maintained by social attention. An intervention consisting of differential reinforcement of appropriate verbal responses and extinction of perseverative verbal responding was effective in decreasing verbal perseverations.

Background: Patients who require positive pressure ventilation through a tracheostomy are unable to phonate due to the inflated tracheostomy cuff. Whilst a speaking valve (SV) can be used on a tracheostomy tube, its use in ventilated ICU patients has been inhibited by concerns regarding potential deleterious effects to recovering lungs. The objective of this study was to assess end expiratory lung impedance (EELI) and standard bedside respiratory parameters before, during and after SV use in tracheostomised patients weaning from mechanical ventilation. Methods: A prospective observational study was conducted in a cardio-thoracic adult ICU. 20 consecutive tracheostomised patients weaning from mechanical ventilation and using a SV were recruited. Electrical Impedance Tomography (EIT) was used to monitor patients' EELI. Changes in lung impedance and standard bedside respiratory data were analysed pre, during and post SV use. Results: Use of in-line SVs resulted in significant increase of EELI. This effect grew and was maintained for at least 15 minutes after removal of the SV (p < 0.001). EtCO 2 showed a significant drop during SV use (p = 0.01) whilst SpO 2 remained unchanged. Respiratory rate (RR (breaths per minute)) decreased whilst the SV was in situ (p <0.001), and heart rate (HR (beats per minute)) was unchanged. All results were similar regardless of the patients' respiratory requirements at time of recruitment. Conclusions: In this cohort of critically ill ventilated patients, SVs did not cause derecruitment of the lungs when used in the ventilator weaning period. Deflating the tracheostomy cuff and restoring the airflow via the upper airway with a one-way valve may facilitate lung recruitment during and after SV use, as indicated by increased EELI.

Multi-Modal Interface Systems (MMIS) have proliferated in the last few decades, since they provide a direct interface for both Human Computer Interaction (HCI) and face-to-face communication. Our aim is to provide users without any prior 3D modelling experience, with a multi-modal interface to create a 3D object. The system also incorporates help throughout the drawing process and identifies simple words and gestures to accomplish a range of (simple to complex) modeling tasks. We have developed a multi-modal interface that allows users to design objects in 3D, using AutoCAD commands as well as speech and gesture. We have used a microphone to collect speech input and a Leap Motion sensor to collect gesture input in real time. Two sets of experiments were conducted to investigate the usability of the system and evaluate the system performance using Leap Motion versus keyboard and mouse. Our results indicate that performing a task using speech is perceived exhausting, when there is no shared vocabulary between man and machine, and the usability of traditional input devices supersedes * Farzana Alibay was a MRes exchange student at Macquarie University from Institut Image, France.

The recent literature on language intervention has become increasingly focused upon developing treatments that more closely parallel normal language acquisition. However, there have been relatively few reports that directly compare imitative procedures to conversational-interactive interventions. The purpose of the present study was to compare the relative effectiveness of imitative intervention and conversational recast language intervention applied to a wide range of grammatical morpheme and complex sentence targets in 21 children with specific language impairment. The results indicated that although both kinds of treatments were effective in triggering acquisition of most targets, consistently fewer presentations to first spontaneous use were required in the conversational procedure. In addition, the transition from elicited production to generalized spontaneous production was more rapid under conversationinteractive treatment. Finally, although imitation treatment was more effec...

Extractive speech summarization, which purports to select an indicative set of sentences from a spoken document so as to succinctly represent the most important aspects of the document, has garnered much research over the years. In this paper, we cast extractive speech summarization as an ad-hoc information retrieval (IR) problem and investigate various language modeling (LM) methods for important sentence selection. The main contributions of this paper are four-fold. First, we explore a novel sentence modeling paradigm built on top of the notion of relevance, where the relationship between a candidate summary sentence and a spoken document to be summarized is discovered through different granularities of context for relevance modeling. Second, not only lexical but also topical cues inherent in the spoken document are exploited for sentence modeling. Third, we propose a novel clarity measure for use in important sentence selection, which can help quantify the thematic specificity of each individual sentence that is deemed to be a crucial indicator orthogonal to the relevance measure provided by the LM-based methods. Fourth, in an attempt to lessen summarization performance degradation caused by imperfect speech recognition, we investigate making use of different levels of index features for LM-based sentence modeling, including words, subword-level units, and their combination. Experiments on broadcast news summarization seem to demonstrate the performance merits of our methods when compared to several existing well-developed and/or state-of-the-art methods.

American English ͑AE͒ speakers' perceptual assimilation of 14 North German ͑NG͒ and 9 Parisian French ͑PF͒ vowels was examined in two studies using citation-form disyllables ͑study 1͒ and sentences with vowels surrounded by labial and alveolar consonants in multisyllabic nonsense words ͑study 2͒. Listeners categorized multiple tokens of each NG and PF vowel as most similar to selected AE vowels and rated their category "goodness" on a nine-point Likert scale. Front, rounded vowels were assimilated primarily to back AE vowels, despite their acoustic similarity to front AE vowels. In study 1, they were considered poorer exemplars of AE vowels than were NG and PF back, rounded vowels; in study 2, front and back, rounded vowels were perceived as similar to each other. Assimilation of some front, unrounded and back, rounded NG and PF vowels varied with language, speaking style, and consonantal context. Differences in perceived similarity often could not be predicted from context-specific cross-language spectral similarities. Results suggest that listeners can access context-specific, phonetic details when listening to citation-form materials, but assimilate non-native vowels on the basis of context-independent phonological equivalence categories when processing continuous speech. Results are interpreted within the Automatic Selective Perception model of speech perception.

Breathing disorders are generally associated with the lung cancer disease. Through daily treatment, lung cancer patients often use a traditional spirometer to measure their lung capacity. However, the use of a spirometer device for accurate measurement requires some sort of training and adjustment, which may be inconvenient for certain groups of patients, especially the elderly. In addition, the spirometer readings can become unreliable if the measurements are not taken as instructed. On the other hand, a microphone, say the microphone on a hand-held device such as a smart-phone, can easily capture the acoustic signal of breath without certain instructions. This signal can then be processed to estimate the lung capacity. In this paper, we propose a methodology through the simply recorded acoustic signal of breath, splitting the breathing cycle to inhale, pause, exhale and pause phases to measure the depth of the breath along with the time duration and signal energy of the breathing phases. We show how these computed parameters are used to estimate the lung size with a high degree of accuracy. This work is part of a virtual reality platform embedded within a smart-phone to assist lung cancer patients regulate their breath. Furthermore, the lung capacity estimation methodology proposed in this paper can also be used to aid patients with other breathing disorders.

This paper proposes and evaluates the use of linguistic information in the pre-processing phase of text classification. We present several experiments evaluating the selection of terms based on different measures and linguistic knowledge. To build the classifier we used Support Vector Machines (SVM), which are known to produce good results on text classification tasks. Our proposals were applied to two different datasets written in the Portuguese language: articles from a Brazilian newspaper (Folha de So Paulo) and juridical documents from the Portuguese Attorney General's Office. The results show the relevance of part-of-speech information for the pre-processing phase of text classification allowing for a strong reduction of the number of features needed in the text classification.

The goal of this study was to examine the linguistic correlates of social anxiety disorder (SAD). Twenty-four individuals with SAD (8 of them with a generalised subtype) and 21 non-anxious controls were asked to give speeches in front of an audience. The transcribed speeches were examined for the frequency of negations, I-statements, we-statements, negative emotion words, and positive emotion words. During their speech, individuals with either SAD subtype used positive emotion words more often than controls. No significant differences were observed in the other linguistic categories. These results are discussed in the context of evolutionary and cognitive perspectives of SAD.

We conducted two experiments to assess the role of referential speech during sign training in which the spoken words corresponding to signs were receptively known to the participants. An alternating treatments design was used to compare sign acquisition across two teaching conditions in which referents were presented either with or without the corresponding verbal label. During the first experiment, signs were taught concurrently; during the second experiment, signs within each of the respective conditions were taught in a serial fashion. In both experiments, signs taught by total communication were acquired faster than those taught by sign-alone training.